INQUA-COMMISSION FOR THE STUDY OF THE HOLOCENE Sub-Commission on Data-Handling Methods Newsletter 15 January 1997 CONTENTS Note from the Old Coordinator------------------ p. 1 Note from the Old and New Coordinators--------- p. 1 The Future of Tilia, Eric Grimm --------------- p. 2 The Environmental Challenge for Numerical Palynology, David G. Green ------------------ p. 3 Plot2, Plotover, and Plotlim; Specialty Programs for Special Problems, Louis Maher -- p. 6 Useful Programs for the Psion series 3a/c, K. D. Bennett ------------------------------- p. 9 New Bookshelf 12, H. J. B. Birks -------------- p. 11 Multimedia Science Presentations and Teaching Aids, John Matthews ----------- p. 12 Voices from the Past; Pollen Analysis and Pollen & Spore Circulars-------------------- p. 16 Tst2norm Determines the Overlap in Two Normal Populations, Louis Maher ----- p. 16 CNT-WIS.XLM: An Excel Macro for Concatenating Polcount Files for Importation into Tilia, Richard Telford --------------------- p. 18 How do I get On | Off the List?---------------- p. 19 ---------------------------------------------------------------------------- NOTE FROM THE OLD COORDINATOR When the INQUA Commission for the Study of the Holocene changed the Working-Group on Data-Handling Methods to a Sub-Commission, I agreed to coordinate a couple of newsletters before passing the task (and the honor) on to somebody else. I am very pleased to inform you that Keith Bennett has agreed to accept the job, and I am asking Holocene Commission President John Dodson to name Keith President of the Sub-Commission; I am pleased to remain on the Advisory Panel. The INQUA File Boutique has moved to Cambridge, England (see below), but I will maintain a mirror site for program downloads at http://www.geology.wisc.edu/~maher.html. It is a pleasure to have served as Newsletter Coordinator for the past six years; it is an experience I treasure. LJM ---------------------------------------------------------------------------- NOTE FROM OLD AND NEW COORDINATORS THE FUTURE OF THE DATA-HANDLING NEWSLETTER Louis J. Maher Department of Geology and Geophysics University of Wisconsin 1215 West Dayton Street Madison, WI 53706 USA E-mail: maher@geology.wisc.edu Keith Bennett Department of Plant Sciences University of Cambridge Downing Street Cambridge CB2 3EA, UK E-mail: kdb2@cam.ac.uk The increasing circulation of the newsletter has brought with it increasing costs, especially for mailing. INQUA's funding is limited, so it will not be possible to distribute future issues as paper copy by first-class post/air-mail to all recipients as now. An obvious alternative is to use various forms of internet distribution as the default mode for circulation, backed-up by paper distribution to those for whom internet access is currently difficult. A trial scheme is being set up (see below), but the first priority is to establish the scale of need for paper copies. This issue will be the last to be distributed on paper to the whole circulation list. If you are unable to receive future issues by any internet means, please let Keith Bennett know as soon as possible. This is essential, both to ensure that you continue to receive the newsletter, and for costing the circulation of paper copies. Keith would also be interested to have names and addresses of possible non-internet recipients who are not currently receiving the newsletter. The circulation list is heavily skewed towards North America, western Europe, and Australasia; we would like to broaden it in other parts of the world. The proposed scheme for future default distribution is to publish the newsletter on the World Wide Web. A trial has been set up using the past newsletters (http://www-palecol.plantsci.cam.ac.uk/inqua), which enables access to each newsletter (index page, then, sequentially, through the articles), and directly to all articles indexed by author or subject. You will even get to view graphics in color which can be a great help in conveying complex charts. There is also a search system, making it possible to find straightaway what Glen MacDonald is doing with his Jimi Hendrix CD (for example). The idea is that future issues will be added in [*p.1/p.2*] entirety, but fully linked to the rest of the system. Each issue will also be packaged in such a way that it can be downloaded as WWW pages, including any graphics, or as text pages (no graphics). Everyone on the distribution list will be notified that there is a new issue. It would be possible to e-mail issues directly to anyone whose WWW access is not adequate for on-line browsing, but it will be necessary to explicitly request this. All articles past and present will be marked as copyright of the author. If past authors are unhappy with the placing of their articles on WWW, those articles will be removed forthwith. But all future articles will be published in WWW format. WWW publishing of the newsletter has two major advantages over the current paper distribution: a potential considerable widening of the readership, and making all articles instantly available and searchable. The major disadvantage is enhancing the present skew in the readership towards more developed parts of the world (and those with better facilities in those parts). We are acutely conscious of this difficulty, and would be extremely grateful for any suggestions of names for paper distribution, or of ways to mitigate the difficulty. And, of course, please contact Keith Bennett with any comments on the WWW pages, and with copy for new articles! ---------------------------------------------------------------------------- THE FUTURE OF TILIA Eric Grimm Illinois State Museum Research and Collections Center 1011 East Ash St Springfield, IL 62703 E-mail: grimm@museum.state.il.us As all lovers (and haters) of Tilia know, one of the shortcomings has been the lack of a full-fledged manual. Thus, Lou asked if I would write a Tilia 2 "starter manual" for the Data-Handling Methods Newsletter. While this is a great idea, it would soon be obsolete because of upcoming changes in Tilia. So, instead, at the risk of being accused of producing vapor-ware, I am providing a report of Tilia plans and developments. Tilia and Tilia graph are DOS programs. However, in the current world of Graphical User Interfaces and Windows 95, such "legacy" DOS programs are ancient technology and are rapidly disappearing from the scene. Thus, I have begun work on the Windows versions of these programs. But first, I will describe the status of Tilia 2. I have been making Tilia 2 available as various beta versions for some time. These are available over the Internet by anonymous FTP and WWW. Tilia 2 is fully functional, and anybody can download and use it. However, the companion Tilia graph requires the licensed device drivers originally purchased with Tilia 1.x, and thus cannot be used unless you have previously installed an earlier version of Tilia. Soon, hopefully by the time this newsletter reaches you, a "final" version of Tilia 2 will be available. Although this version will look much like the earlier versions, it has some substantial differences. In particular, it is a DOS 16-bit protected mode application, which means it will run in extended memory. Thus, it will not be limited to the RAM CRAM problems many have had with earlier versions, which were confined to the lower 640 kb available to normal DOS programs. Enormous spreadsheets will now be possible. Also, CONISS and CA are completely integrated into Tilia, and are not independent applications as they were previously. Tilia graph will again be assessable from the main Tilia menu. A possible disadvantage is that the program will require a probable minimum of 4 MB memory, and consequently will not run on older 286 and 386 machines with less than this amount, which, in any case, are rapidly becoming obsolete. The "final" Tilia 2 version will also use an updated graphics library and will be released with some new device drivers, including an HP LaserJet driver capable of 600 dpi and a CGI Metafile driver. The later will facilitate simple transfer of TG diagrams to other graphics software. Microsoft PowerPoint, for instance, will import CGI metafiles. If you use one of the standard GKS fonts (e.g. simplex), PowerPoint will allow you to easily change the font to any standard Windows font. While these features will be apparent to the user, many of the changes are "under the hood." I have converted the program from C to C++ and have been encapsulating much of the functionality to facilitate the move to Windows. Although I could be spending my time writing a manual for this "final" version of Tilia 2, I feel I am much more productive by using my limited time for programming on the Windows version. The menus and dialog boxes will be significantly different from the DOS version, although the functionality will be much the same. The new menu system will follow Windows standards and, therefore, ought to be more user-friendly. The new version will be a 32-bit Windows 95 application, which will not run under Windows 3.1 (or Win32), but Windows 95 is rapidly [*p.2/p.3*] becoming standard and comes with virtually all new computers. The Windows 95 version will also have a completely integrated help system and on-line manual (yeah!). It will be backwardly compatible with earlier versions of Tilia (i.e. it will read your old .til files). The plan is to first release a Windows 95 version of Tilia graph, the program most limited by DOS. The company that produced the graphics library I used for the DOS version has now produced a Windows 95 DLL, which is greatly facilitating the conversion. The two greatest advantages of Windows version to the user is that the program will use standard Windows device drivers (and print manager) and TrueType fonts. Thus, any device with a Windows 95 driver (virtually everything) will be compatible with Tilia graph. The Windows 95 version of Tilia will follow. In fact, this program will be a combined Tilia/Tilia graph application. I plan to make the new versions available over the Internet to previous owners of Tilia/Tilia graph. Watch the Tilia list-server for announcements and details (send a message to listproc@lists.colorado.edu with the contents of the message being "subscribe tilia-l " without quotes and <> symbols. Example: subscribe tilia- l Paul N. Pieker). ---------------------------------------------------------------------------- THE ENVIRONMENTAL CHALLENGE FOR NUMERICAL PALYNOLOGY David G. Green Johnstone Centre Charles Sturt University PO Box 789 Albury, NSW 2640 Australia E-mail: dgreen@life.csu.edu.au Introduction From Galileo turning a telescope on the heavens to DNA recombination, new analytic methods in science have always led to new discoveries. The past two decades have seen the introduction of many new techniques in numerical palynology. In some cases these have produced new insights; in other cases the potential has not yet been fully realized. In any field of endeavour it is useful to take stock from time to time and identify major opportunities and challenges. In this short account I discuss some recent developments and the challenges that they pose for future research in palynology. This is in no sense a review, but rather an exploration of how numerical palynology could help to address some of the major environmental issues of our time. Some environmental needs and challenges Environmental management is one of the great issues of our time. The problems include both sustaining the productivity of the world's agricultural systems and conserving the world's ecosystems and biodiversity. The critical nature of these issues has seen them become an integral part of our culture and brought them to the forefront of economic and political decision-making. Most areas of environmental science are today linked to these issues in one way or another. Palynology has much to offer environmental research - and management. Numerical palynologists especially should try to meet the challenges posed by trying to provide environmental information of the kind needed. Environmental monitoring One area where pollen records have always been used is in tracing past climate changes. Given recent concerns about the greenhouse effect, the need for accurate data of this kind has never been more strongly felt. Most importantly pollen analysis offers the possibility of placing current climatic fluctuations in the context of "natural" climatic variations across a range of time scales. A number of authors, such as Webb and Birks, have developed numerical methods for inferring climatic variables from pollen spectra. As well as recording climate change, pollen records also indicate some of the effects of climate change. They offer a means to monitor how forests and agricultural systems respond to global changes of all kinds. Besides climate effects, these changes include atmospheric pollution, land clearing, spread of weeds, increasing salinity, soil degradation and fire. Pollen records provide a temporal dimension that complements the kinds of data that can be gathered by remote sensing or field surveys. Pollen analysis also provides rich sets of historical data with which environmental models may be calibrated, tested and improved. Studies of this kind were pioneered by Al Solomon and his colleagues during the 1970s. The link to field ecology There is a real need to link palynological studies to field ecology. The resolution of pollen records - often hundreds of years worth of sediment between samples - is usually so poor that most ecological processes are obscured. Most [*p.3/p.4*] ecologists tend to ignore pollen analysis for this reason. And yet at the same time ecology suffers because field studies tend to focus on very short time scales. However we cannot hope to understand current processes without placing them in the context of historical vegetation change. In Australia for instance, bushfires often leave behind the charred skeletons of trees that predated European settlement. A vast gap exists between the time scales considered by field ecologists (at most a few decades) and those normally considered by pollen analysts (usually thousands of years). Even a single sample (typically 10-20 years worth of lake mud) would average out all the data in most long-term field studies, and the time period represented by the spacing between adjacent samples (often 100-200 years) is greater than the time span within which many ecological processes occur. For these reasons, the potential of pollen analysis to be a source of present-day environmental information has often been overlooked. During the 1980s I was involved in several studies that demonstrated the feasibility of combining studies of contemporary pollen records with field ecology (e.g. Green et al, 1988). Biodiversity Many programmes are now underway around the world to gather information about the distributions of plants and animals (Green 1994). See for example, the International Organization for Plant Information (IOPI), which is trying to database the world's plants (http://iopi.csu.edu.au/iopi). These data are needed for many purposes, such as identifying endangered (or potentially endangered) species. Conservation programmes need to be able to predict potential distributions of species, either to locate rare species or to assess possible relocation areas. Some good examples can be found on the site run by Australia's Environmental Resources Information Network (ERIN) - http://www.erin.gov.au/. Pollen studies can help to calibrate and test distribution models by providing data on current presence or absence, former distributions, and especially the effects of disturbance and other non-equilibrium factors. Methodological issues Automated pollen counting One of the greatest handicaps to really detailed studies, such as fine resolution pollen analysis (see below), is the sheer time and effort involved in counting pollen. To make such studies practical on a routine basis, pollen counting really needs to be automated. Over the last 30 years there have been sporadic attempts to automate various aspects of pollen analysis, especially counting (e.g. Walker et al. 1968; Langford, et al. 1986). These have met with varying degrees of success, but no system has yet achieved widespread use. Identifying pollen grains is one of the more difficult exercises for pattern recognition. Every palynologist is well aware of how difficult pollen identification can be, even for a human expert. Although it may be years before any system can match human performance, there are many simple tasks that systems could already tackle, such as estimating total number of grains, or the volume of charcoal in a sample. One alternative to full automation is computer-assisted counting. For instance, the computer could locate grains and list its best guesses for each one. This would be particularly effective if the system could tally the grains that it can identify with high confidence and refer only the more obscure cases to the operator. There are now many systems of this kinds, using machine-learning systems such as artificial neural networks, bayesian networks, and decision trees. Automation remains the key technical issue in pollen analysis. Without automation many of the prospects discussed here remain time-consuming and expensive. With automation all sorts of studies become cost-effective. On-line analytic resources One of the more exciting aspects of the Internet is that is makes possible collaboration on a previously unimaginable scale. Elsewhere I have argued the case for researchers to form subject-based information networks (Green and Croft, 1994; Green 1995), especially in palynology (Green and Stocker, 1995). Such a network would provide an efficient medium for communicating results to colleagues. Sharing resources could also help to cut research costs. For instance at present each research group has to duplicate a effort required to compile pollen identification keys. On-line keys could reduce the time and effort required. Furthermore having keys available from different areas would simplify the problem of finding references to identify exotic types. Other novel possibilities include on-line data processing (Green Wesley & Peters, 1994). One of the great problems for numerical palynology is to make methods widely accessible to palynologists. The network is a perfect medium for providing specialized analyses that most [*p.4/p.5*] people would use only occasionally (e.g. statistical zonation). Data mining One of the exciting prospects that stems from cooperation on the Internet is that new types of research become possible. One of these is data mining - the extraction of useful information by sifting and analysing large databases. One important principle is what I term the "serendipity effect". If you combine any two pieces of data (e.g. overlay a vegetation map on a soils map), then you can immediately ask new questions with your data. If you go on combining more and more datasets, then the number of potential questions that can be addressed - and the number of unexpected insights - goes up exponentially. For instance, an important stage in the interpretation of a pollen profile is to compare it with records from other sites nearby. However if we go further and combine the data sets from sites throughout a region, then we can address even more questions. A good example of this was the compilation, during the 1970s and 1980s, of Holocene pollen records for North America and their use to reconstruct past climates and postglacial migration patterns (Davis 1976; Webb 1979). Note that while this example required a special effort to compile a one-off database, the Internet provides a medium in which potentially all relevant datasets could be published in a standard format on a routine basis. Perhaps the supreme example of what can be achieved through public domain resources are molecular biology databases, such as Genbank and EMBL (Green 1993). Comparing new DNA and protein sequences against these databases is now a standard part of any genomic study. So useful have the databases proved that most leading journals now make it compulsory for authors to submit their data to one of the major databases before their results are accepted for publication. Fine resolution pollen analysis During the 1970s and 80s, research by a host of researchers led to a new sub-discipline known as fine resolution pollen analysis (Green, 1983; Green and Dolman, 1988). Technical advances, especially in methods of dating sediments, now make it possible to sample sediments, and to interpret the results, in very fine time steps. Several studies (e.g. Green et al. 1988) have yielded pollen records with single-year time steps. Even a modest improvement in sampling can reveal short-term events that are missed entirely by coarser sampling. Whilst the problems of achieving precise chronological control are formidable, innovations in field and laboratory techniques have resolved many difficulties. However much research is still needed about interpreting pollen records at fine time intervals: it is far more demanding than simply "eye-balling" a pollen diagram. For instance, by altering both the rate at which trees produce pollen and the rate at which pollen is delivered into the sediment, variations in annual rainfall may affect pollen quantities without affecting the size of a tree population at all. The approach that I have promoted to cope with such problems is to represent hypotheses about plant ecology as mathematical models, which can then provide guidelines for interpreting pollen diagrams. An important consequence of fine time control is that it is possible to compare pollen records directly with independent sources of information, such as weather records, field observations or local histories. Instead of documenting long-term vegetation histories, fine resolution pollen studies can identify short-term vegetation events and processes. The emphasis on small-scale events applies to spatial scales, as well as time. An example is work (e.g. by Richard Bradshaw in Britain) on the local events that are recorded in boggy forest hollows and small ponds. Perhaps we should be broadening the focus of traditional pollen studies on lakes and bogs as study sites. In his PhD research Gary Dolman showed that in dam sediments it is sometimes possible to identify individual storm events. During the 1980s Bob Wasson and Robin Clark at CSIRO Canberra showed that because of the fast sedimentation rates reservoirs can be a source of useful records about the recent past. Conclusion Palynology can make a significant contribution to environmental management. Pollen analysis can detect the impacts of human activity on vegetation in sufficient detail to relate historical events with their effects on vegetation. Perhaps its greatest contribution so far has been to show that environmental degradation has led to the fall of past civilizations. The lesson is, if we are wise enough to listen, that it is a grave mistake to confuse slow environmental change with stability. We cannot continue to abuse the environment with impunity. My argument here is that palynology has the potential to contribute much more than it already has. To summarize the main points raised above I argue that: [*p.5/p.6*] ù High priority should be given to solving certain methodological hurdles in palynology, especially automation. ù Sharing of information and resources via the Internet has the potential to make pollen studies more cost effective and to make possible various new types of studies. ù Palynological studies can and should contribute to efforts to address environmental issues. In particular greater effort should be devoted to monitoring current processes and their recent history. Others may disagree with the above assessment or priorities and there are many issues that I have not addressed here. If nothing I hope that this discussion will lead others to think about where their field of research is heading as we enter the Twenty-first Century. References. Davis, M. B. 1976. Pleistocene biogeography of temperate deciduous forests. Geoscience and Man 13, 13-26. Green, D. G. 1983. The ecological interpretation of fine resolution pollen records. The New Phytologist 94, 459-477. Green, D. G. and Dolman, G. S. 1988. Fine resolution pollen analysis. Journal of Biogeography 15, 685-701. Green, D. G., Singh, G., Polach, H., Banks, J. C. G., Moss, D. and Geissler, E. 1988. A fine resolution palaeoecology and palaeoclimatology from southeastern Australia. Journal of Ecology 76, 790-806. Green, D. G. 1993. Databasing the world. INQUA - Commission for the Study of the Holocene, Working Group on Data-Handling Methods 9, 12-17. Green, D. G. 1994. Databasing diversity - a distributed, public-domain approach. Taxon 43, 51-62. http://life.csu.edu.au/~dgreen/papers/taxon.html Green, D. G. 1995. From honeypots to a web of SIN - building the world-wide information system. In: Tsang, P., Weckert, J., Harris, J. and Tse, S. (eds.): Proceedings of AUUG'95 and Asia-Pacific World Wide Web '95 Conference Charles Sturt University, Wagga Wagga, pp. 11-18. http://www.csu.edu.au/special/conference/apwww95/papers95/dgreen/dgreen.html Green, D. G. and Croft, J. R. 1994. Proposal for Implementing a Biodiversity Information Network. In Linking Mechanisms for Biodiversity Information. Proceedings of a Workshop for the Biodiversity Information Network, Campinas, Sao Paulo, Brasil. http://life.csu.edu.au/~dgreen/papers/proposal.html Green, D. G., Wesley, A. & Peters, D. 1994. Information analysis via the World Wide Web. INQUA - Commission for the Study of the Holocene, Working Group on Data-Handling Methods 12, 7-9. Green, D. G. and Stocker, R. 1995. SINQUA - a special interest network for Quaternary research. INQUA - Commission for the Study of the Holocene, Working Group on Data-Handling Methods 13, 23-27. Langford, M., Taylor, G. & Flenley, J. R. 1986. The application of texture analysis for automated pollen identification. Proceedings of a Conference on Identification and Pattern Recognition, Toulouse, pp.729-739. Univ. Paul Sabatini. Walker, D., Milne, P., Guppy, J. & Williams, J. 1968. Computer-assisted storage and retrieval of pollen morphological data. Pollen et Spores 10, 252-262. Webb, T. III. 1979. The past 11,000 years of vegetational change in eastern North America. BioScience 31, 501-506. ---------------------------------------------------------------------------- PLOT2.EXE, PLOTOVER.EXE, AND PLOTLIM.EXE; Specialty Programs for Special Problems By Louis Maher In an earlier newsletter, I described SLOTDEEP.EXE which I had written to use for correlating (slotting) pairs of pollen diagrams (Maher, 1993a). It utilized the two cores' dissimilarity coefficients while preserving their depth information. The user is shown an "exploded" representation of the dissimilarity matrix and allowed either to accept the computer's solution or manually to interact with patterns in the data. I have found many uses for SLOTDEEP, not the least for tracing sediment facies by their pollen assemblages in a single lake. There were two things about SLOTDEEP's presentation graphics that I dislike, although both were done for a purpose. In the initial view one site's pollen diagram is displayed above the other with each taxon aligned; this lets the user compare the sites visually. But to allow the same screen control information file (SLOTSEE.INF) to be used for many sites within a region, the diagram is plotted at a reduced scale so that unusual abundances do not result in overwriting one taxon's column by another. After the sites are correlated, it would be nice to display the diagrams at a larger scale. The second problem deals with showing the resulting correlation by intercalating the sites' levels by color-coded bars without an actual depth scale. Although this makes [*p.6/p.7*] an effective screen display, it would be nice to overlay the correlated site on top of the principal site at a common depth (or age) scale at the same time increasing the abundance scale for the percentage plots. I have corrected both problems by following my usual philosophy of making two small separate programs rather than combining everything into one large program that hogs so much memory on the older/simpler computers that there is no room left for the data. PLOT2.EXE solves the first problem, and PLOTOVER.EXE handles the second. As an example of their use, I will compare my Devils Lake diagram with the odd levels of the Kellners Lake diagram produced several years earlier by Bob Goodwin, a former student (Goodwin, 1976). Kellners Lake lies 110 miles (177 km) northeast of Devils Lake across Wisconsin's Tension Zone a significant vegetation boundary. Both sites had a number of 14C dates; their depths were converted to age in decades using DEP- AGE.EXE (Maher, 1992). Second-order functions produced good correlation coefficients for both sites. The two sites were then loaded into SLOTDEEP, and points of similar age were set and marked on the exploded matrix. For samples older than about 5000 yr B.P., the selected points closely matched the pattern of dissimilarity coefficients. The dissimilarity between the sites was much higher during the last half of the Holocene. SLOTDEEP then converted Kellners Lake's time scale to the Devils Lake scale, and the converted KELLCONV.DAT file was saved to disk. This final conversion was not really needed in this particular case, because both sites had already been converted to age from "well-behaved" sets of dates. When the two sites are loaded into PLOT2.EXE, the resulting display is shown in Fig. 1. The short horizontal line segments at the left and right margins measure 1000's of years along the y-axes (as they would measure meters had either site's y-axis been measured in cm rather than decades). The diagrams can be shown in separate colors, and there is a provision for making small changes in scale and site names without reloading the data.. It can be seen at a glance that both sites start with a spruce (Picea) zone and end with the ragweed (Ambrosia) zone that signals European settlement; i.e. they both span the Holocene. And at a first glance the two diagrams are very much alike, especially the lower halves. But on a second glance there are a number of differences: Quercus drops off in the upper part of Kellners; Fagus has lower values in Devils, and Ambrosia has a higher average at Devils. But an easier way to compare the correlated diagrams would be to superimpose them (Fig. 2), and that is the job of PLOTOVER.EXE. The user first loads the diagram that is to serve as a reference background. The second diagram is then loaded and displayed as an overlay. On the color screen the reference diagram is normally printed in a muted color whereas the overlay is printed in a bright contrasting color. When done in this manner, it is easy to see that there is a reasonably good match in the older part of the record; however, Fraxinus achieves higher values at Devils about 10 K B.P. at a time when there is more Pinus at Kellners. But the real differences start after about 5 K B.P. Betula and Pinus increase more at Kellners than at Devils. And the decreases in Ulmus, Ostrya, and Acer are less at Kellners than at Devils. Fagus and Tsuga are both much more abundant at Kellners than at Devils; both taxa appear [*p.7/p.8*] to have migrated westward into Wisconsin around the north side of Lake Michigan, reaching the Kellners region first. The pollen of both Fagus and Tsuga show up at Devils Lake, but only Tsuga trees have reached its vicinity. The "Other (7)" category includes four trees (Abies, Alnus, Carya, and Juglans) and the Poaceae, Cyperaceae, and Chenopodiaceae; the NAP taxa are more common in the Devils Lake record during the latter part of the Holocene than they are at Kellners. I find these two sites to be instructive. Although they occur on opposite sides of the Tension Zone vegetation/climate boundary, they show that pollen gets across the boundary, and that it mutes the differences of the vegetation at both sites. That is good in that the blurring helps us correlate the sites. It is bad because we receive mixed signals at both sites; did the proportion of, say, Betula increase in the vegetation around Devils Lake during the last three thousand years, or is it the result of more north winds or more birch trees growing farther north? And do the differences of the pollen diagrams at these two sites show that Wisconsin's Tension Zone came into existence half way through the Holocene? While on the topic of specialty programs, I have long been interested in how the size of the pollen count affects the precision with which a taxon's abundance can be estimated. Mosi mann (1965) had worked out the mathematical relationships at least ten years before microcomputers made it feasible to calculate and display confidence limits on pollen diagrams on a routine basis. About eight years ago I modified my PLOTSITE.EXE [*p.8/p.9*] program so that it would calculate and plot 0.95 confidence limits for the pollen taxa. I call it PLOTLIM.EXE. I found it made a good vehicle for demonstrating the built-in uncertainty in pollen data even should all the grains be identified perfectly. Fig. 3 is an example plot made with PLOTLIM using the Kirchner Marsh Diagram that Tom Winter and Herb Wright made famous; its data are available in the file T50RAW.EXE that is in the INQUA File Boutique (Maher, 1993b). On the computer screen the text, the pollen columns, and the 0.95 C.I. can be colored independently for clarity. The portion of the percentage column to the left of the confidence interval is dotted to help differentiate it from the interval itself in black and white plots. Note the very wide confidence intervals in the lower three or four levels; this provides a quick clue to the low pollen counts in those samples. I have placed PLOT2.EXE, PLOTOVER.EXE, PLOTLIM.EXE, and some example data in the INQUA File Boutique in a self- extracting file called PLOTSZ.EXE. They all use VGA graphics. References. Goodwin, R. G., 1976, Vegetation response the Two Rivers Till advance based on a pollen diagram from Kellners Lake, Manitowoc County, Wisconsin, Unpublished M.S. Thesis (Geology & Geophysics), University of Wisconsin, Madison, 39 p. Maher, L. J., Jr., 1992, Depth-age conversion of pollen data, INQUA - Commission for the Study of the Holocene, Working Group on Data-Handling Methods Newsletter 7:13-17. Maher, L. J., Jr., 1993a, SLOTDEEP.EXE: Manual correlation using the dissimilarity matrix, INQUA - Commission for the Study of the Holocene, Working Group on Data-Handling Methods Newsletter 9:21-26. Maher, L. J., Jr., 1993b, Keep the heart of the NAPD on a single HD floppy, INQUA - Commission for the Study of the Holocene, Working Group on Data- Handling Methods Newsletter 10:15-17. Mosimann, J. E. 1965, Statistical methods for the pollen analyst, in B. Kummel and D. Raup (eds). Handbook of Paleontological Techniques, Freeman and Co., San Francisco, pp 636-673. ---------------------------------------------------------------------------- USEFUL PROGRAMS FOR THE PSION SERIES 3A/C K. D. Bennett Department of Plant Sciences University of Cambridge Downing Street Cambridge CB2 3EA, UK E-mail: kdb2@cam.ac.uk In the newsletter of January 1990, I described a program called POLLTAX for a pocket computer produced by Psion plc (Bennett 1990). The same company subsequently introduced a new series of machines, Series 3, which are a substantial improvement in all respects. The current model, introduced in October 1996, is the Series 3c, but it represents only minor changes on the Series 3a, which has been available for some time. This note describes the basic features of the Series 3a, provides access to more information, and introduces three programs of mine that might be useful to readers of this newsletter. The Psion Series 3a is a palmtop, or hand-held, computer, weighing only 275g (including batteries) and measuring 165x85x22mm. It is available in four RAM sizes, 256kb, 512kb, 1Mb, and 2Mb, and has two slots for addition of extra memory, usable as RAM. Much the cheapest way to buy memory is buy it in the machine, so I shall discuss a 2Mb Series 3a for the rest of this note. The operating system is windows-like: a series of icons, with the active icon highlighted. The keyboard has a 57-key QWERTY-pattern, capable of generating 244 characters (including the IBM extended character set). Power comes from two AA batteries, with a life, so Psion say, of up to 80 hours, but I rarely get more than 50 hours. There is also a backup power supply in the form of a small camera-type battery, which supports the system if the regular batteries fail for any reason. An external power supply is also available, but I have not tried this. The Series 3a comes with a number of applications pre-loaded. These include a Word-compatible word processor, a Lotus-compatible spreadsheet, a diary, a database, a calculator, and a clock/alarm system. Multiple application can be run simultaneously. A serial link can connect it to a PC or Mac to download files, and a modem plus terminal emulation software make it possible to connect to any dial-up service, including an internet provider. Psion have just announced a product called PsiMail Internet, which is compatible with a standard Internet account, and allows on-line browsing of the World Wide Web: [*p.9/p.10*] (http://www.psion.com/productguide/scpsimailint.html). Two irritating deficiencies of my Series 3a are the rather clumsy means of seeing the directory structure (now much improved in the Series 3c), and the lack of a backlight (remedied in US models of the Series 3c, much to the irritation of UK Psion users). This machine is therefore a fully functioning computer, capable of most of the things one would expect from a computer, including connection to another computer for the things it cannot do. For scientific purposes, as with the old Organizer II, the key feature is a programming language, called OPL. This has been upgraded to handle the new windowing environment. Apart from the inbuilt application, there is a vast array of software available for the Series 3a, much of it free- or shareware. One that I find especially useful is a scientific calculator Calc 3a v 1.6 by Richard Schmidt (http://home.worldonline.nl/~rschmidt/). A good way to get a flavour of the quantity and quality of this software is to take a look at Steve Litchfield's WWW site, and especially his reviews of Psion software: (http://ourworld.compuserve.com/homepages/slitchfield/). I have written or modified three programs that may be of interest to palaeoecological users of the Series 3a or 3c. POLLTAX3 I have converted and extended my pollen-counting program, POLLTAX, into a system that runs on the Series 3a, renaming it POLLTAX3. It should also work on the Series 3c, but I have not tried it (yet). This program works similarly to the original, but differs in the following respects: ù Windows-like environment ù Increased number of pollen types (287) ù Choice of input and output files, to facilitate use of different lists of taxa codes for different samples and also for more user-friendly organization of output files ù On-line help ù Optional inclusion of condition codes (four) with each grain counted ù System of searching the list of taxa for a particular code It can be downloaded from URL: http://www-palecol.plantsci.cam.ac.uk/psion3a.html, where further details will be found. LLTOGR This package consists of four utility programs and a driver module for the Psion 3a/3c to make conversions from latitude and longitude to National Grid co-ordinates for Great Britain and Ireland. They are not finely polished programs, but working utilities that do a job as simply and as quickly as possible. Conversion from longitude and latitude is by exact calculation using formulae and constants from the publications of the Ordnance Surveys of the two areas. The reverse calculation is by successive approximation (make a guess, see how close it is, make a better guess, etc), and thus takes longer. I have tried to minimize this time by rearranging the equations, but there a number of trigonometrical calculations that seem to be unavoidable. Output has been tested against examples given by the Ordnance Surveys and against values from maps. The programs are available by anonymous ftp to ftp-palecol.plantsci.cam.ac.uk, in directory pub/psion/lltogr. There is a zipped package of the five executables, together with additional details about the program (lltogr.zip). The details are available separately as a file called README. PSGPS This is a program for connecting the Psion Series 3a/3c to a Magellan GPS and analysing the data. Steve Litchfield's WWW site: (http://ourworld.compuserve.com/homepages/slitchfield/) provides details of how to connect several other GPS systems to the Psion 3a/3c. The heart of PSGPS is built around Lou Maher's GPS.BAS (Maher 1994), converted to OPL, but then extended to take advantage of some of the Psion's features. The program works either by receiving data (NMEA 0183B format) from the GPS, and plotting it (so you can follow a route), or by analysing data previously collected, as described by Lou Maher. The Magellan XL is very handy to use in the field, and works comfortably within a car, preferably tucked above the dashboard, giving a clear view of the sky through the windscreen. This seems to work well on most cars (I have tested it in a Fiat Panda, Vauxhall Astra, and Trabant), and at speeds of up to 140km per hour (in Germany, but not in the Trabant). The Magellan has an output lead that can be connected to the Psion's serial lead, and once both machines are up and running, the Psion display plots the data received, at a scale chosen by the user. The advantage of doing this is that the Psion display is much better than the GPS provides, and is more flexible (because it is programmable). I have included a set of databases (Psion [*p.10/p.11*] .dbf format) of the latitudes and longitudes of the main cities in all continents: the locations of these can be indicated, as can marks at any point chosen by the user while the program is running or from a personal database. The utility of such a system for basic route-finding or following on the open road or in cities has to be seen to be believed. One day, all vehicles will have something like this built in. The system can, of course, also be used in a static location to show the spread of points received by the GPS (as Lou illustrated). For users in the UK or Ireland, the output display figures can be in their National Grid units instead of Latitude and Longitude. The program and associated databases are available by anonymous ftp to ftp-palecol.plantsci.cam.ac.uk, in directory pub/psion/psgps. There is a zipped package of the executable, together with additional details about the program and databases (psgps.zip). Further details of the Psion Series 3a and Psion Series 3c are available from Psion plc: (UK URL: http://www.psion.com/, US URL: http://www.psioninc.com/). I have set up a WWW page with details of other sources of information, including suppliers, newsgroups, and software: see http://www-palecol.plantsci.cam.ac.uk/psion.html. References. Bennett, K. D. 1990. Pollen Counting on a Pocket Computer. INQUA Working Group on Data-Handling Methods, Newsletter 3, 5. Maher, L. J. 1994. Using the Global Positioning System. INQUA Working Group on Data-Handling Methods, Newsletter 11, 23-28. ---------------------------------------------------------------------------- NEW BOOKSHELF 12 H.J.B. Birks E-mail: John.Birks@bot.uib.no The following recently published books may be of interest to readers of this Newsletter. Bailey, T. C. & Gatrell, A. C. 1995 Interactive spatial data analysis. Longman, Harlow. 413 pp. Paperback. (with INFO-MAP software diskette). Baxter, M. J. 1994 Exploratory multivariate analysis in archaeology. Edinburgh University Press, Edinburgh. 307 pp. Paperback. Christensen, R. 1996 Analysis of Variance, Design and Regression. Chapman and Hall, London. 587 pp. Everitt, B. S. 1996 Making sense of statistics in psychology. A second-level course. Oxford University Press, Oxford. 350 pp. Paperback. (with datasets on diskette). Good, P. 1994 Permutation tests. A practical guide to resampling methods for testing hypotheses. Springer, New York. 226 pp. Handley, M. & Crowcroft, J. 1995 The World Wide Web. Beneath the Surf. University College London Press, London. 198 pp. Paperback. Lindsey, J. K. 1995 Introductory Statistics. A Modelling Approach. Oxford Science Publications, Oxford. 214 pp. Paperback. Moore, P. D., Chaloner, B. & Stott, P. 1996 Global environmental change. Blackwell Science, Oxford. 244 pp. Paperback. Shao, J. & Tu, D. 1995 The Jackknife and the Bootstrap. Springer, New York. 516 pp. Wackernagel, H. 1995 Multivariate geostatistics. Springer, Berlin. 256 pp. Wildi, O & Orloci, L. 1996 Numerical exploration of community patterns (Second edition). SPB Academic Publishing. Amsterdam. 171 pp. Paperback. Williamson, M. 1996 Biological Invasions. Chapman and Hall, London. 244 pp. Zar, J. H. 1996 Biostatistical Analysis (Third edition). Prentice-Hall International, London. 662 pp + Appendices. Paperback. ---------------------------------------------------------------------------- Keith Bennett will e-mail you when the next newsletter is published. Does he have your current e-mail address? kdb2@cam.ac.uk ---------------------------------------------------------------------------- [*p.11/p.12*] MULTIMEDIA SCIENCE PRESENTATIONS AND TEACHING AIDS John Matthews Ohana Productions 23 Sherry Lane Nepean, ON K2G 3L4, Canada E-mail: af763@freenet.carleton.ca During the past several years, as a paleoecologist in the Terrain Sciences Division of the Geological Survey of Canada, I participated in the creation of several multimedia CD productions. All are designed to present large science projects in a manner that is suitable for both the layperson as well as useful to the specialist. To achieve this goal, they employ several software applications not often used together by scientists. These pages contain sample screens from these presentations as well as examples of how similar multimedia techniques could be used for CD presentations of databases. One of these proposed database presentations a Digital Pollen Reference Collection will likely be of special interest to palynologists among the Newsletter readers. Our initial exploration of the potential of the CD medium for presenting large science projects dealt with The Mackenzie Delta Borehole Project, a large multidisciplinary effort involving three deep boreholes drilled in the Mackenzie Delta region in 1992. The longest of the boreholes bottoms out in the mid-Miocene. All three yielded a wealth of data too much for reasonable presentation as a traditional paper publication. Our goal was to design a multimedia presentation that allowed the reader to extract snippets of information from various levels in each core while at the same time preserving a journal-style presentation for the technical reports. We also deemed it essential that the user have access to all of the raw data associated with the project as well as to background documents, such as those normally appearing only as titles in the reference section of a printed monograph. After examining existing science-based CDs, we decided to construct the presentation in a manner that it could run on all of the popular computer platforms without requiring the reader to possess any special software other than that provided on the CD. The figures shown here are tiny black and white replicas of the views one would see on the screen. After some introductory material, the user sees Fig. 1 of the Borehole CD. The three buttons are entry points for examination of the results and data from each of the boreholes. Fig. 2 appears when the Fossils and then Microfossils buttons are clicked. If one of the icons next to the core image is [*p.12/p.13*] clicked (e.g., the pollen icon at 5.93m level next page), additional information becomes available via the buttons at the bottom of the central window. Core Photo shows the sediments that were sampled. Diagram opens the full-screen pollen diagram. It is generalized and modified for a screen presentation; however, a more complete version, similar in style to a published pollen diagram, is found in the Reports section, which is reached by going to the Borehole Menu page. This redundancy illustrates one of the unique features of the Borehole presentation and others discussed below: information is presented in traditional document format as well as in small doses via icons, buttons and other links. The Palliser Triangle Global Change Project is another large multidisciplinary effort that is ideal for a multimedia CD presentation. It deals with the past climate of a critical part of western Canada. Unlike the Delta Borehole presentation, the target audience for the Palliser CD is very broad, including laypersons, residents of the Palliser region, students, teachers and research scientists. The Palliser presentation has benefitted from lessons learned in the construction of the Borehole CD. For example, the reader moves through the presentation by selecting icons or text labels rather than buttons. Much more attention has been devoted to the artistic style of the presentation, resulting in a product much more pleasing to the user. Experience also taught us the importance of producing a self-running demo of the presentation. This demo serves first as a very effective planning guide and storyboard. Its construction also provides most of the artwork required for the project. Finally, the demo very effectively promotes the fully interactive, final product. This last use is very important if one requires financial partners to produce the CD. In the case of the Palliser CD we require such partners because it is our intention to distribute the final CD free of charge. Each of the sponsors is offered space on the CD to advertise their activities/services. Many users of the Palliser Triangle CD will explore no further than the Overview section (Fig. 3), reached by the binocular icon. But it is our hope that a few readers will become curious enough to dig deeper and visit the Scientific Studies section. Use of jargon or references to reports not ordinarily available to the layperson pose the biggest barriers to such exploration. We deal with this problem in two ways. First, the Reports section contains a number of background documents, among them some very general "how to" and "why" documents. Second, the presentation includes a comprehensive glossary. Clicking on the words Parabolic Sand Dunes, Fig. 4, opens a glossary "tablet" containing a definition of the term. But note that the definition is not restricted to text. Like the example, it can also include diagrams or photographs, and some items might well be illustrated by voice-overs or video clips. Kenosee Lake, Fig. 5, is one of several that were cored during paleobiology studies in the Palliser region. Each of the coring sites shown on the map is a clickable item. Selecting one opens a screen describing the general features [*p.13/p.14*] of the core at that location. Then, by selecting one of the options listed on the right, the reader can obtain additional information on that particular core. For example, clicking the photos option opens the screen seen in Fig. 6. It shows both normal and x-ray photographs of part of the core and a photo of the coring operation. The Palliser Triangle and Mackenzie Delta borehole CDs are large and complex presentations involving many linked pages and considerable scripting. Multimedia techniques can also be used very effectively and with little design effort to introduce a series of publications or a database. For example, Fig. 7 could be the opening screen for the Geological Survey of Canada paleontological database that is now in preparation. The two windows at the bottom allow one to query the database by entering a stratigraphic unit name or a taxon name. The database can also be opened, and the search narrowed, via the icons presented along the left side of the screen. The options shown on the lower right side of the screen offer extra features without significantly increasing the multimedia design overhead of the project. Reports would lead to a series of titles that could include all previously published CSC publications in paleontology. The papers would be in the Adobe Acrobat pdf format, meaning that they would contain electric bookmarks and links to other documents. Sponsors would be a link to a section on the activities of the Geological Survey and any other partners that contributed to the project, and Data would take the reader to a file containing images of all GSC type specimens. In other words, use of a multimedia interface opens the way for a much richer database than is ordinarily presented on CD. The same is true for another presentation now in the planning stages: the Digital Pollen Reference Collection. Almost all desk-top and many lap-top computers sold today have internal CD players. A CD pollen reference collection, running on a computer placed next to a counting scope, would be a powerful instructional and research tool. No longer would students and staff at small institutions be crippled by lack of an adequate reference collection. In fact, because a CD reference collection would drawn from a number of existing collections, it would be more complete than any single collection even those at large institutions. A digital pollen reference collection is not a new idea. Several have been constructed including one designed for the world wide web. However, what we are proposing [*p.14/p.15*] here would be far more than a simple collection of pollen images. A mock-up of one of the screens in the proposed Digital Pollen Reference Collection is shown in Fig. 8. Pollen images are presented in the central window, with additional comments in the smaller text box at the bottom. The user obtains an image of a pollen type in one of three ways: typing the taxon name in the top window, working through the key on the left, or selecting the taxon name from the list presented on the right side. One group of the options will simulate a microscope: use of "up" and "down" arrows on the computer key board will allow the user to focus up and down on the image in the center window and other keys will change magnification. The buttons shown along the row at the bottom on the mock-up screen add features that make the Digital Pollen Reference Collection much more than a series of pollen photos. For example, the button labeled Regional allows the user to select the region from which the core they are counting comes. Once this is done, a series of small windows show the 20 most common pollen types that have been found in other pollen studies in that region. In addition, this section will contain images of fossil pollen from selected sites in the region in order to illustrate regional preservation characteristics. A toggle button will allow the user to switch to a count sheet designed for that region and to add to a count database, using the keyboard function keys or by clicking on pollen images shown on screen. These count data are then available for numerical manipulation using the various statistical methods presented in the Data Room. In the Tutorial module, students will be able to test the various statistical tools in order to see how their output changes as a function of differences in the number of pollen types and pollen sums. This would also be the place where different processing techniques are presented, possibly accompanied by video clips. Another option would be to present a video of several traverses of a pollen preparation slide, with pop-up tags indicating which objects are pollen, which are fungi, trash, etc. Reports leads to a screen with a number of selectable titles. They might include papers of historical interest or even entire monographs, e.g., Erdtman's entire book on Pollen Morphology and Taxonomy of Angiosperms (with publisher's permission of course). If the reader clicks the Reports button while in one of the regional sections, the titles listed are studies conducted within that particular region. To be useful for students, the Digital Pollen Reference Collection should be priced no higher than a typical text book. This means that sponsors will be required to help defray production costs. But even more critical than financial partners are the many collaborators required in order to amass a suitable collection of pollen images. We are at present looking for both types of partners but especially the latter. The format used for the Digital Pollen Reference Collection might easily be adapted for presentations dealing with other types of fossils, such as plant macrofossils, insects, diatoms, forams etc. In fact, the first test of the presentation model will likely deal with marine molluscs. Because the Digital Pollen Reference Collection is still in the planning stages, there is plenty of time to alter the method of presentation or to rethink the content and options. We welcome comments and suggestions on this project or any of the others discussed above. A CD with [*p.15/p.16*] demos of the Borehole and Palliser projects is available for viewing upon request. ---------------------------------------------------------------------------- VOICES FROM THE PAST Pollen Analysis Circular v. 1-8 (1943-1944) and Pollen and Spore Circular v. 9-18 (1945 - 1954) Written and coordinated by Paul Bigelow Sears (Reprinted from master copies derived from originals held by: University of Minnesota, Limnological Research Center, 310 Pillsbury Drive, Minneapolis, MN 55455 USA.) (Coordinated November 1996 by: Linda C. K. Shane (shane002@maroon.tc.umn.edu) and Rebecca E. Teed (teed0003@gold.tc.umn,edu)). The coordinator suspects a good number of the readers have heard about the above project of Linda C. K. Shane and Rebecca E. Teed to reprint the whole run of the Pollen Analysis Circular (later the Pollen and Spore Circular) from the 1940s and 50s. I have been looking through my copy and found it to be a treasure. Although I am probably biased by my Upper Midwest U.S. education and my age, it was hard not to be impressed by the names of the contributors, 35 of whom I knew (know) and/or revered: Ernest Antevs, Elso S. Barghoorn, William S. Benninghoff, Kirk Bryan, M. L. Buell, Stanley A. Cain, Kathryn Clisby, W. S. Cooper, Lucy Cranwell, A. Orville Dahl, Pierre Dansereau, Edward S. Deevey, Jr., O. C. Durham, Knut Faegri, David G. Frey, Harry Godwin, Jane Gray, Henry P. Hansen, Calvin Heusser, G. E. Hutchinson, H. A. Hyde, Johs. Iversen, Donald B. Lawrence, Estella Leopold, Daniel Livingstone, G. F. Mitchell, Ruth Patrick, Winifred Pennington, J. E. Potzger, James M. Schopf, Paul B. Sears, A. F. Traverse, W. H. Twenhofel, L. R. Wilson, R. P. Wodehouse On the basis of an earlier request for interest, Shane and Teed decided to have sets made on acid-free archival paper and bound with a continuous coil binding. There are 180 pages (single side). The original cost per book was $20.00 US plus $6.50 postage and handling within the USA (total $26.50). Shipping out of the country costs more: air packet $21.00 (7-14 days) or surface printed matter $6.50 (6 to 8 weeks) (totals $41.00 US or $26.50 US). This is a volunteer-based project and they can not accept credit cards or the extra expense of handling POs or international funds. If you find yourself interested in a copy of the Pollen and Spores Circular, send an e-mail note to Shane see whether she has any more copies. Her address is: Dr. Linda C. K. Shane, University of Minnesota, Limnological Research Center, 310 Pillsbury Drive SE, Minneapolis, MN 55455, Tel: 612-624-8526 (new number), Fax: 612-625-3819, email: shane004@maroon.tc.umn.edu. ---------------------------------------------------------------------------- TST2NORM.EXE DETERMINES THE OVERLAP IN TWO NORMAL POPULATIONS By Louis J. Maher The concept of a normal population seems ubiquitous in science and life in general. Our data are often validated with means and standard deviations based on the assumption that the normal curve applies or that the data can be transformed to approximate a normal distribution. I was studying some sites in Wisconsin that had 14C dates associated with them, and I was trying to determine if the dates were sufficiently alike that I could accept the null hypothesis that they were of the same age and that the reported difference could have arisen by chance. I was also going to calibrate the dates with CALIB3c (Reimer, 1994), and that program has a null test based on the chi- square criterion. But more than a simple null test, I wanted to estimate the extent to which the populations overlapped. One site at Valders quarry had the date 12,965 ñ 200 yr B.P. The event-of-interest at Devils Lake was bracketed by the three dates: 12,260 ñ 115 yr B.P., 12,520 ñ 160 yr B.P., and 12,880 ñ 125 yr B.P. The mean of the three Devils Lake dates is 12550 ñ 233 yr B.P. (mean of the three dates ñ the square root of the sum of the squares of the dates' standard errors). Assuming that the two sites' dates are normally distributed with means 12965 ñ 200 yr B.P. and 12550 ñ 233 yr B.P., how much do the two distributions overlap? I thought about this a good deal, and I even considered getting the proportions by plotting the distributions to scale on paper, cutting out the parts, and weighing them. But then I recalled the idea of circular normal probability and the normal joint-probability density function that I had once used in calculating confidence limits for microfossil concentration measurements using samples spiked with marker grains; the details can be found in Maher (1981). [*p.16/p.17*] The basic reasoning and proof involves coding and standardizing the two populations so that they can be plotted in units of their respective standard deviations that is, in Z units. This can be done for any value of a variable by subtracting the population's mean and dividing the result by the population's standard deviation. (Of course, any Z unit can be transformed back into the original units by multiplying it by the population's standard deviation and adding its mean.) Fig. 1 shows the situation for the two 14C dates. The one with the larger mean has been standardized and plotted at the left side; the date with the smaller mean is plotted at the top. The standard normal joint-probability density function defines a bell-shaped surface lying over the (x, y) plane, and centered at the middle of the graph. The joint- probability density (JPD) decreases away from the center; 0.9998 of the JPD occurs within 4.1 Z units of the center, and that limit is shown by the large dashed circle. We will ignore the trivial probability lying outside that circle. Now we can set up a series of confidence intervals about the larger mean 0 ñ 0.2Z, 0 ñ 0.4Z, etc.; we will refer to these intervals as C2. Likewise a similar set of confidence intervals about the smaller mean will be referred to as C1. An example is shown in Fig. 2. As the value of the intervals' Z gets larger, the confidence limits of the two distributions increase and will finally merge. What will be the value of Z when the merger occurs? If we increased Z in small increments, at each step we could calculate the minimum difference between the confidence intervals by subtracting the upper limit of C1 from the lower limit of C2. That difference will start as a positive number, but with each step it will decrease until it becomes negative. When the difference is zero the intervals just touch; when the difference is negative the intervals overlap. Referring back to Fig. 1, if we decoded the standardized data and plotted all the possible values of C2 - C1 on the graph, they would form a series of isolines sloping from lower left to upper right; their numeric values would decrease along a gradient from upper left to lower right. The only isoline plotted on Fig. 1 is the one where C2 - C1 = 0; it separates the positive and negative (shaded) regions on the graph. The gradient of C2 - C1 extends at right angles to the isolines. Its orientation angle can be predicted from the ratio of C2's standard deviation to C1's standard deviation: Theta = arctan [200/233] = 40.6ø , measured from a line parallel to the C1 scale and passing through the figure's origin. If the standard deviations were the same, Theta would equal 45ø, and, of course, if C1's standard deviation was less than C2's, Theta would be greater than 45ø. Given this background, it can be shown that the small circle at the center of Fig. 1 has a radius of Z = 1.351, and it is tangent to the isoline C2 - C1 = 0. That is the instant when the two confidence intervals shown in Fig. 2 start to merge, and for which the JPD can be calculated. In the general case, if we know the JPD where the confidence limits of the two populations have no common values, then we can subtract this from unity to find the JPD that the two populations share. This is the kind of problem that can be solved quickly by computer. I have written a program called TST2 NORM.EXE to find how much two normal populations overlap. The procedure requires the populations to be ordered based on the size of their means; the computer handles this once the means and standard deviations are entered. [*p.17/p.18*] Given the means and standard deviations of two normal distributions, the program raises Z by increments of 0.001 over the range from 0 to 4.0. For each value of Z, and for each distribution, it calculates confidence limits ñ Z units about each of the two means. The confidence limits steadily widen as Z increases. For each value of Z the program subtracts the upper limit of the distribution with the lesser mean from the lower limit of the distribution with the greater mean, and it assigns the difference to the variable "Test." In the general case, "Test" starts as a positive variable that gets smaller as Z increases. The program searches for the value of Z for which "Test" first becomes negative; that is to say, the confidence limits of the two means begin to overlap. The last positive value of "Test" corresponds to a value of Z where the confidence ranges of the two means do not overlap, and for which the two distributions' joint-probability density can be calculated. The program then calculates the probability under the normal curve ñ Z units from the mean, and it subtracts that probability from 1.0 (unity). The difference represents the joint-probability density that the two distributions share. For the problem with the two 14C samples, TST2NORM's final screen shows the following data: 'Test' 1.071198 Z = 1.348 .7641187 Z = 1.349 .4570392 Z = 1.350 .1499597 Z = 1.351 -.1571197 Z = 1.352 There is no overlap of the range of the two distributions through Z = 1.351, at which time the unshared Joint Probability Density reaches 0.8233. At the interpolated value of Z where 'Test'= 0, 0.1765 of the Joint Probability Density is shared between the two distributions. FIRST SECOND Mean: 12550 12965 SD: 233 200 Do you wish to run another pair? (Y/N) Almost 18 percent of the JPD of the two populations is shared. Although we can use this as a null test (à > 0.05), in this particular case I am more interested in the actual amount of the overlap of the two populations. I have put a self-unzipping copy of TST2NORM.EXE in the INQUA File Boutique as TST2NOMZ.EXE in case it might be useful to you. Remember that the program assumes the underlying populations really are distributed normally, and that the means and standard deviations are accurate. That is a lot to assume, but one routinely makes such assumptions with tests. References. Maher, L. J., Jr., 1981, Statistics for microfossil concentration measurements employing samples spiked with marker grains, Review of Palaeobotany and Palynology 32, 153-191. Reimer, Paula, 1994, Radiocarbon calibration news. INQUA-Commission for the Study of the Holocene, Working Group on Data-Handling Methods 11, 21-23. ---------------------------------------------------------------------------- CNT-WIS.XLM: AN EXCEL MACRO FOR CONCATENATING POLCOUNT FILES FOR IMPORTATION INTO TILIA Richard Telford Institute of Earth Studies University of Wales Aberystwyth, UK E-mail: rjt95@aber.ac.uk I have recently been using POLCOUNT (Maher, 1992) for recording diatoms in cores from Ethiopia. I find the program very useful. However, combining the count files into Wisconsin format, prior to importing them into Tilia, is both time consuming and tedious as I have over 50 samples per core. With this in mind, I have written an Excel macro that performs this task. CNT-WIS.XLM is an Excel 4 macro, but it works on Microsoft Excel 7 for Windows 95. Before being used on an Apple Macintosh it might need a little editing. POLCOUNT produces a series of output files, each beginning with a title, the number "102" and then data for [*p.18/p.19*] 102 taxa, like the following sample: Mud Lake = 01-12-97 102 1 250 1.5 0 7 12 1 1 16 5 0 33.5 13 172 5 3 5 10 5 1 1 1 1 2 0 0 2 0 0 0 0 0 0 0 0 24 11 5 153 4 13 0 1 1 4 2 0 1 2 3 0 0 0 0 0 7 5 2 21 38 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 595 POLCOUNT also uses a file that contains the names of a site's taxa. It starts with a title, the number "101," and a taxon list like the following sample scrap that might be called MUDTAXA.TAX: Mud Lake Taxa 101 Marker Grain 00 Taxon 01 Taxon 02 ... Taxon 98 Taxon 99 Sum(Taxa 01-99) 100 Save the POLCOUNT files without any extension; you might name your Mud Lake files ML0001, ML0010, ..., ML1500 to identify both site and sample number in the file name. Put both CNT-WIS.XLM and the *.TAX file in a directory, and put copies of the data files in a sub-directory by themselves. To combine the count files, start Excel, select "Files," and open the macro CNT-WIS.XLM. The macro will load and announce that it concatenates POLCOUNT files into the Wisconsin format, that it assumes the count files will have no extension (and that all other files will), and that to start, you should type the three-key combination Ctrl-Shift-W. In sequence you will be asked for 1) the path to the directory with the count files, 2) the name of the taxon file, 3) the name of the site, and 4) the name of the Wisconsin- format file (with extension ".RAW") that you wish to produce. Finally, the macro will remind you to run REMZEROS on the *.RAW file to remove the taxa that never received any counts, before importing the Wisconsin file into Tilia. Owing to the number of files the macro has to open and close, it is rather slow, taking about 3 seconds per count on my 386. It is important to remember that only "extensionless" files that you wish to include in the Wisconsin file should be in the directory. The *.TAX file should be in the form listed above; that is: Title, 101, Markers.... The macro takes care of cutting off the first two lines and inserting the sample depth. The macro may not place all the POLCOUNT files in correct order in the *.RAW file, but that can be quickly taken care of in Tilia. After running REMZEROS.EXE to remove unused taxa, import the *.RAW file into Tilia (Load data file, Wisconsin format). (In 1.x versions of Tilia, you may have to delete the first line (Ctrl-Y) which is the sample depths.) Then press Alt-E to edit the samples | F3 to sort | N to sort on sample number. Then save as a Tilia file. CNT-WIS.XLM and some sample data can be found in the self-extracting zipped file called CNT-WISZ.EXE in the INQUA File Boutique. Reference. Maher, L. J., 1992, Turn your expensive old pc into a dumb pollen counter, INQUA-Commission for the Study of the Holocene, Working Group on Data-Handling Methods Newsletter 8:24-27. ---------------------------------------------------------------------------- HOW DO I GET ON | OFF THE LIST? From time to time almost everyone wants to get on a List or to get off a list that no longer holds our interest. Alwynne B. Beaudoin recently reminded those on the Quaternary list that she has this useful information at her Canadian Association of Palynologists (CAP) WWW site at http://www.ualberta.ca/~abeaudoi/cap/cap.html and follow the link to "Internet Resources." You will find two files - Discussion Lists and More Discussion Lists - which list the subscription instructions for this list plus many other geological and Quaternary-related ones. ---------------------------------------------------------------------------- [*p.19/p.20*] NEW NEWSLETTER COORDINATOR Keith Bennett Department of Plant Sciences University of Cambridge Downing Street Cambridge CB2 3EA, UK Phone: (0)1223 333948 (direct) (0)1223 333900 (messages) (0)1223 462416 (home) Fax (0)1223 333953 E-mail kdb2@cam.ac.uk WWW http://www-palecol.plantsci.cam.ac.uk/ Anon ftp ftp-palecol.plantsci.cam.ac.uk New WWW Address of the INQUA File Boutique http://www-palecol.plantsci.cam.ac.uk/inqua ---------------------------------------------------------------------------- Mirror INQUA File Boutique in North America ftp ftp.geology.wisc.edu Logon: anonymous; Password: your e-mail address Path: /pub/inqua World Wide Web http://www.geology.wisc.edu/~maher/inqua.html ----------------------------------------------------------------------------