Download PC-ORD Version 4 Fixes

These are Windows PC-ORD version 4 and associated fixes.

You will get all the fixes when you download.

Note these fixes will only work if you have purchased Windows PC-ORD Version 4.

These are the last fixes for Version 4.  All future fixes will apply to Version 5.

* ver. 4.01 Corrected problem with import of relaxed Cornell (Decorana/Twinspan) format. Fix for hiding cluster dendrogram legend when unchecked in preferences. "Minimum" misspelling fixed in Scree Plot legend. Fixed Twinspan failure when no title was provided. Fixed Graph Ordination problem: rotated axis scores were not used when viewed in plane other than the plane of rotation.

* ver. 4.02 Fix for international users plotting overlays with regional settings using comma for decimal separator. Corrected output error when NMS Autopilot finds 4-D solution. NMS Autopilot changed to no longer increase the stringency of the stability criterion for the final run. This had resulted in the final run giving results that were slightly different than the best preceding trial run. Fix errors occurring when menu items are quickly selected before matrix files are fully indexed in background.

* ver. 4.03 Fixed NMS logic error in previous fix (4.02) -- prevented graph of stress vs. iteration. Improved handling of multicollinearity by multiple regression in CCA. Improved error handling for CV values from Row and Column Summary.

* ver. 4.04 Corrected divisor for frequency and abundance in species list broken down by groups; improved handling of non-integer grouping variables in species lists.

* ver. 4.05 In Indicator Species Analysis improved table of monte carlo results by adding group identifier for group with maximum observed IV. Compact-format summary with breakpoints changed. If "separate summary for each break group" is checked, summary tables reflect only those species included in that break group.

* ver. 4.06 Two changes in Indicator Species Analysis: fixed option to exclude groups which in some cases misinterpreted request; MaxGrp in final table changed from group sequence number to group identifier.

* ver. 4.07 NMS, fixed memory allocation problem that causing varimax rotation crash on NT systems. Logic changed for choosing which run to rotate. If the number of real runs is greater than one, the rotation is deferred til the final "best" solution. That rotated solution is written into the result file and the graph file. As before, if the number of real runs is one, the rotation is applied to the maximum number of dimensions requested. Fixed species-area curve failure to print standard deviations with small data sets. Transparent progress window on NT systems fixed. Matrix multiplication (AE) wrote incorrect empty columns for some matrices because of memory allocation error.

* ver. 4.08 Fixed PC-ORD not remembering your last choices (opened matrices, analysis options, etc.) from introduced logic error in previous fix (4.07)

* ver. 4.09 Fixed Joint Plot and Overlay Second Matrix graphs when rotating via variable from the main matrix.

* ver. 4.10 Fixed problem with MRPP hanging with huge data sets (several thousand items).

* ver. 4.11 Fixed problem with import of compact-format individual tree data when converting to counts. Species were ordered incorrectly.

* ver. 4.12 Fixed premature termination of indicator species analysis with moderately large data sets. The problem arose in version 4.06.

* ver. 4.13 Fixed opening the project file Default.prj causing PC-ORD reporting version error message and refusing to do analysis. A corrected Default.prj file is included in the download.

* ver. 4.14 Corrected TWINSPAN printing only partial table in some data sets with more than 5 times as many sample units as species. Fixed NMS elapsed time clock so it does not reset at midnight.

* ver. 4.15 Added clarifying message and statistics to Graph Ordination | Statistics | Percent of variance explained. The statistics now also show the degree to which ordination axes are not completely independent of each other, as often happens with NMS. Added text file to NMS output for producing Shepard diagram. Activated Append button of main window after doing a statistic in the Graph window.

* ver. 4.16 Fixed Twinspan quitting prematurely with more than 7 levels of division.

* ver. 4.17 Added the ability for users to add and run their own modules. See Add-Ins. Improved speed of importing *.csv files.

* ver. 4.18 Fix for out-of-memory error message on computer systems with extra large memory capacity. Improved matrix display for large numbers from 8 significant digits to 16 significant digits. Note, the analysis has always been high precision; only the display in PC-ORD was enhanced.

* ver. 4.19 For 1-D ordinations only, removed spurious information for "Axis 2" from Graph Ordination | Percent of variance in distance matrix.

* ver. 4.20 Revised estimate of Twinspan memory use in File | Memory Requirements (previously overestimated). Fixed problem with NMS Scores (predictive mode NMS) producing varying but equally valid results for identical objects, when many ties existed in the dissimilarity matrix. Module altered so that the same ranking results each time. Adjusted precision of display of matrices to 7 digits. Added capability of turning on an individual graph point label temporally by left clicking or permanently by right clicking. You also can right click on any empty space in the graph and a point labeling dialog will pop up.

* ver. 4.21 Changed axis labels for low-resolution NMS plot of distance vs. dissimilarity. Added capability of selecting ordination point labels from second matrix, group numbers, or the graph file. See FAQs Categories with text labels and Groups with greater than 15 categories.

* ver. 4.22 Fixed problem with NMS Scores not completing with Relative Sorensen distance. Changed output detail in NMS. Listing of the magnitude of the gradient vector changed from sum of squares (SS) of elements to the Sqrt(SS/N) where N is the number of objects. This appears only when you select "List stress, etc. for each iteration." Changed graph preferences options file format to fix problem with graph fonts and file size.

* ver. 4.23 Changed import of list (database) format to remove error reading negative numbers.

* ver. 4.24 Fixed problem with matrix editing of large numbers. Fixed matrix cell truncation to fit length of data displayed to cell size; does not affect analysis high precision.

* ver. 4.25 Species lists: trap error with same species name appearing more than once in main matrix.

* ver. 4.26 Results file text search feature added. Help file updated. Appending results file now scrolls down to appended material. Arrows added to the end of vector overlays on ordinations and vectors color coded by group. Expanded format and improved error trapping in row and column summary. CCA in PC-ORD no longer reports p-values for axes 2 and 3 of Monte Carlo tests. Results for these axes are not independent of axis 1. Pierre Legendre has demonstrated that p-values for eigenvalues of axes after axis 1 may be biased, when based on a simple randomization test.

* ver. 4.27 Fix for graphing errors with large matrices of 10,000 or more rows or columns.

* ver. 4.28 Fixed problem with CCA not completing Monte Carlo tests with user-selected random number seeds. Improved recognition of delimiters when importing data files. Fix 'All files (*.*)' file dialog 'Files of type:' filter. Fixed matrix editing of large integers greater than 32767 turning into negative numbers when saved. Maximum length of file names increased to 246 characters. Input of spreadsheets made more forgiving for cells not used by PC-ORD.

* ver. 4.29 Improved tolerance to unstable solutions in DCA output. Fixed problem with importing two matrices with a matrix swap in between.

* ver. 4.30 Summary statistics from compact data format, fixed problem not completing large data sets. Changed MRBP to prevent users from applying to data sets with more than one instance of each treatment in one or more blocks. Improvements with the progress thermometer. Improved output of distance matrix in NMS.

* ver. 4.31 Fix for problem in editing labels. Fix for problem in saving graph statistics.

* ver. 4.32 Fix for problem in calculating File | Memory Requirements for large data sets. Does not affect whether or not an analysis will actually run.

* ver. 4.33 Changes to the help file.

* ver. 4.34 Improved formatting of row and column summaries. Default printer error fixed. If your matrices windows are closed down too small to see the data, delete the pcord.ini file in the pcord4 folder. Consistent international decimal character fixed in dialogs.

* ver. 4.35 Fixed Window | Arrange.

* ver. 4.36 Fixed for results file remembering font preference and some other minor issues.

* ver. 4.37 Fixed Twinspan failing to write final table when lowest nonzero value smaller than lowest cut level. Fixed graph legend font name preference.

* ver 4.38 Improved progress messages and cancellation procedure in NMS Scores. Fixed "output conversion error" with cluster analysis of huge data sets.

* ver 4.39 Fixed graphing CCA Joint plot with 'M'ixed variables in second matrix.

* ver 4.40 Fixed problem with NMS not calculating species scores by weighted averaging with a particular combination of options.

* ver.4.41 Fixed axes labeling precision to two digits when axis range is between 0 and 0.1