General
Empty variables or zero variance error
Q. I received an error message "Error: Empty variables; attempt to divide by
zero." What's wrong?
A. If your spreadsheet has rows or columns that are all zeros, many analyses will fail
because in the underlying math, something is divided by sums or sums of squares that equal
zero. In general you should check for empty rows and columns before running
analyses. The easiest way to spot these is to run the Row & Column Summary under
"Summary". A related problem is when a column has a variance of zero (all
values are the same in that column). Any analysis requiring division by the variance
or standard deviation of that column will, therefore, produce an error message.
Second matrix and graph file do not match
Q. We get the following error message. "The number of items in your second
matrix does not match the number of items in your graph file. Second matrix will be
ignored. This error came after doing the ordination and while trying to graph the
ordination.
A. The most common cause of this problem is when one or more sample units (rows) are
deleted from the main matrix. If you do this, you also need to delete the
corresponding rows from the second matrix. The easiest way to avoid this problem is
to use Modify Data  Delete Rows  Both Matrices.
Negative CV's
Q. In Row and Column Summary I obtained a negative coefficient of variation (CV).
How can that be?
A. One can get wild results like CV = 833% if the data include negative numbers, or
were standardized in some way that produces negative numbers (such as relativization by
standard deviates). But the CV is always nonnegative if the data are all nonnegative.
The CV makes sense only with nonnegative data.
Resampling tests
Q. Are resampling or randomization tests (e.g., jackknife, bootstrap) provided for
ordination procedures other than CCA?
A. Resampling or randomization procedures used in PCORD for several other applications
(NMS, Indicator species analysis, Mantel test, PCA, and perMANOVA). Jackknife
estimates of species diversity are included. But at present we have not included
randomization tests for ordinations other than CCA, PCA, and NMS.
Spreadsheet Columns
Q. How many columns of data can I use? My spreadsheet program allows a maximum
255 columns. I want to enter a data set with more species than that.
A. In PCORD you can use up to 32,000 columns if your machine can handle it. But
most spreadsheet software is limited to many fewer columns than that (255 columns in old
versions of Excel). You can circumvent that by using compact format or database
format. PCORD imports these into a spreadsheet. Or, depending on the shape of your
data matrix, it may be advantageous to enter the transpose of your matrix in a spreadsheet
program. For example, if you had 100 plots and 300 species, you could enter the
species as rows and the plots as columns, then transpose the matrix in PCORD. If
you have a very large data set we recommend either compact format or database format.
To use the database format, first export your data from your database to a delimited
text file. PCORD expects three columns, in the order:
SampleUnitID SpeciesID Abundance
Then PCORD will build a matrix from this.
PCORD crashes on laptop
Q. When I launch PCORD on the laptop the machine behaves strangely and eventually
crashes. Is there a known problem?
A. Check to see if you have a network printer as the default printer. (Start 
Settings  Control  Panel  Printers) A check mark will appear next to the default
printer. When you are not hooked up to the network, make sure the default printer is
not a network printer.
Error reading PCORD options file
Q. I received an error message: Check for possible mixed or multiple version numbers of
PCORD. Error reading PCORD options file. What is the solution?
A. Use Tools  View Application Data Folder. Close PCORD, but leave the folder
window open. Delete the file Pcord7.opt. Restart PCORD. Then if you
haven't already, download the latest fixes (Help  Check for software updates).
File not found error
Q. When attempting to run NMS using starting configuration I provide in a
"graphrow" file PCORD is unable to open or read the file. The error
message I get is "Error reading graph file. Graph file not found." What is
wrong?
A. Any analysis related file has to be stored in a folder that has a filename
(including the path) less then 240 characters long. Open Windows Explorer and
navigate to the folder where the file is stored to get the complete path in the Explorer
Address window. If the path plus file name is longer than this limit move the files
not found to a shorter path and/or use a shorter file name.
Analyze second matrix
Q. How do I get row and columns summaries, outlier analyses, etc. on the second matrix?
A. Use "File  Switch Matrix" to swap the main and second matrices.
Think of the main matrix as the active matrix. This is the matrix on which the
operations are performed, unless an operation specifically requires the second matrix.
Editing spreadsheets
Q. Why doesn't PCORD allow cellbycell editing of spreadsheets? I have to go
back and forth between PCORD and a spreadsheet program when I am cleaning my data.
A. PCORD doesn't attempt to duplicate the spreadsheet functions. However PCORD
versions 4 and above allow simple editing of the main and second matrices. More
complicated editing and sorting is best done in your spreadsheet software. Be sure
to close a file in PCORD before editing it in spreadsheet software and viceversa.
Otherwise file sharing conflicts may result that will cause problems in either Excel or
PCORD.
Convert output to tables
Q. How can I convert tables in PCORD’s output files into tables in my word
processor?
A. Your word processor probably has functions for converting tabular text files into
tables. For example, using Microsoft Word, you can highlight a table in a PCORD
output file, then select Table  Convert Text To Table, to convert the output into a
standard tabular format. Be sure to set the delimiter to a space character and
adjust the desired number of columns.
Convert Output to Spreadsheets
Q. How can I convert tables in PCORD’s result files into spreadsheets?
A. It is easy and fast to convert tables in your results file to columnar data in
Excel. That works really well with PCORD output, since the result file
tables are fixed width, matching one of the options of Excel.
Highlight the table that you want to export, then type Ctrlc to copy. Put the
cursor on a single cell in Excel where you want the upper left corner of the table, then
type Ctrlv to paste. The results will all go in a single column. Then, before doing
anything else, while the column is still highlighted, use Data Text to Columns  Fixed
width, to create the separate columns.
Note that PCORD also provides direct conversion to spreadsheet in a number of cases, such
as saving an ordination graph file (containing the ordination scores) as a spreadsheet.
Groups on ordinations
Q. I have a question about grouping. I have measurements of habitat along three
transects of 250 plots, one transect along each of three creeks (total of 750 plots).
How can I make PCORD group these plots so I can best run a ordination on them?
A. Ordination doesn't necessarily have a builtin concept of groups, but you can
overlay groups on any ordination. To do this, create a second matrix with 750 rows
and 1 column (or you can add this column to an existing second matrix). Use this
column to create a categorical ("C") variable that codes for "creek"
(1, 2, 3). Open this as the second matrix and when graphing your ordination, select
"creek" as an overlay variable. You can use different colors and/or
symbols to indicate each group. If you want to compare species and communities among
groups, consider MRPP and Indicator Species Analysis.
Multifactor Experimental Designs
Q. How can I using purely nonparametric perMANOVA or MRPP to analyze experimental
designs with more than one or two factors in the design?
A. PCORD uses at most 2 factors in perMANOVA, 1 factor in MRPP, and 1 factor plus
blocking in blocked MRPP. For more complicated designs, consider the following
possiblities.
 Aggregate subsamples by averaging. This will decrease noise in the estimates.
This applies to nested designs.
 Combine multiple levels of blocks or nesting into a single level representing control of
spatial effects. For example, with blocks nested within sites, use block/site
combinations as a new compound blocking variable.
 Identify weak factors by oneway analyses and drop the weakest factors from the design,
in effect increasing replication.
 Identify the strongest factor and split the analysis into a separate analysis within
each factor. This is particularly effective if the factor interacts with another
(e.g. Factor 2 makes a difference only at one level of Factor 1).
 With repeated sampling of the same design, calculate the statistic representing
treatment effect at each date separately. Then plot effect size (e.g. A from MRPP or
F from perMANOVA) against date.
 For repeated sampling subtracting the first date from the second, within each sample
unit. This results in values that represent the change in response variables.
Accordingly, this matrix will have both positive and negative numbers. If you do
this, you need to use a Euclidean distance measure, or another distance measure that
tolerates negative numbers, because the negative numbers resulting from the subtraction
don't make sense with Sorensen distance or other proportional cityblock distance
measures.
 Other kinds of subtractionbased data adjustments are also possible—for example,
subtracting the mean (or median) for each variable in each block from each observations
that block. Thus each observation becomes a deviation from the block median or mean.
These adjusted observations could then be used in an analysis of a more complex
design. For example, if you had twoway factorial design with blocking, using the
deviations from withinblock means would allow a twoway perMANOVA. What you would
lose is a measure of the size of the block effect (since you have subtracted it out).
Mac version?
Q. Are versions of PCORD available for the Macintosh?
A. No. There is only a Windows versions of PCORD. However, PCORD will run
on Virtual PC for Mac. It will also run on the Windows side of a partition on a Mac.
Mac emulation 3D crash
Q. I am using PCORD with Mac OS X running Windows XP via Parallels. Graphing 3D
ordinations makes PCORD crash. Is there a solution?
A. Open XP Control Panels  Appearance and Themes  Display Properties  Settings 
Advanced  Troubleshoot and set the Hardware acceleration slider to the second position
from the left that says "Disable all but basic accelerations. Use this setting to
correct more severe problems". Note this position also disables DirectDraw and
Direct3D used by any other software.
PCORD CD will not install
Q. PCORD CD will not install on my system. What is the solution.
A. Be sure to disable any antivirus program such as Norton before trying to install.
If the CD still won't install, you may have the CD autorun feature disabled.
You can install manually from the Start Menu by selecting Run and entering x::Setup
where x is the letter of your CD drive. You also can click the Browse button and
navigate to the Setup program.
Exponential formats for numbers
Q. I’ve attached an output file from row and column summaries. Why are the numbers
on my species summaries so odd?
Summary of 53 species
N = 9 points

No. Name Mean
Stand.Dev. Sum
Minimum
Maximum (etc.)

1 AGRPAR 0.111E+00 0.333E+00
0.1000E+01 0.000E+00 0.100E+01
2 ALLPET 0.100E+01 0.112E+01
0.9000E+01 0.000E+00 0.300E+01
3 ASPPLA 0.111E+00 0.333E+00
0.1000E+01 0.000E+00 0.100E+01
A. The numbers in your output are in exponential format ("E fields"). PCORD
switches to this for row/column summaries when the number of decimal places in the regular
format has insufficient precision in some portion of the table. In this specific case,
you're on the borderline  the regular format probably would have been ok too.
The following item from the builtin help system explains how to interpret the exponential
format:
E Fields
Numbers are sometimes listed with the computer version of scientific notation. This is
used for maximum flexibility in portions of the output. The numbers following the E are
exponents. For example,
.7071E+01 = 7.071
.9169E+00 = .9169
.1234E05 = .000001234
Discriminant Analysis
Q. Does PCORD contain Discriminant Analysis? If not, why not? I could use
MRPP or perMANOVA but I am more interested in the graphs than the statistical tests.
A. PCORD does not contain discriminant analysis. We have emphasized techniques
that are more appropriate for community data, especially nonparametric techniques that are
not readily available elsewhere. So DA falls in the category of a parametric
technique that is available in virtually every major statistical package and is seldom
useful for community data. If you are mainly interested in graphical approaches, you
can use other ordination techniques, such as NMS, then superimpose group memberships on
the symbols in the ordination. If you are using community data, you should also
consider indicator species analysis, which will perform better than discriminant analysis
in telling you which species differ among groups. MRPP has no graphical component,
but when you have a priori groups MRPP makes a nice companion to a graphical approach such
as ordination.
Won't work with Vista or Win 7
Q. I just installed PCORD and nothing seems to work.
A. The PCORD folder might not have write access.
 Log in as Administrator
 Right click the PCORD icon or menu item that you run PCORD
 Select Properties
 Select the Security tab
 In the Permissions for SYSTEM box, check Allow for everything except Special Permissions
 Click Apply and OK
Can't find saved files in Vista or Win 7
Q. Whether saving a matrix, graph, or results file, when I try to find the file, it is
nowhere to be found.
A. Sounds like you are saving to the C:\Program Files\PCORD folder or a subfolder. (In
64bit Windows: C:\Program Files(x86)\PCORD) For security reasons Microsoft doesn't
think you should write any files to the Program Folder or subfolders so it writes to a
Virtual Store that you cannot see or get to.
Here are some solutions:
 Save your files to any folder other than a Program Files folder, for example, Documents.
 With Windows Explorer navigate to the Program Files\PCORD folder (with 64bit Windows to
Program Files (x86)\PCORD folder) and click Compatibility Files in the Windows Explorer
bar.
 Run PCORD as Administrator. Even if you are logged in as Administrator you still
have to right click the PCORD icon on the Desktop or on the Start Menu and select Run as
Administrator.
 One time, right click the PCORD icon and instead of selecting Run as Administrator,
select Properties  Compatibility  Privilege Level and check Run this program as an
administrator.
Out of memory
Q. I am running a large matrix and I get the "out of memory” error. What can
I do?
A. PCORD can use all of your available memory. If you have plenty of memory and you
get this error, your operating system has encountered some conflict in system resources
and you may need to reboot.
There are several steps you can take to maximize the amount of available memory (check
"Memory Requirements" under the File menu).
Avoid loading other memoryresident software, such as network drivers, if possible.
 Avoid loading other memoryresident software, such as network drivers, if possible.
 Reduce the size of your matrix by:
 analyzing subsets of the data,
 aggregating the observations to a higher level in your sampling design, or
 eliminating rare species (say those in less than 5% of your sample units).
 Add more memory (RAM) to your machine.
 See additional suggestions for using large data sets.
Editing and Searching Result File
Q. How can I edit, search in, or insert comments in my result files?
A. Simple editing and commenting of your result file is possible directly within
PCORD. For more elaborate editing, just open the result file with the notepad or
other text editor. You can do this without closing down PCORD.
Data Structure
Analyses and transformations of the second matrix
Q. How do I get row and columns summaries, outlier analyses, etc. on the second matrix?
A. Use File  Switch Matrix to swap the main and second matrices. Think of the
main matrix as the active matrix  the matrix on which the operations are performed,
unless an operation specifically requires the second matrix. Alternatively, you can
get a quick summary or profile of main and second matrices by selecting Advisor  Show
Current Profile.
Delete rare species
Q. Is there an easy way to delete rare species from a matrix?
A. Assume that species are columns. Select Modify  Delete Columns  Fewer Than N
Nonzero Values. Then you enter a value for N. If N=1 then empty columns are
deleted.
Delete rows from second matrix
Q. I deleted rows from my main matrix, but when I performed an analysis, I received an
error message that the second matrix didn't match the main matrix. How can I delete the
corresponding rows from the second matrix?
A. To delete rows from both matrices, use Modify  Delete Rows  Both Matrices.
Missing data
Q. What can I do about missing data (cells in my matrices for which I have no data)?
A. You must take steps to replace missing cells with numerical values, or you must
delete the offending rows or columns. If you ignore warnings about missing data,
they are most likely to be interpreted as zeros, which may or may not be reasonable.
There are several approaches to substituting values for missing cells. We
recommend reading Tabachnik and Fidell (1989) for a good discussion of the pros and cons
of various remedies for missing data. PCORD expects you to deal with the missing value
problem rather than having an automatic way of dealing with it. The reason for this
is that for most multivariate analyses there simply is no good universal way of handling
missing values.
Error message about a cell outside data matrix
Q. When PCORD reads my data I get an error message about a cell that is outside of my
data matrix. Why?
A. There is something in your spreadsheet outside the matrix. Even if the cell
looks blank, there may be something there. For example, if you go to a cell in a
spreadsheet, hit the space bar, then move to another cell, the cell will look blank, but
there are actually data written in that cell: first a code saying there is a label field,
next the "blank" character. When you put the cursor on the cell in
question the spreadsheet will display one of the characters that signals a label field:
usually ' " or ^. This is displayed not in the cell itself, but in the separate
display that shows the contents and any imbedded codes in the current cell. You can
get rid of these by erasing the cell in question or a whole block of cells.
Error in species file, number not first in line
Q. I entered my data in the compact format. When I run "Summary" or
attempt to import the data I get an error on ever line: "Error in species file, line
1, Species number not first in line." My species file looks correct. What
is wrong?
A. In this particular case, one possibility is that each line begins with a TAB
character. PCORD is looking for a number but finding something that is not a
number. Check for and remove TAB characters (or other nonnumeric character) at the
beginning of each line.
Subplots in each plot
Q. I have 40 plots and 60 subplots in each plot. Each of the plots has been
recorded at 3 dates. How can I enter and analyze this large data set in PCORD?
A. Your data set has a nested structure which is best handled in PCORD with the
compact data format using "break groups." If each combination of
date/plot/subplot is analyzed separately, then you will have way too many rows for
analysis (40 x 60 x 3 = 7200 sample units). We suggest combining some of the sample
units  essentially collapsing the nested structure somewhat. How you do this
depends on the question of interest. For example, if you are interested in the
differences between dates, you might aggregate the 60 subplots within each combination of
date/plot, leaving you with a manageable 80 sample units. This would be done by
creating a compact format file. If you already have the data in spreadsheets, you
can export the data to compact format, then use a text editor to append the break groups
that you choose, then import the data into PCORD, selecting aggregation of subsamples.
The groups of plots to be combined are separated by the keyword "break".
However you set up your main matrix, you should set up a corresponding second
matrix that codes for each level in your design.
Using numbers as row labels
Q. I wish to label my sample units with numbers, but PCORD insists on character data
in the first column. It seems that the only way to get number labels into PCORD is to
either enter them in compact format and import, or reformat each cell by hand in Quattro
or Lotus, or insert a character before the first number in Excel.
A. This is a problem with the Excel export to *.wk1. It is too smart for its own good,
converting numbers defined as characters into numbers defined as numbers. The
problem is irrelevant for PCORD 7, because it no longer uses the *.wk1 format by default.
One solution to this and many other data formatting problems is to export the file from
Excel in commaseparatedvalues format (*.csv). This puretext format strips out all
formatting codes. Then import the *.csv file into PCORD.
Compact format vs. spreadsheet
Q. I have a large data set in a spreadsheet program. The help file states that
the compact data format should be used to optimize storage of this type of data. Do
I need to convert my spreadsheet to compact format?
A. Compact format is mainly a time saver with data entry. You CAN use it to store
data in a more compact way, but unless you are low on disk space, it is usually more
convenient to keep the data in spreadsheet format. There are, however, a number of
situations in which using compact format can be very helpful. If you are managing
large data sets with many subsets, you will find compact format convenient for combining
subsets in various ways. You can easily combine different subsets of a large data
set, just by appending the compactformat text files to each other. But if you (a)
have plenty of disk space, and (b) don't need to fiddle a lot with subsets of data, and
(c) don't need to compact format for data entry, then you should stick with the
spreadsheet format.
Export compact format with breakpoints
Q. I have a large spreadsheet data set that I wish to export to compact format.
How can I insert breakpoints in the compact format file?
A. When you export to compact format, breakpoints are not inserted. If you wish
to insert those, do so with a text editor after you have created the compact file.
Append matrix
Q. How can I append one species matrix to another in PCORD?
A. Combine two spreadsheets by selecting Modify  Append Matrix.
How many columns of data can I use
Q. How many columns of data can I use? My spreadsheet program allows a maximum
255 columns. I want to enter a data set with more species than that.
A. In PCORD you can use up to 32,000 columns if your machine can handle it. But most
spreadsheet software is limited to many fewer columns than that (255 columns in old
versions of Excel). You can circumvent that by using compact format or database
format. PCORD imports these into a spreadsheet. Or, depending on the shape of your
data matrix, it may be advantageous to enter the transpose of your matrix in a spreadsheet
program. For example, if you had 100 plots and 300 species, you could enter the
species as rows and the plots as columns, then transpose the matrix in PCORD. If
you have a very large data set we recommend either compact format or database format.
To use the database format, first export your data from your database to a delimited text
file. PCORD expects three columns, in the order:
SampleUnitID SpeciesID Abundance
Then PCORD will build a matrix from this. Both compact format and database (list)
format are described in detail in the online help system.
Bad entry in cell x. Expecting number
Q. I get the error: "Bad entry in cell x. Expecting number." I looked
at the spreadsheet and can find nothing wrong. What is the problem?
A. PCORD is expecting a number but finding something else. Sometimes the problem
can be "invisible." Examples:
 the letter o instead of the number zero
 an imbedded space in a number (the spreadsheet automatically considers this character
data)
 an equation instead of a number
If you can't see the problem, reenter the data in the cell and explicitly format the cell
to be a number. This problem can also produce the error message: "Error reading
main matrix. row x. Label encountered in spreadsheet when expecting number."
Another way to quickly remove all cell formatting that PCORD does not recognize is to
save the file in Excel as a .CSV (comma delimited) file type rather than .wk1 and then
import that file into PCORD.
Combining two spreadsheets
Q. My dataset is in Excel, and is contained in two worksheets, one for each year of
data. Each worksheet has a slightly different species list, and the species are not
in the same order. How can I combine the worksheets?
A. Use Modify Data  Append Matrices to append a matrix of your choice to the main matrix.
The combined matrix is then offered for approval (as temp.wk1) as the new main
matrix. The main use of this procedure will be to combine community matrices that
differ somewhat in the species present. Using Append Matrix will correctly align the
columns of the matrices, regardless of the order of species in the two matrices.
Standardizing data
Q. When I use Ward's method with the Euclidean distance as similarity index to cluster
my data, are my column variables then automatically standardized? Or doesn't it
standardize my variables?
A. Your column variables are not automatically standardized. If you wish to do
that, you should use Modify data  Relativizations, then choose the method for
relativization that you want.
Cornell full format tables
Q. Am I correct that PCORD does not include importing Cornell full format tables
(species x frequency/abundance)?
A. A. PCORD imports Cornell condensed files but does not import the Cornell full
format table directly. That format is, however, easy to place directly into a wk1
spreadsheet, using one of the data tools with your spreadsheet. With Excel, paste
the data into a single column, then use Data  Texttocolumns (fixedwidth fields), then
add the header rows and save as a .wk1 file
Error reading main matrix. Data read error
Q. I have had no success in importing a CSV file into PCORD. I get the message
"Error reading main matrix. Data read error." How can I correct the
problem?
A. Check to be sure you have included the header rows. See the help system under CSV
format. Also, you can look at the text file "tempdat.tmp" in the
Application Data folder. You can easily find this file by selecting Tools 
Application Data Folder. You can open this plain text file with Notepad.
This file contains information for the the last item at the time the import failed.
This file contains one data item per line. You should be able to see the data for
the row that failed or any other information out of place.
Number of rows
Q. I'm having some problems reading my data matrix (*.wk1) and I'm hoping you can help.
The matrix is 13228 rows X 12 columns. When I try to open it, I get an error
message saying that "13228 rows" doesn't match the actual number of rows (8188).
There are no missing data values and no apparent reason why it should stop at row
8188. It always stops at that exact row, no matter what I try.
A. Some older versions of Excel export only the first 8192 rows (8188 not counting the
4 header rows). You can get all 13228 rows (actually up to 32,000 rows) of your data
into PCORD by saving the file from excel in CSV format. Then, the first time you
open it in PCORD, you will need to import it (File  Import  CSV). Then save it in
PCORD file. From then on you can use it as a normal data matrix in PCORD.
But if you reopen it and change it with that same version of Excel, you will need to go
the CSV route again to get it to save the whole thing.
Can not import Excel
Q. I get an "Invalid variant operation" when I try to import an Excel .xls or
.xlsx file. Is there a solution?
A. First download the latest modules and see if that solves
the problem.
Microsoft Excel must be on the computer in order to import. (Open Office will not
work.) During import, Excel is loaded into memory invisible and sent instructions to
open the file, and copy the data to the clipboard in text format so it can be
accessed. On a few systems this does not work for unknown reasons. If you
continue to have problems, save the file from Excel in CSV format and then import as CSV.
Can not import CSV
Q. I received a data file from a colleague overseas and cannot import the file.
What is the problem?
A. It sounds like the problem is due to the difference in the list and decimal
separators, one using comma and period respectively, and the other using semicolon and
comma. You can temporarily match your colleague's language platform by Start 
Control Panel  Regional and Language  Regional Options. Then import the file.
Error exporting to Excel spreadsheet
Q. When I try to export a matrix to *.xls or *.xls I get an "OLE error" if
there are more than 256 columns. This happens in both Excel 2007 and Excel 2010.
What is wrong?
A. You may have set the default format for saving to *.xls (Excel 2003). This
will cause an error in later versions of Excel if there are more than 256 columns.
To fix this:
In Excel 2007:
 Click the Office button in the upper left corner
 Click the Excel Options button at the bottom of the dropdown menu
 Select Save  Save file in this format: Excel Workbook (*.xlsx).
In Excel 2010:
 Select File tab
 Options
 Save  Save file in this format: Excel Workbook (*.xlsx).
Then in PCORD, export as .xlsx file type.
Convert Database Format to Spread
Q. How can I convert my data in a database to PCORD spreadsheet format?
A. PCORD versions 4 and above allow you to import/export files in a list (or database)
format. You can choose your field delimiters. In the help under file formats, see
List (Database) Format.
Error reading main matrix. Data read error
Q. I have had no success in importing a CSV file into PCORD. I get the
message "Error reading main matrix. Data read error." How can I correct the
problem?
A. Check to be sure that you have included the header rows in your CSV file. Also, look
at the text file "tempdat.tmp" in the application data folder. This file
contains info for the the last item at the time the import failed. This file
contains one data item per line. You should be able to see the data for the row that
failed or any other information out of place.
Graphs
Construct a dendrogram from TWINSPAN
Q. How do I construct a dendrogram (tree diagram) from the output of TWINSPAN?
A. It is possible to construct a dendrogram from the pattern of zeros and ones in the
margins of the TWINSPAN table. This pattern carries information on the series of
divisions performed by TWINSPAN. Observe the righthand side of the table first.
The first column of zeros and ones shows the first division  rows receiving a
zero were separated from rows receiving ones. Rows receiving the same digit are more
similar to each other than rows in different groups. Moving to the right, the next
column shows the next two divisions. Each of the preceding groups is again divided,
as shown by the pattern of zeros and ones. You can continue these sequential
divisions until you reach the rightmost columns  these groups were too small to divide
further. You can sketch in a traditional dendrogram to the right of the TWINSPAN
table. But it differs a little from the dendrogram you would get by cluster
analysis, because in TWINSPAN the finest branches of the tree usually have more than one
item on them.
View two graphs at once
Q. How can I view two graphs at once?
A. Yes. You can start two instances of PCORD, putting each in a window occupying half
of your screen. There is some potential for confusion if the two
"sessions" are working from the same graph files. Be sure to File 
Save As  Graph and give the graph file a unique name for each before graphing.
Insert graphs into MS Word
Q. What is the best way to insert PCORD graphics into word documents?
A. Consider the following tips:
 Edit the ordination completely in PCORD  do not attempt to reedit the picture in
Word.
 Use "save black & white" option (Graph  Options  Preferences) to avoid
gray scales for figures destined for blackandwhite reports and manuscripts.
 Drag joint plot labels into clear spaces for legibility.
 Edit title, axes, and legend in PCORD for clarity and informativeness
 When a figure is going to be reduced greatly in its final form, select larger font sizes
and larger symbols in PCORD.
 Use options for Legend Symbol/Color to select easily distinguished symbols.
 Results may differ in sizing and other ways if you insert pictures from saved files vs.
cut/paste using the clipboard. You may need to experiment with this.
 Resize the pictures in Word only by dragging the corners, not the sides, so that the
proportions of the graphic are not altered.
 To include additional text, symbols, or graphics, do it on top of the imported
pictures, not within them. This avoids editing the picture with Word, which will
usually force unwanted changes on the graphic.
Vertical axis becomes horizontal in MS Word
Q. When I insert a graphic into Word, the text in the vertical axis label becomes
horizontal. How do I fix this?
A.
1. Right click the graphic and select Edit Picture from the popup menu
2. Select the text box that you want vertical
3. From the menu select Format  Text Direction
Categories with text labels
Q. Can categories be given short text labels on ordination graphs, so that points are
labeled with text indicating their category?
A. Yes, PCORD 7 allows text as data for categorical variables. In versions 4 and
5 you can plot the row names from the second matrix, instead of getting those names from
the graph file. You can use this facility to assign ANY labels you wish, to be
included in the ordination. One application would be to give text names to groups,
as you suggest. First create a second matrix with row names that reflect group
categories and open this matrix as the second matrix in PCORD. Assuming you already
have done the analysis and have a graph file, open the graph window. Select preferences in
the menu and then select Labels From Second Matrix.
Groups with greater than 32 categories
Q. How can I overlay more than the 32 maximum categories on ordinations?
A. You can do this in two ways: 1) In the Graph  Preferences dialog, select Labels
From Group Variable. The values of your categorical (C) variable will then
label the points. If you have more than 32 categories there will not, however, be a
legend, but you can create one easily with your word processor by superimposing a text box
containing a legend of your own design. 2) You can change the row names in your second
matrix to reflect a group's categories and then select from Graph  Preferences  Labels
From Second Matrix. See Categories with text labels.
"Not a valid floating point value" error
Q. When I choose Overlay from Second Matrix or Joint Plot from the Graph menu I get the
following error message and the program freezes: '2,0' or ',002' is not a valid floating
point value? What is the problem?
A. The problem is the "," comma as a decimal point in the cutoff value.
International handling of decimals in PCORD works for most countries, but not all.
The solution is to make sure the Cutoff r2 value has a "." and not a
",". The easiest way to do that is change the international settings
to English (United States). From the Windows Start menu select Settings  Control
Panel  Regional Settings.
Species areas curves by group
Q. My dataset consists of data from 6 regions. I would like to run a species area
curve for each of the six regions (overlayed). Can PCORD run species areas curves
by groups and how would you do it?
A. There is no builtin option for species area curves by groups. Create a
separate matrix for each region, then run spparea curves for each. Each time you
run it, copy/paste the numerical output into a spreadsheet. Use "text to
columns" in Excel. Then you can plot the selected columns on a single graph
with Excel.
Ordination plot by group
Q. When I ran an ordination plot, I could not get the six groups to plot out as
different symbols or colors. Once I graph the ordination, the group button at the
top is not on. How do I fix this problem?
A. With the ordination graph menu up, go to Options  Preferences and make sure that
either "Color code categories" or "symbol code categories" is turned
on. If that doesn't work, be sure that you have declared the variable in the second
matrix as "C" for categorical.
Rotation of axes
Q. Does the new version permit rotation of PCA axes?
A. PCORD allows rotation of axes from ANY of the ordination methods. Varimax
rotation is, however, available only with NMS.
Edit and save graphics
Q. How can I edit and save graphic images of ordinations and overlays?
A. Graph an ordination. Then select File  Save Graph AsSave_Graph. The graph is
saved as a Windows metafile (*.wmf or *.emf). You can also save it as a bitmap
(*.bmp) or a jpeg (*.jpg) or a tiff (*.tif). This file can then be inserted into
many word processors and other programs where you can easily resize and edit it. For
high resolution bitmaps, see Resolution for Save and Copy.
Convert graphics to PDF
Q. Sometimes figures do not convert well to PDF. The left axis label consistently
ends up horizontal. How can I fix this?
A. Two suggestions:
 Use the enhanced windows metafile format (*.emf) if you are not doing so already.
 Try Adobe Acrobat Distiller. Graphics are, in general, more reliably converted to a PDF
by Distiller than by regular Adobe Acrobat.
3D graph label quality
Q. The labels on my 3D graphs are not crisp and have ragged edges. Is there a
solution?
A. This is a problem with Windows smoothing of fonts.
In Windows XP:
 Select Start
 Control Panel
 Display
 Appearance
 Effects and uncheck Smooth edges of screen fonts
 Click Apply and OK
In Windows Vista:
 Select Start
 Control Panel
 Personalization
 Window Color and Appearance
 Click on Open classic appearance properties for more options
 Click on Effects
 Uncheck Use the following method to smooth edges of screen fonts
 Click Apply and OK
In Windows 7:
 Select Start
 Control Panel
 System and Security
 System
 Advanced system settings
 Settings under Performance
 (or Performance Information and Tools
 Advanced tools
 Adjust the appearance and performance of Windows)
 Uncheck Smooth edges of screen fonts
 Click Apply and OK
If Windows 7 Control Panel is Viewed by icons instead of Category then
 Select Start
 Control Panel
 Performance Information and Tools
 Advanced tools
 Adjust the appearance and performance of Windows
 Uncheck Smooth edges of screen fonts
 Click Apply and OK
See also 3D graph label holes or rotation trails
3D graph label holes or rotation trails
Q. Labels that are behind my 3D graph show through and/or all labels leave trails when
rotating the graph. Is there a solution?
A. In the Graph  Preferences  Format tab check Hide
labels that are behind graph.
Alternate solutions to try are:
Check Label one axis side so there are no labels behind the
graph
Uncheck Adjustable labels, especially with a joint plot graph.
If you are Vista or Win 7 user and getting rotation trails, you need to take an
additional step:
 Right click on empty part of screen
 Select Personalize
 Select Window color and appearance
 Select any of the Color schemes other than Windows Vista Standard or Windows
Aero
Note, for Vista or 7 users, the problem only applies to what you see on the screen.
Labels will be properly hidden without the extra step of changing the screen
appearance with Copy, Save, Print Preview, Print, and Save
Animated GIF.
See also 3D graph label quality
Add points/symbols for individual species
Q. How can I add in the points/symbols for individual species on the NMS or other
ordination plots.
A. Two things need to happen to get points for species on ordinations:
 1. You need to have generated a graphcol.gph file, which should be visible in the window
in the upper right. If you are using NMS, you get this automatically if you use autopilot.
If you turn off autopilot, to get the graphcol.gph file you must be sure to check the box
in NMS Setup  Ouptut options Calculate scores for (species) by weighted averaging. A
similar checkbox is available for other ordination methods. Note that if you save the
graphcol.gph and graphrow.gph files, you don't need to rerun the analysis to do this, just
open the saved files (File  Open  Graph Row File, Graph Col File.
Sometimes even when you request the scores for species, they are not produced. This
happens if one or more of your columns is empty (all zeros), such that the weighted
averaging step fails.
2. Graph  Options, preferences  Format tab  check Plot Points, Species (or whatever you
called your columns).
3D Plots Possible?
Q. Is it possible to generate 3dimensional (XYZ) ordination diagrams in PCORD?
A. Yes. This feature was added in version 5.
CCA
How CCA in PCORD differs
Q. How does CCA in PCORD differ from that available in other statistical packages?
A. CCA is not yet available in major statistical software packages (do not confuse it
with canonical correlation, an older eigenanalysis technique based on a linear response
model). CCA in PCORD does not have some of the options available in another
implementation (such as inclusion of covariates and detrending), but we think you will be
pleased with the spreadsheet data format, modern interface, integrated graphing, modest
price, and our fully integrated set of data transformations and other multivariate tools.
CCA in PCORD is easier to use than CANOCO but doesn't have some of the options in
CANOCO (e.g. inclusion of covariates). On the other hand, CCA in PCORD is
integrated into a relatively comprehensive package of community analysis, providing many
more classes of analyses than CANOCO and including numerous nonparametric multivariate
techniques. Even so, PCORD is much less expensive than CANOCO, allows larger data
sets, and a wider selection of graphical representations of community data (e.g. 3D
ordination, cluster analysis, twoway clustering, speciesarea curves, successional
vectors, etc.).
Comparison of PCORD and Canoco
CCA: cannot solve weighted least squares
Q. When trying CCA I get the message "cannot solve weighted least squares, please
check contents of matrices." What should I check for?
A. CCA cannot solve the regression equation under several circumstances. One
common cause is when two or more variables are essentially identical or identical.
Another possibility is that the same variables are included in both the main and
second matrices.
Order of variables in correspondence analysis
Q. Concerns have be raised about results from correspondence analyses depending on
order in which variables are entered  is this a problem? If so, has it been fixed?
A. This was mainly a problem with DCA and TWINSPAN, less so with CCA. The Windows
version has implemented Oksanen and Minchin's (1997) "super strict" criteria for
stability of CCA, DCA, and TWINSPAN.
Error message: delete empty rows
Q. I find an inexplicable error message with CCA using a monte carlo test. I am
told that I need to delete empty rows in my main matrix. The rows specified are not
empty. Indeed, none of the rows in the matrix are empty. The problem occurs only
when testing the hypothesis, "no structure in main matrix..." What is
wrong?
A. Probably you have many zeros in your main matrix. This is common, but if the
problem is extreme, then shuffling within columns of the main matrix will result in empty
rows. When this happens, the monte carlo procedure discards that particular
randomization. But if it happens many many times, eventually PCORD gives up and
reports the error. You may be able to solve the problem by deleting some rare
species (for example those occurring in less than 5% of the sample units). This
problem can also occur with any randomization test involving shuffling elements within
columns (species) of a community matrix (e.g. the monte carlo test in NMS)
Downweight in CCA
Q. To compare my DCA analysis with CCA, I would like to downweight the species data
also in CCA. Since this option is not available for CCA, how can I do it?
A. In PCORD you cannot downweight in CCA in exactly the same way as in DCA. But
you can downweight rare species prior to ANY analysis by appropriate adjustment of the
data. For example, a simple approach is to delete species that occur in fewer than
5% of the sample units. Other transformations, such as the information function of
ubiquity, downweight both very rare and very common species, maximizing the influence of
species that occur in half of the sample units.
CCA error printing coefficients
Q. When checking the result file of CCA I found following message: "Error printing
coefficients. Possible multicollinearity problem" inside a table. What
does it mean and what should I do?
A. It means you have variables in your second matrix that are very strongly correlated
with each other. Check their correlations (use checkbox for "list correlation
coefficients for second matrix"), then remove a redundant variable. For
example, if you have two variables, percent canopy cover and percent canopy open, one is a
linear function of the other. This can also arise with groups of related variables,
for example %sand, %silt, and %clay in soil samples. For most data sets, those three
will add to 100%, so only two variables are providing new information.
Passive samples in CCA
Q. In a CCA ordination, is there a way to enter some samples passively so as to see
where they fall out in a preexisting ordination?
A. PCORD doesn't have an automatic way of doing this, but you can easily do it with a
spreadsheet. Plug in the values for the environmental variables into the multiple
regression equation. If you use the raw environmental variables, be sure to use the
unstandardized regression coefficients. If you express the environmental variables
as standard deviates from the mean of the original data set, then you can use the
standardized regression coefficients.
DCA or CCA with negative values for data
Q. Is it not possible to use DCA with negative values for my data? I always get a
warning that there are negative values found.
A. No, you should not run DCA or CCA on a main matrix with negative values.
NMS
NMS setup stability criterion value
Q. How low can I set the NMS stability criterion? What is the consequence of
making it smaller?
A. You can set the stability value as low as you want, down to zero. The only
cost is in computing time. With a small to medium data set you should seek a stable
solution by setting it very low (e.g. 0.0000001) and set the number of iterations very
high (say 500).
NMS monte carlo shuffling error
Q. When I run NMS with a Monte Carlo test, the shuffling routine ends with a message:
"Shuffling data resulted in undefined distances over 10,000 times. Abandoning
efforts to shuffle." The results for the real data are ok, but I do not get
results from the Monte Carlo test. What should I do?
A. When the randomization routine creates one or more empty sample units, that
particular shuffle is discarded. If your matrix is very sparse (very many zero
values), it is very difficult to almost impossible to obtain many shuffles of the data
with no empty sample units. NMS is basically giving up on the Monte Carlo test for
your data. Consider the following possibilities:
 Can you aggregate your data and analyze it at a higher level? If your sampling is
hierarchical, the smallest units (say microplots within plots) are often empty. By
aggregating microplots up to the plot level, the density of zeros in the data becomes much
lower.
 Can you partition your data set into more homogenous pieces and analyzing them
separately? After removing empty columns from each submatrix, they should have a
lower fraction of zeros than the whole matrix.
 Can you discard rare species? Discarding species that occur in fewer than 5% of
the sample units will typically not change the results much. If none of those is
possible, then you're pretty much stuck without a randomization test.
NMS 3D solution
Q. I am trying to run a 3D NMS with autopilot off. It runs fine, but only two
axes are written into the graph file. Can you tell me how I can force a 3D solution
in NMS and see all three axes in the graph?
A. It sounds like you are requesting a 3D solution but PCORD is deciding a 2D
solution is better, based on the builtin criteria for selecting the "best"
solution. You can prevent this by unchecking the box "Step down in
dimensionality" and specifying a 3D solution. Then, if you request, for
example, 100 runs with the real data, it will choose the best 3D solution from those
runs.
Different solutions with NMS
Q. Every time I run NMS, 2D graphs come out differently, and I cannot figure out how to
get the same result repeatedly.
A. NMS differs fundamentally from other ordination techniques in a number of ways (see McCune & Grace (2002), chapter 16). One important difference is
that NMS is typically run from random starting configurations. This means that the results
will vary across different runs. Usually the variation is small, even though sometimes at
first glance it appears big. For example, axis 2 in one run may be nearly identical to
axis 1 in another run. Or an axis in one run may be essentially a mirror image of an axis
from another run. You can also get different solutions that are essentially the same, but
one is a rotation of the other. For other fundamental differences from other ordination
techniques, please see upper right of p. 131 in McCune & Grace
(2002).
Significance of correlations with ordination axes
Q. Does PCORD provide pvalues for the significance of correlations between scores on
ordination axes and other variables, such as 1) correlations between community ordinations
and environmental data and 2) correlations between ordinations and individual species? Or,
can the significance of the correlation be determined using tables with significance
levels for r?
A. PCORD does not provide pvalues for these correlations. It is important to understand
the reason why not. Please see the first column on p. 107 in the book, Analysis
of Ecological Communities.
This section starts out, "These correlations should primarily be used for descriptive
purposes. It is both desirable and possible to resist the temptation to assign a pvalue
for the null hypothesis of no relationship between ordination scores and some other
variable..."
The remedy for this  how to describe its importance  lies in the _effect size_ rather
than the p value. Please see the middle paragraph in that column. A short excerpt:
"...set your own standards for how small an effect size, as indicated by the rvalue
or rank correlation coefficient, you are willing to interpret. In almost all cases, this
threshold will be more conservative than one determined by the pvalue..."
If you don't like this logic, or an editor or reviewer insists on a pvalue, you can look
up and report p values (see end of that paragraph), but that makes you vulnerable to the
arguments presented earlier.
TWINSPAN
Construct a dendrogram from TWINSPAN
Q. How do I construct a dendrogram (tree diagram) from the output of TWINSPAN?
A. It is possible to construct a dendrogram from the pattern of zeros and ones in the
margins of the TWINSPAN table. See the example in McCune and
Grace (2002). The pattern or zeros and ones carries information on the series of
divisions performed by TWINSPAN. Observe the righthand side of the table first.
The first column of zeros and ones shows the first division  rows receiving a
zero were separated from rows receiving ones. Rows receiving the same digit are more
similar to each other than rows in different groups. Moving to the right, the next
column shows the next two divisions. Each of the preceding groups is again divided,
as shown by the pattern of zeros and ones. You can continue these sequential
divisions until you reach the rightmost columns  these groups were too small to divide
further. You can sketch in a traditional dendrogram to the right of the TWINSPAN
table. But it differs a little from the dendrogram you would get by cluster
analysis, because in TWINSPAN the finest branches of the tree usually have more than one
item on them.
Enhancements of TWINSPAN
Q. I would like to see dendrograms produced by TWINSPAN as well as other enhancements
to the program. Will TWINSPAN be improved in PCORD?
A. We decided not to spend much time trying to upgrade Twinspan, other than expanding
its memory capabilities, because of serious, fundamental flaws with the concepts behind
the method and the resulting poor performance of Twinspan with heterogeneous data sets
(see for example van Groenewoud 1992, Belbin, L. and C. McDonald. 1993). The problem
is how well the underlying structure of the data is represented by the twoway table
produced by Twinspan. Have you tried twoway cluster analysis? Or simple
cluster analysis? If you are not familiar with cluster analysis, we recommend using
the following options: Ward's method of linkage and Euclidean distance. If you get
groups of sample units that are acceptable, then save the classification as a new variable
(see option under cluster analysis), then run Indicator Species Analysis to tell you which
species are most strongly separated among the groups. Or use twoway clustering to
get a graphical representation of how groups of sample units and species are related.
