BugStats produces multi-sample environmental summary diagrams from coded
habitat data. When used with stratigraphic sequences or repeated pitfall
trapping, these diagrams can be used to identify environmental changes
through time as reflected in the taxa present in the samples. When used
with archaeological samples they can be used to illustrate spatial variation
in human impact.

The method is simple, transparent and uses broad (eco)habitat classifications.
Ca. 4995 taxa have been assigned to 22 habitats (BugsEcoCodes, ca.
8317 designations) through a combination of examination of various
published habitat descriptions and the detailed ecology codes assigned
by Koch (1989-92). 125 Koch codes are included
in BugsCEP for ca. 2920 species (ca. 11192 designations), but are
yet to be included in graphical outputs. |
 |
The categories are based on those most commonly used in Quaternary
research, with a little adjustment to cater for archaeological investigations.
In particular, the work of the Bugs authors, and Mark Robinson (e.g.
2000), Philippe Ponel (e.g. 1995)
and Harry Kenward (e.g. 2001) have influenced
the category definitions. The species designations are those of
the Bugs authors, with a considerable influence from Klaus Koch's
work. Due to the nature of these influences the codes are most likely
to be more ecologically useful in Central to Northern Europe.
Taxa may be assigned to more than one habitat, but may exist in
only one 'indicator' class at a time.
|
Despite a probable northern bias the use of a constant system enables
reproducibility in any fauna, and thus the system could even be useful
where the habitat coding is less accurate.
More information is available in Buckland (2007).
The BugsEcoCode system produces quantitative reconstructions from qualitative
habitat descriptions/designations. It is thus a reproducible semi-quantitative
system which allows advanced inter-site comparisons. Raw and calculated
data is always output in the export files, and a sample by sample breakdown
of species and their codes and abundances can be obtained.
BugStats is to some extent a work in progress, and we would very
much appreciate feedback on the usefulness and problems with the BugsEcoCode
calculation system and outputs.
Why BugsEcoCodes &
BugStats |
- There is need for a standardized (i.e. regionally comparable)
& transparent system for summarizing palaeoentomological interpretations
& reconstructions (from 100's of taxa, or very few).
- Ordination method results (PCA, CA etc.) not comparable in
terms of inter-site reconstruction - habitat definitions derived
in this way are only applicable within the context of the original
study.
- There is currently no existing routinely used software for
numerically reconstructing habitats from fossil insects - there
is a need for (at least semi-)quantitative environmental reconstruction
software.
- No existing system works as well on modern data as fossil -
allowing long term (palaeo)biodiversity studies.
- Palaeo-data often violates/confuses assumptions needed for
many advanced statistical methods (differential taphonomy, uncertain
representation, many 0's). BugStats does not.
- No current habitat summary system has its species designations
fully and publicly available.
|
Generate environmental
reconstructions for a site & produce BugsEcoGraphs |
BugStats can produce sample habitat
summary diagrams for any site which includes taxa that have been assigned
BugsEcoCodes. The graphs allow direct comparison of the environments
represented by samples between samples and sites. A number of standardization
and calculation options are available.
BugStats uses the concept of environmental representation (env. rep.)
as follows:
1 environmental representation = 1 environment represented by 1 individual
or taxon. A taxon can represent more than one environment.
The raw results are counts of environmental representation in each
habitat class for each sample.
These can (optionally) be standardized by dividing by the sample sum
to allow samples to be compared on a common (percent) scale.
|
- To open the BugStats interface click the [BugStats] button
on the main toolbar.
- Make sure MS Excel is closed and not running in the background.
- If you have created
a site, and entered
abundance data for it then it will be available from the 'Site'
drop down box towards the top of the screen. (Use the [Refresh
sites list] button if it is not there).
Sites without abundance data are not shown.
BugsCEP already contains abundance data for over 550 published
sites - try some!
- If the site has more than one countsheet, then select the required
one from the 'Select Countsheet' drop down box.
- Select the desired options, as described below (and in more
detail further below)...
|
 |
Calculations
- To species id's only - only use taxa identified to species
level (i.e. omit sp. spp. and indet. from calculations). Creates
a more precise reconstruction, but may miss information in poorly
resolved faunas.
- Transform abundance log(n+1) - applies a natural logarithm
transformation to the raw reconstruction counts (plus 1) before
standardization. Some authors believe this compensates for inter-sample
variation in a number of ways, including log/normal abundance
distributions.
Note that this will not be indicated on the output file, so
please name your file accordingly.
|
|
- Standardization
- No abundance, % sumrep - Taxa occurrence only (presence/absence).
Env. rep. sum for habitat class expressed as percent of sum
of env. rep. for sample (sumrep).
- No abundance, raw - Taxa occurrence only (presence/absence).
Env. Rep. counts for habitat class only.
- Abund. weighted, % sumrep - Taxa env. Rep. multiplied
by its abundance in sample.
Env. Rep. sum for habitat class expressed as percent of sum
of env. Rep. for sample (sumrep).
- Abundance weighted, raw - Taxa env. Rep. multiplied
by its abundance in sample.
Env. Rep. counts for habitat class only
|
|
Charts
- Use Ref as sample names- Use sample Reference/Context
field to label samples rather than database sample code (recommended).
- Descriptive titles - use real words for habitat titles,
rather than BugsEcoCodes.
- Autoscale x-axis - BugStats will scale diagrams to fit
all data. Uncheck this to specify a maximum and interval yourself.
- Include sum graphs - Output graphs showing (shown to
the right):
Sumrep - sample environmental representation sums.
Abundance - sample abundance sums.
NSpec - number of taxa per sample.
- Graph title - Use this box to give the graph a meaningful
title (BugStats will append the chosen standardization option
to this name on output).
|
 |
- Click the [Create Bugs Ecofig] button, choose a save
location and provide a filename, click [Save] then stand
back and watch it calculate.
Do not touch the computer until the 'Completed' message appears
- BugStats needs full control of your system, and could crash
if you interrupt it.
- When finished, the name and path of the export file will be
given in the 'Ecofig file created' box towards the bottom of the
screen. Click the [Open Results in Excel] button to the
right of this to see the results.
See below for an explanation of the results
and some tips for preparing the graphs for presentations/publications.
|
See a sample by sample species breakdown
of BugsEcoCodes for a site
|
The BugStats species breakdown report allows
you to see exactly which represent which environments, and how abundant
they are, in each sample. This is a useful too when trying to understand
the environmental reconstruction diagrams.
|
- To open the BugStats interface click the [BugStats] button
on the main toolbar.
- Make sure MS Excel is closed and not running in the background.
- If you have created
a site, and entered
abundance data for it then it will be available from the 'Site'
drop down box towards the top of the screen. (Use the [Refresh
sites list] button if it is not there).
Sites without abundance data are not shown.
- If the site has more than one countsheet, then select the required
one from the 'Select Countsheet' drop down box.
- Click the [Explore EcoCodes for Selected Site] button,
and the report will be shown in preview mode...
|

|
- As always in BugsCEP the report can be browsed in the preview
window; printed directly (check the number of pages first!), exported
to MS Word or exported to MS Excel using the buttons at the top
of the window.
Remember to save the file after exporting in the application
exported to.
|
Compare samples
at a site using correlation coefficients |
Correlation coefficients can be used to compare
the similarity/dissimilarity of the species composition of samples.
This is especially useful when trying to identify samples that represent
similar environments in Quaternary geology and environmental archaeology.
BugsCEP currently only supports one correlation
coefficient - the 'modified Sørensen's' coefficient of similarity
of Southwood (1978), which is the inverse
(1-B) of the Bray-Curtis coefficient of dissimilarity (Krebs,
1989). Additional coefficients will be added with time.
The results are exported as MS Excel files and
can be used as the basis for building cluster diagrams/dendrograms,
but no such feature is present (yet...) in BugsCEP.
NOTE: This feature is currently passworded
and locked due to lack of development time - it's ugly, and we really
don't know if it works on all systems! If you are particularly interested
in the feature then please contact us and we will provide you with
the password if you agree to help us debug the module... it will
be opened up just as soon as we know it works properly.
|
- To open the BugStats interface click the [BugStats] button
on the main toolbar.
- Make sure MS Excel is closed and not running in the background.
- Click the [Calculate Coefficients] button and enter the
required password to open the calculations pop-up window.
- Select a site from the 'Site' drop down box.
- If the site has more than one countsheet, then select the required
one from the 'Select Countsheet' drop down box.
- Check [] Transform abundance ln(n+1) if required.
- Click [Calculate & Export to XLS], choose a save
location and provide a filename, then click [Save].
- Stand back and let BugStats calculate - do not disturb it!
- You will be informed that calculations are complete and asked
whether you want to open the results in Excel.
See below for an explanation of the results
file.
|
BugStats output files
explained |
BugStats produces three types of output:
- BugsEcoGraph files (XLS format)
- Site EcoCode breakdown reports (preview, print, export)
- Coefficient files (XLS format)
|
BugsEcoGraph files
Three worksheets: Graphs; PctResults; RawResults
|
Graphs
A horizontal series of 24-27 bar chart like figures (depending
on options) as shown below:

The first figure shows sample names and the diagram title provided
by the user. The next 23 figures show the BugStats results for each
sample in each of the 23 habitat classes of the BugsEcoCode system.
Each figure is titled by the class name, and has a scale at the
bottom which is either in percent or counts depending on the options
chosen.
The final three figures provide summary information for each sample
if requested. See above for details.
|
PctResults and RawResults
RawResults shows the raw counts of environmental representations
for each sample in each habitat class.
PctResults shows the raw counts recalculated according to the options
selected in the BugStats interface. These values are used to construct
the figures on the graphs worksheet.
'Indicator' classes are not included in standardization calculations
as they are subclasses.
The last few columns show sample data, including reference/name,
coordinates and depth.
|
Coefficient files
Results are exported as a matrix, with sample names in the first
row and first column. The remaining cells show the coefficient of
similarity values for each sample compared with every other one.
The data is formatted and colour scaled here for illustration.
A value of 1 represents total similarity, and 0 no similarity.
|
 |
Tips for improving the graphs for presentation/publication
Copy and paste the figures into vector graphics editing software
(e.g. CorelDraw)
- With the 'Graphs' worksheet selected in a BugStats output, activate
the 'Drawing' toolbar and select the 'Select Objects' tool.
|
- Select all the figures - easiest done by scrolling to the bottom
right and dragging a box around them.
- Press CTRL+C to copy the images to the clipboard.
- Switch to your graphics package and paste special as 'picture
(metafile)'.
- The figures will probably be grouped - ungroup the objects and
then polish the image as you see fit for your purpose.
We recommend deleting empty figures (environments not represented
by the faunas), enlarging titles and scale texts.
|
The BugStats methods explained
|
BugsEcoGraphs
Please see Buckland (2007) for a detailed description.
|
Correlation coefficients
Please see Buckland (2007) for a detailed description.
|
BUCKLAND, P.I. (2007). "The Development and Implementation of
Software for Palaeoenvironmental and Palaeoclimatological Research: The
Bugs Coleopteran Ecology Package (BugsCEP)". PhD thesis, Environmental
Archaeology Lab., Department of Archaeology & Sámi Studies.
University of Umeå, Sweden. Archaeology and Environment 23, 236
pp + CD. Available online: http://www.diva-portal.org/umu/abstract.xsql?dbid=1105
KENWARD, H (2001). "Insect remains from the Romano-British ditch
terminal at the Flodden Hill Rectilinear Enclosure". Reports from
the Environmental Archaeology Unit, York 2001/49, 15pp.
KOCH, K. (1989-92). Die Käfer Mitteleuropas. Ökologie, 1-3.
Goecke & Evers, Krefeld.
KREBS CJ, (1989), Ecological methodology. Harper Collins Publishers,
New York, USA, p. 654.
PONEL, P. (1995). Rissian, Eemian and Würmian Coleoptera
assemblages from La Grande Pile (Vosges, France). Palaeogeography,
Palaeoclimatology, Palaeoecology, 114, 1-41.
ROBINSON, M. A. (2001) Insects as palaeoenvironmental indicators.
In, D. R. Brothwell & A. M. Pollard (eds.) Handbook of archaeological
sciences, 121-133. J. Wiley & sons, Chichester.
SOUTHWOOD TRE, (1978), Ecological methods, with particular reference
to the study of insect populations. John Wiley & Sons, New York, 2nd
ed., 524p.
|