About 3i Interactive Key and Taxonomic Database Software

Keys    Downloads    Home page   

3i – Internet-accessible Interactive Identification.


3i is an easy-to-use set of software tools for streamlining the process of producing morphology-based, Internet-accessible taxonomic revisions. To complete a revisionary project using 3i, users enter or import taxonomic nomenclature, morphological character data, illustrations, and specimen occurrence (collection) data into 3i’s standards-compliant relational taxonomic database. 3i then automatically generates online interactive keys and species (and higher taxon) pages directly from the database. When 3i is installed on a web server, these pages become globally accessible via the Internet. Virtual online revisions generated by 3i may also be used to create manuscripts, including dichotomous keys, synonymies, and natural language taxon descriptions for traditional print publication. Taxon pages generated by 3i may be automatically exported to the Encyclopedia of Life (EOL), Catalog of Life (CoL), and 3i is also a registered data provider to the Global Biodiversity Information Facility (GBIF). Thus, 3i is one of the most versatile and comprehensive cybertaxonomic software systems currently available.

3i is designed to accelerate the process of producing a taxonomic revision (both online and in print), especially for highly diverse groups of organisms. By linking and organizing the large amounts of nomenclatural, morphological and specimen-level data, plus images, needed to produce a revisionary synthesis, 3i automates many of the mechanical and repetitive aspects of revisionary studies, allowing taxonomists to focus on the core intellectual pursuits of their profession: comparative morphology and species discovery. Specific features of 3i are described in more detail below:

  1. Multi-entrance polytomous (matrix-based) keys, with unlimited numbers of characters, character states, and taxa.
  2. Support for several different languages (Example 1 – Eng, Example 2 – Rus).
  3. Dynamic updating of character list after each step of identification: characters no longer relevant removed; states not relevant for further identification marked.
  4. Support for numeric (continuous) characters.
  5. Characters may be sorted by morphology, by rank (assigned by the author), or by separating power recalculated after each step of identification.
  6. Support for linking characters and taxa to explanatory images (Example).
  7. Taxon names are hyperlinked to autogenerated taxon pages (Example).
  8. Taxon pages may include autogenerated descriptions (Example), synonymies (Example 1, Example 2), distribution maps (Example 1, Example 2), lists of specimens examined, and bibliographies (Example).
  9. Host associations, if included in specimen records, may be displayed as bar charts showing abundance on each host. (Example).
  10. Keys can include taxa of different hierarchic levels (species, genera, tribes, etc.). The number of hierarchic levels is unlimited. Users may limit identification to a single hierarchic level.
  11. Simple and advanced (Boolean) database search for names, geographic distribution, host plants, collector, etc. Results of a search may be applied as a filter to create a customized interactive key. (Example).
  12. Export of data matrices as Nexus files for phylogenetic analysis. A phenetic tree may be generated on-line from the data matrix.
  13. Import of pre-existing data matrices (from Excel spreadsheet).
  14. Visit statistics with recognition of users. Storage of individual preferences for users visiting the web site.
  15. Compliance with DarwinCore standard for specimen-level taxonomic databases and export of DarwinCore and SDD archives.
  16. Creation of multiple keys from the same data matrix.
  17. Creation of customized keys using search-based filters (e.g., species from a particular locality or host plant).
  18. Interactive dichotomous key building utility for creating print-publishable keys.
  19. A validation utility for nomenclatural data, that follows the Code of Zoological Nomenclature.

For a more detailed list of features and comparison of different programs for interactive identification see the table by A. Lobanov et al. This is a modified version of the table originally compiled by M. Dallwitz including additional features and expanded list of programs.

The 3i program has already been used to create several interactive keys and taxonomic databases, which are listed below. These databases are still under development, and do not present all of the above features (numbers in brackets indicate the features supported by each key below):

 

To use the 3i keys, Internet Explorer 5.0+ browser is recommended for better performance. Other browsers may cause undesirable changes in the Key interface. However,  the program has been tested and works correctly with the following browsers: Internet Explorer 5–6, Netscape 4–7, Opera 6–7, Mozilla, Safari, Firefox, and should work with any other browser supporting frames and Java Script.

The 3i key is based on a character state matrix stored in a MS Access database. The program uses the capability of ASP-pages to query the database and retrieve the names of taxa matching the states specified by the user. If a character is not scored for a taxon, this taxon will remain in the list whatever state of this character is chosen.

The main work field presents the list of characters. Each character has a drop-down box with a set of states. The characters can be sorted in three ways. 1. By rank. The author of the Key specifies which characters are the best for identification, such as the characters which practically cannot be misidentified, or characters which vary less (but not always the easiest to see). The rank of the character is specified in parentheses after the character name. 2. By separating power.  The author ranks are disregarded and the program calculates which characters better split the remaining taxa into groups of equal size (this order is recalculated after each step of identification). 3. By morphology. The characters in this case are grouped according to morphological regions, such as head, thorax, etc. Within each region, they are ordered by rank.

Identification can be started with any character. A state of a character can be chosen directly from the drop-down box. Characters can have hyperlinks to explanatory images. After the states of one or several characters have been chosen, the Proceed button should be pressed in order to renew the list of taxa matching the search criteria and the list of eliminated taxa. Characters irrelevant for further identification are removed from the list. The states not relevant for further identification are not removed, but instead are marked with minus symbol. The number after each state shows how many taxa will stay in the list after selecting it. If a character has multiple states and there is uncertainty about which one to chose, the NOT option can be helpful; the taxa not having this state will remain in the list. Some characters may have no states, requiring an exact numeric value, such as a measurement, entered into the field (the range of valid values is shown in parentheses after the character name). The set of taxa has hyperlinks to the taxa descriptions.  The button Clear all is used to return the list of characters to its original state and start identification from the beginning.

The Preferences button allows one to modify the interface and the list of characters displayed. For example, if your specimen is a female, you can choose to display only characters applicable to females by deselecting the check box next to "Males". The applicability of each character based on stage/sex is shown in parentheses together with the importance index: M - male characters, F - female characters, N - nymphal characters, L - larval characters, P - pupal characters, E - egg characters, and D - damage characters. If your Internet connection is slow, you can slightly increase the download speed by choosing not to display eliminated characters or/and taxa. The preferences also include options for sorting the character list described above, and the error tolerance drop-down box. A few words should be said about the last option. If, after several steps of identification, you have no taxon left matching to the criteria of search, you can either go one or more steps back and try other states or characters, or increase the error tolerance. The error tolerance is the number of mismatching characters allowed during identification (default value is "0"). After each step of identification you may see the number of mismatching characters in parentheses before the taxon names. By increasing the error tolerance you relax the criteria of search and thus increase the list of taxa remained after each step of identification. Taxa are eliminated only when the mismatch exceeds the specified error tolerance. Optional features are Language switch and Rank limit. The last one allows to limit identification to higher taxonomic rank (for example, if the key is for identification of species, the list of taxa can be limited to genera or tribes).

To refresh the screen with new options, the Proceed button should be pressed. Some old browsers may need to be closed and restarted in order for the pages with new settings be downloaded correctly; otherwise the computer will keep displaying the pages with old settings stored in the browser's cache. User preferences are saved after the end of each work session.

Collaboration

The 3i program was designed to work with any taxa at any level of the biological hierarchy. It is distributed for free on the following conditions:

  1. Program is used only by the recipient, and is not for further distribution.
  2. Program is used as a whole package, without any changes or modifications.
  3. You agree to test the program and send me feedback with description of problems experienced while using the program, how easy or difficult it was to learn to work with the database, what other features are desirable for the Internet-accessible key, or any other suggestions on improvement.
  4. In a month after the program has been installed on a server, it will ask for an on-line registration. Currently, I need only the Http-address of the web site, the name of interactive key, and the author name(s). I would like to maintain a list of interactive keys created using the 3i program.

 

Gregor Hagerdor developed 3i - SDD converter. The export is indirect, the original 3i database should be converted into DiversityDescriptions (converter available since version 2.0), SDD could be created from there. As a result, some details (specimen, nomenclature, publication data), which could in principle be expressed in SDD 1.1, are getting lost because they were not fully supported by DiversityDescriptions.

The taxon pages generated by 3i may be automatically exported to the Encyclopedia of Life, Catalog of Life, GBiF, Discover Life, Global Name Index, etc.

To create a new 3i interactive key, a PC computer with Microsoft Access 2000, XP, or 2003 (for the database editing) and any Microsoft Web Server (for the Internet publication) are needed. The 3i interactive keys can be viewed through the Internet only. If you do not have Microsoft Access or you have an earlier version, you can download and install MS Access 2000 Runtime, which is a freely distributed version of Access that will allow you to run the 3i database, but not design databases.

Download 3i program and description (3i.zip, 5.7 mB, Updated December 15, 2011).

Installation under Windows XP Professional.

Installation under Windows Vista.

Installation under Windows 7.

How to create the first interactive key using 3i.

Export from the 3i.

Download MS Access 2000 Runtime (ART2KMin.zip, 33 mB).

The 3i program was downloaded 1404 times.

Acknowledgements

Development of the 3i program and work on interactive keys was supported by NSF-REVSYS grants DEB0315373, DEB050529679, and DEB0715499 and Hatch award ILLU-875-361.