Publicly available DELTA data; Watson's Angiosperm families

[Mike Dallwitz]

                                                                  7 March 1995

> From: bwilliam at oyster.smcm.edu (Bill Williams)
> To: Delta-l
>
> Is there somewhere a list of all publicly available DELTA data sets? I have
> a couple of undergraduate students trying to get INTKEY up and running for
> possible student use, and they'd very much like to find some North American
> angiosperm data (they [and I!] find the "ANGIO" data intimidating because it
> has so many families [and characters!] they've never heard of).

The only publicly available data sets I know of are those on the DELTA
gopher/ftp areas (ftp://muse.bio.cornell.edu/pub/delta and
ftp://spider.ento.csiro.au/delta). I would be happy to add other interactive
identification packages, whether using DELTA/INTKEY or other software. If you
send me a contribution, please archive the files in a self-extracting file
(preferably using LHA). Also, please include a short file Download.1st
containing a BRIEF description of the package and instructions for downloading
and installing it.

You should be able to overcome your problems with Leslie Watson's Angiosperm-
families INTKEY package. Most of the information you need is in the file
ANGIO.1ST which accompanies the package. Here are some relevant extracts,
together with my additional comments. (For brevity, the INTKEY operations are
given in their command-line forms, but they can all be invoked via the menus.)

    Subsets [of the taxa] which should be more or less representative of large
    world regions can be obtained by using the `geographical distribution'
    character (accessible via the keyword `distrib'), which has been provided
    especially for this purpose. For example, to obtain a subset for North
    America, enter state 12 of this character, followed by INCLUDE TAXA
    REMAINING or SET FIX ON.

INCLUDE TAXA REMAINING will make all subsequent operations apply to only the
North American taxa (as recorded). SET FIX ON will `fix' the distribution
character so that it is not cleared when a new identification is started. The
latter method has the advantage that the `fixed' character is subject to the
error-tolerance mechanism, so that a family not recorded from North America
can still be identified.

    [Alternatively,] taxon keywords, representing subsets of the taxa, can be
    defined via lists of names. Lists for a few geographical regions are
    provided with the data. For example, the file AUSTRALIA.TAX defines a
    keyword AUSTRALIA for all the the families represented in Australia. The
    definition can be invoked in INTKEY, when required, by the FILE command:
        FILE INPUT AUSTRALIA.TAX
    The keyword can then be used with any command requiring a set of taxa. In
    particular, if you enter
        INCLUDE TAXA AUSTRALIA
    you will be working exclusively with the Australian subset. To resume
    working with the full data set, enter
        INCLUDE TAXA ALL

To make things easier for students, you can use this method to restrict
operation to even smaller subsets of taxa, for example, for the state, the
university campus, or a set of specimens being used in a practical class. Any
of the above methods can be incorporated in the initialization file,
INTKEY.INI, so that the restricted set of taxa is the one seen when the
program is started.

    Detailed information on particular characters and states is available via
    the character notes. If you are uncertain how particular character states
    are being applied, use the program to obtain lists of families
    exemplifying them, and examine representatives available locally (SET
    MATCH E (Exact) should be invoked, to eliminate taxa that are variable for
    the characters in question). Some of the `characters', though clearly
    contributing to the usefulness of the data, are not taxonomic characters
    in the normal sense (geographical distributions, numbers of genera and
    species per family, synonyms etc.). ... When attempting identifications,
    you will normally wish to exclude from consideration such quasi-
    characters, and esoteric characters such as anatomy, physiology and
    cytology. The character keyword NONDIFF has been provided, for convenience
    in excluding the kinds of characters which will not normally be required
    for routine identification. Enter
        EXCLUDE CHAR NONDIFF
    to invoke it.

You can take this further by defining and using additional keywords to exclude
(or include) characters that you think inappropriate (or appropriate) for the
level of knowledge of the users. Again, this reduction in the number of
available characters can be done in INTKEY.INI.

    Botanists with sufficient knowledge of plant families to select characters
    for their known reliability and diagnostic value will use the system more
    efficiently than others. The interactive technique, however, makes
    reliable identification more accessible for persons without specialist
    knowledge, and its use will help them acquire it. If you are at a loss
    where to start with a specimen, first scan the character list for any
    characters you feel you can confidently use. ... Such obvious and
    relatively unambiguous features as phyllotaxy, habit (herb/tree), aromatic
    foliage, gland dots, will be found to have great diagnostic value. ...

    Contributions of descriptive data and illustrations suitable for improving
    and extending the database would be welcomed, and appropriately
    acknowledged in the displays and in the accompanying literature.

Character illustrations, and additions or improvements to the character notes
would also be welcomed, and quickly incorporated in new releases of the
package.

Mike Dallwitz                                  Internet md at ento.csiro.au
CSIRO Division of Entomology                   Fax +61 6 246 4000
GPO Box 1700, Canberra ACT 2601, Australia     Phone +61 6 246 4075