
History

Tue Feb 20 17:03:38 1996

To resolve a 'NO SUITABLE CHARACTERS' error report from the key program,
in the ITEMS file the state of character 30 of the
wahlgreni item was changed as such : 30,-  -> 30,1

To allow the key program to take into account all characters,
all numeric characters are mapped to artificial states (look for *KEY STATES 
in the TOKEY file).  As a result of this it was required to increase the 
MAXIMUM NUMBER OF STATES in the SPECS file from 10 to 20.

To be able to actually use character 33 in the key, the 
ITEMS file needed modification from 33,- to 33,1.5 (= a kind of default 
max length...) for the following items : confusa, handschini, petterseni, 
subilliciens, vandeli meridionalis, vandeli praealpina and vandeli relicta.

Finally, character 34 is excluded from the key to increase the
readability of the key (possibly not a good choice).

The result is an incomplete key with special emphasis of incompletness
with regard to the European species...

Fri Feb 23 14:25:01 1996

> [Dr Christiansen]:
> I think that 34 is 
> not a good choice since most species have this information available. 36 or 
> 37 would be better since many species do not have this information 
> available.

Character 34 is reintroduced. At the same time
all character states '-'(=not applicable) are replaced by 'U'(=unavailable).
This results in a much better key.  The change is dramatic.  Almost all species
can be keyed out now. None of the characters were excluded and the key program
selected 37 out of the 40 available characacters to build the key.

Two problems still remain to be solved :
1. a lot of species appear more than once in the key.  This is caused by
	the intra-taxon variability. And this is caused by the great number
	of unavailable character states.
2. some species appear at the same place in the key.  This also is caused
	by the unavailability of character states for these species.  It
	could also be an indication of a possible synonym.

This is the list of 'undistinguishable' species with an indication of the
unavailable character state that could bring a solution :

....... inaequalis	34,U
        styriaca	34,40-50
 
....... aggtelekiensis	34,20
        caladairensis	34,U
 
....... kersmaekersi	34,U
        aggtelekiensis	34,20

....... aggtelekiensis	34,20
        melatensis	34,U
 
....... aggtelekiensis	35,2
        chapmani	35,U

....... virei		possible synomym?
        subvirei
 
....... flatua		possible synonym?
        sp.V
 
....... fasciata	34,U
        spelunca	34,U

Thu Feb 29 12:23:59 1996

> [Dr Christiansen]:
> The missing values for character 34 should be as follows; inaequalis 67; 
> caladairensis 34; kersmaekeri 15; melatensis 25; fasciata 72 and spelunca 
> 15. chapmani character 35 is 1.7-2.0. Eliminate species V.

Introducing these character states solved all but the virei/subvirei
problem.  Based on the currently available character states for these
species they cannot be keyed out seperatly.

Some species apear more then once in the key due to unavailable
character states :

aggtelekiensis            3 times
binoculata                2
encrusae                  2
fallax                    2
goughi                    2
handschini                8
imparipunctata            2
joupani                   3
lleidensis                2
marteli                   7
passauri                  3
picta                     2
propinqua                 3
stachi                    8
styriaca                  2
subduodecima              4
subilliciens             10
tarraconensis longicorni  3

Frans Janssens
frans.janssens@advalvas.be
http://www.geocities.com/~fransjanssens/
