continuous characters (fwd)
jiel at BCC.ORST.EDU
Wed Aug 31 23:36:46 CDT 1994
On Wed, 31 Aug 1994, Liu Jie wrote:
> This is Jie Liu again, I would like to forward following message to the
> list. Some of you may be interested in.
> Kevin Thiele (Cladistics 1993, 9:275-304) state that: Data scored for
> cladistic analyses may be quantitative or qualitative, continuous or
> discrete, and show overlapping or non-overlapping values between taxa.
> Pimentel and Riggins (Cladistics 1987, 3:201-209) state that: Only
> character state derived from characters showing discrete variation can
> provide phylogenetic information. Continuous characters are
> transformational homologies not subject to test and, therefore, invalid
> for phylogenetic analysis.
> I am working with nematode taxonomy, wish to relate species by
> phylogenetic analysis. In some groups of nematodes, differences between
> some taxa may often be "dramatic", e.g. by length of infective juveniles
> and many of the characters that can be used to separate taxa are quantitative
> characters and show continuous variation. If I agree with Thiele, I can try
> to use generalised gap-coding method ( or other methods) for coding
> continuous data into discrete characters. If I agree with Pimentel et
> al., I don't know of any method might be useful to deal with
> non-descretely characters for phylogenetic analysis. Is that possible
> for us to develop methods for coding the data to certain states and/or
> a computer program which can run with these data or with continuous
> data for phylogenetic analysis ?? Do you agree that phylogenetisists
> should be interested in all heritable features of the organisms they are
> studying, any observed trends that have a genetic basis should
> potentially be available for analysis ?
> I would be very appreciated your suggestions on my problems.
> J. Liu
> ---------- Forwarded message ----------
> Date: 31 Aug 94 10:48:16 U
> From: p stevens <p_stevens at nocmsmgw.harvard.edu>
> To: jiel at BCC.ORST.EDU
> Subject: continuous characters
> Dear Liu Jie,
> Your message on continuous characters on TAXACOM was forwarded to me; perhaps
> you in turn might forward this message to that list...
> The issue of continuous characters is to me one of the more interesting we
> face. You probably know about the dimension of the problem (Syst. Bot. 16:
> 553-583. 1991. oulines it): many co-called qualitative characters are in fact
> overlapping quantitative variation in disguise.
> With Nancy Gift, I have recently been carrying out experiments on how people
> score more or less continuous overlapping quantitative variation; the same
> data, presented in three different fashions, may be scored in up to 11
> different ways (19 people were involved in the experiment). In one case,
> changing the ordinate to log meant that a different division of the data into
> states was adopted... We have not finished our study, but earlier analyses of
> character states derived from the subdivision of such overlapping variation
> gave a rather different tree from the analysis of more discrete/qualitative
> characters; preliminary analysis of molecular data by Kathy Kron yields a
> topology similar to that of the discrete/qualitative tree. I myself would
> definitely be inclined to stay away from using such overlapping variation, or,
> at the very least, experiment -very- comprehensively with its subdivision.
> You will also probably be aware of a recent paper by Kevin Thiele (Cladistics
> 9: 275-304. 1994). He clarifies the confusion of terms used when discussing
> this problem, although I cannot say that I agree with his solution. However,
> the ultimate arbiter of all attempts to deal with such "nasty" variation is
> going to be congruence with other data sets, which is why the work I mention
> above is so exciting to me - we are getting several data sets bearing on the
> one problem. I have not yet had time to see how to deal with our data in the
> context of Kevin T's study, and so do not know if preparing the states in his
> way would yield yet another topology or not.
> A final point to note is that there is a substantial body of literature
> with vision, hearing, and language that is relevant to this issue of getting
> character states from overlapping continuous variation - the issue there is
> categorical perception, does it exist, and in what form. This is exciting
> stuff, because I think some of the broader issues here spill over into such
> fields, as well as ethnobiological classsifications and typological thought.
> I would be very interested in your ideas about this problem.
> Peter Stevens (pstevens at harvard.edu).
> Forward by
> Jie Liu e-Mail:jiel at bcc.orst.edu
> Department of Entomology Phone:503-737-5526 (O)
> Oregon State University 503-753-4642 (H)
> Corvallis, OR 97331 Fax: 503-737-3643
More information about the Taxacom