Nhu Nguyen xerantheum@gmail.com
Tue, 05 Oct 2010 17:04:16 PDT
Hi Cody,

Thanks for pointing out this paper. I have not seen this paper but just now
had a look.  You said that "It deals specifically with the data used in the
'04 analysis" but actually, there is no overlap in the datasets at all! Let
me explain.

In Manning et al. (2004), the authors used two genes, *trn*L-F and *rbc*L in
their analyses. In Lebatha et al. (2006), the authors used morphology and *
trn*L-F. They found that using just morphology alone, their analyses support
the 3 groups. If they use the DNA data alone, the analyses support Manning
et al. (i.e. a unified Ledebouria). In a combined morphology and DNA, their
data support the 3 group scheme.

Here are my questions: what happened to the gene rbcL? Why don't they re-use
the data that was published by Manning et al. in combination with their new
morphological data? Instead added totally new taxa to their analyses and
removed the taxa used by Manning?

These two questions may seem mundane but they have important implications in
what an analysis will tell you. For the first question, Manning uses 2
genes, which is always better than just 1 gene. The more gene you add to
these types of analyses, the better your data becomes and the more reliable
your hypotheses. So instead of keeping the 2 genes used by Manning and
supplemented with morphology, which will make the analysis even better than
Manning's, Lebatha chose to use one gene and morphology. I wonder if adding
the two genes + morphology will make a stronger case of support for
Manning's paper? Lebatha is obviously in support of morphology and still
relies heavily on it in her work.

The second part of my question concerns taxon sampling. In these analyses,
taxon sampling is extremely important. An incomplete taxon sampling can give
you false information. I am curious as to what happens when one adds the
taxa that Manning used to Lebatha's taxa? Will the results come out
differently? Can Lebatha still make the same conclusions?

Lebatha's work added important information to the group, but for me there
are still too many questions in which I would like for them to address
before accepting their work fully.  If a revision is necessary in the
future, one could make sub-generic classifications (e.g. Tribe) within
Ledebouria. Lebatha et al. did not make a revision, which means Manning's
revision in 2004 still stands as the most valid up to date, premature or

Berkeley, CA

On Tue, Oct 5, 2010 at 3:39 PM, Cody Howard <cchowar1027@yahoo.com> wrote:

> Lebatha's (2006) paper was published a couple of years after the Manning
> (2004)
> paper, obviously. It deals specifically with the data used in the '04
> analysis,
> and concludes that Manning's "revised taxonomy" is "premature".
> Nhu, have you seen this paper?
> --

More information about the pbs mailing list