h2g2 The Hitchhiker's Guide to the Galaxy: Earth Edition

When building cladograms how many characters are needed? Also how does one choose which characters to include and which to leave out?
I guess that the real work of sorting the data set into cladograms is done by computer.

Good question! I'm not a professional in the field, unfortunately, so I don't know details such as how many characters are used, but I do know that they have pretty sophisticated computer programmes these days to sort the data, so you guess correctly.

Drat I was hoping to get into a discussion of it here
I was hoping that you would have been able to tell me why certain characters are chosen over others and why some are given higher weightings...
Something my friend who did a PhD on systematics/evolution of fishes couldn't really explain.

Hehe, well I'm no PhD - in fact I flunked Astrophysics at university so I don't even have a degree. However, I suspect that more weight is given to features that are unlikely to have arisen more than once in the manner in which they evolved. Feathers, for instance, carry enormous weight. Any two feathered animals *must* share a common ancestor which also possessed feathers, since such a complex structure is extremely unlikely to have evolved twice. Limblessness among vertebrates, however, is not nearly so important, since it is comparatively easy to lose legs if your lifestyle proves them unnecessary. Pointed, conical teeth I guess aren't too important - a host of fish-eaters have evolved teeth of this shape. Molars, however, are probably unlikely enough to be given more weight in cladistic analysis.

I'm just theorising here. What I've said above depends, obviously, on a certain amount of intuition, which is something that cladists abhor (see my note on preconceptions). So there's probably a more scientific way of approaching the issue.

I have just sent an email to Dr Henry Gee, author of Deep Time. He kindly reviewed my article before it was edited/approved and gave it a nod of approval. If he can provide an answer to your question I'll post it here.

Ta for that.
Your comment about astrophysics makes me curious about why you're writing about cladistics though

Okay - I got it wrong. Here's what Dr Gee had to say:

"...the answer to the question "Why are some characters deemed more important than others in cladistic analysis?" is that "they aren't!". In cladistics, characters are emphatically *not* weighted according to their perceived importance. Anyone who says otherwise has missed the point. If only all questions were so easy to answer."

So there you have it. I should have known this - I'm quite embarrassed that I didn't.

In answer to your question about Astrophysics, well, just you try studying it for 3 years! I used to be passionate about Astronomy, and was a mine of information about planets, stars, etc. But what I am NOT is a physicist, and endless incomprehensible lecture notes full of equations, formulae, proofs and derivations taught me that much at least. It was a hideously dull and awful subject.

Dinosaurs were, once upon a time, a greater passion for me, and having left university I renewed my acquaintance with them. But I quickly became interested in what came before and after dinosaurs (but mainly before), and gradually built up an extensive surface knowledge of the entire history of life. Cladistics is becoming essential in understanding that history, so I've taken quite an interest in that, too. I'm in the process of writing a book about the history of life, but it's just for my own fun, really. It's full of cladograms and pretty pictures of prehistoric beasties. Anyway I'm digressing now.

Fair enough then I didn't know that. Thanks for your answer.

Having studied electronics with all it's formulae and derevations I know what you mean.
So that's where the cladistics stuff comes in, it all makes sense now

Hard core cladists would say that you should not give any weights to different characters, for fear of introducing some sort of bias to the data. Most people working with cladistics though, say that you can bring in weights from outside sources. For example, studies on a different closely related group, etc. But these weights need to be very well defended, just in case a hard core cladist argues.

One of the biggest dissagreements in cladistics right now is how much you should weight different characters. Some methods used are very circular. For example, some weight the most the characters that _follow the tree best_ , which just makes the first tree seem less ambiguous. The tree is not actually less ambiguous, though (all of the data is still the same!).

For a while, building the tree was done by hand, but now it is done by computer. Three programs that are usually used are MacClade, PAUP*, and Mesquite. Mesquite is new and also free (because it is still being debugged)-www.mesquiteproject.org