Note also that, for efficiency, different features are described at different levels of the tree. Thus, whether the ovary is superior or inferior is often most efficiently coded at family level, whether the leaves are opposite or alternate may best be coded at genus level, what colour the flowers are at species level, and how long the seeds are may be best at specimen level. Ultimately, you're right, all coded features 'inherit' to the ends of the tree (the specimens) but it would be inefficient to code, or to store, the data there.

This is exactly the answer I needed -- thanks. So yes, for obvious practical reasons, described characters can apply directly to taxonomic concepts. But I guess the one lingering point I have is that the descriptive data is not *necessarily* used only in the context of taxonomy, and as such perhaps should not be confined to a taxonomic context by whatever moniker is finally selected. I think that a taxon could be thought of as a "Bioliogical Object", so that might be a more encompassing term. But then again, if you don't want to confine it to living objects, then the "Biological" could be discarded.

I quite like SDBO. Is there a courier service from Lisbon to Hawaii?

Nah -- just think warm thoughts next month in Portugal (wish I could attend, but alas, cannot).

Aloha, Rich