Taxacom: Species described based on a single collection

[Laurent.Gautier at ville-ge.ch]

Dear all,

We often hear or read: "A huge amount of Amazon plant species are known 
only from the type" or "Most XIXth century species descriptions are based 
on a single specimen". But has anybody gone any further in metrics?

I am looking for statistics (published or not) on the percentage of 
species descriptions based on a single specimen (or collection), and 
possibly its variation across time or geographical area considered.

Does that ring a bell to anybody? My main focus is on flowering plants, 
but all living organisms welcome!

Thanks in advance,

Laurent Gautier
Conservatoire et Jardin botaniques, Geneva, Switzerland