I thought I understood the guidelines for determining scientific authorship: the individual making the greatest intellectual contribution is the lead author, followed sequentially by those making progressively lesser contributions. In addition, the final-author slot is sometimes reserved for a lab head or project initiator, who may have made little direct contribution to the paper but deserves some vague honour nonetheless.

But now I am confused. A collaborator of mine at the University of Cambridge asked to be moved from second to last position on a four-authored paper. When I asked why, he said the British Research Assessment Exercise (RAE), which determines departmental rankings and government funding, gives greater credit to the final than even the second author on a multi-author paper. My confusion deepened when two other colleagues — both Americans — had a lively disagreement about who would be last author on a paper with seven authors.

The 'communicating' (or 'corresponding') author is often, but not always, the lead author. If he or she is not the lead, is some special significance attached to this? Does it count for something on the RAE? Some disciplines have evolved their own idiosyncratic rules. I have also noted that another common convention is to list authors alphabetically, but does the RAE know if Williams made a lesser or greater contribution than Anderson?

Is there a set of coherent authorship rules written down somewhere that, in my 20-year research career, I have managed to miss? If not, then perhaps there should be.

Please note that I am the first, last and communicating author on this Correspondence.