> That business of passing around the entire node population, as a means to
> get around follow-sibling) is cute.  However, the outer test won't work for
> # cols = 1; it should probably be adjusted to look for mod cols = 0


Dear all,

Let me assure you that I will *never* approach solving a task like

"Arrange a number of nodes in a 1-column table".

Please, do not relate me to the quoted text above.

Aron, please note that my solution is also terribly wrong in the case of N =
0.

And for all cases of N < 0.


Cheers,
Dimitre.