[seqfan] Re: Conway's subprime Fibonacci sequences

[Hans Havermann]

Franklin T. Adams-Watters:

> So submit the cycle lengths ordered by the first starting pair that  
> generates that cycle (in anti-diagonal order).

Wouter Meeussen:

> In selecting the starting points for each cycle, I followed Franklin  
> T. Adams-Watters' suggestion of ordering the pairs in anti-diagonal  
> order.

I have to betray my ignorance of not knowing what 'anti-diagonal  
order' meant exactly. In trying to figure it out (for positive  
integers), I conjectured that it is either {{1,1}, {1,2}, {2,2},  
{1,3}, {2,3}, {3,3}, {1,4}, ...} or {{1,1}, {2,1}, {2,2}, {3,1},  
{3,2}, {3,3}, {4,1}, ...}. But Wouter's A214892-A214896 appears to  
employ two from the former ({10,18}, {23,162}} and three from the  
latter ({4,1}, {18,5}, {382,127}). So I'm still confused.