duplicate hunting, pt. 7

Ray Chandler rayjchandler at sbcglobal.net
Sat Apr 28 17:25:27 CEST 2007

> Possible duplicates:
> A075643 and A076074
> 	-Andrew Plewe-
I don't have a proof, but have convinced myself that these two sequences continue to agree.  The first term of a group
is essentially the least number not previously included.  The previous groups have sieved out multiples of earlier
numbers.  After 2*prime(k)+1 groups have been selected, there are no non-primes available less than prime(k)^2.  It
takes one more group for A076074 since we don't have complete freedom to choose the last two terms of a group, but the
same argument holds.  The least number available continues to be prime, a(n) = prime(n-1) for n>2.  These sequences turn
out to be the non-composites (A008578) with the first two terms transposed.  

Ray Chandler

More information about the SeqFan mailing list