--- In

primenumbers@yahoogroups.com,

"Mike Oakes" <mikeoakes2@...> wrote:

> you are really just finding the probable-prime

> norms of the Gaussian Lucas sequence

> V(1,-3+4*I,n)

Thanks for tidying up, Mike. Note that

Q = -3 + 4*I = (1 + 2*I)^2

is the square of a reduced parameter, q,

inside your small quadrant, which is where

I found these primes in the first place.

The general formula for squaring q by "renaming" is

V(p,q,2*n) = V(p^2-2*q,q^2,n)

That, in turn, is merely the m = 2 case of

V(p,q,m*n) = V(V(p,q,m),q^m,n) ... [1]

which is one of my favourite functional equations.

This bothered Bouk de Water when I showed him

a non-trivial example. He asked me to prove,

at the very outset, that

V(V(p,q,m),q^m,n) = V(V(p,q,n),q^n,m) ... [2]

else he would never believe [1]. (As I recall, he was not

best pleased when I derived [2] as a corollary of [1].)

PS: I noted your neat pair of gigantic probable Gaussian primes

V(1,2+2*I,19979)

U(2+I,1-I,19979)

Not as remarkable as the de Water and Noe probable prime pair

V(1,-1,148091)

U(1,-1,148091)

but yours is still a notable collision.

David