DD - I'm not speaking from exacting experience here, but from a mix of similar situations and theory. It would seem to me that an approach from mortise & tenon joinery would be in order. When you use a flat joint with a solid extension piece, the solid piece is its own object and has the possibility to swell and shrink independent of the stock. The glue joint is not strong enough to exactly matcch the shrink/swell of the stock and the extension. Osmotic forces can be very high; the ancients used dry wooden wedges driven into small undercuts in stone to crack off construction size chunks and shapes. When the wedges were wetted, osmotic pressure exerted sufficient force to split the stone.
So, hollow the extension piece with a thin rimmed mortise. Cut the butt to a matching tenon. This should reduce the shrink/swell of the extension and provide sufficinet glue areaa to force the extension to match the butt's shrink/swell. Along the same vein of ideas, hollowing the butt should reduce its shrink/swell, also. If the mismatch in shrink/swell is severe, the mortise rim could crack. Therefore, I recommend that the mortise be made in the extension as a cracked extension seems less of an issue than a cracked stock. It may also be that extensions must be closely matched as to wood species and grain orientation.
Hope that gives something to consider.