What pressure cap are you running? Measure the distance down to the cap sealing surface in the rad to make sure it is parallel to the top, I've seen a couple where this surface was warped and caused the cap to release into the overflow and not completely close after the initial opening. I found it by putting a rubber gasket cut out of sheet rubber under the cap which effectively raised the pressure where it would release and added a softer seal.
The cap itself needs at least 0.050" compression on the rubber seal ring or it will release early, any compression grooving caused by set on this ring reduces the pressure it will release at. I sounds like you have the rad puke it's guts into the overflow and then suck it back in and the cap is not closing to keep the system sealed. If the seal area in the rad is not parallel it will open and not close properly and you end up with a pulsing pressure release which explains your temp swings.
When it happened to me it was a bugger to diagnose since a new cap didn't solve the problem even though running with the hood off the car allowed a passenger to witness the reservoir alternate between full and empty. Of course this only happened at race loads and rpm. I recommend at least a 16 pounder on your system, I ran a 24 pound cap to maintain good pressure in the system and prevent the dump to reservoir that we were having (a rotary engine turning at 8500 can generate 20 pounds easily just in coolant flow pressure). I would rather overpressure the system than have it puke during a yellow flag and then be out of coolant when the green drops again. Race rpm subjects the rad and coolant system to very high pressures and rads when new can handle 50 psi and higher before they blow out, usually it's the rad hose that goes first anyway. Seal it up tight and then try it under load and see if it solves your problem.
Hope this helps.