It seems to be a question which comes up regularly.
The mechanism by which it occurs is diffusion: where you have two similar metals in intimate contact, with nothing between them, atoms from one can diffuse into the other, and vice versa, across the interface. Essentially they wander about and forget where they are supposed to be, until enough of them have migrated that the boundary between the the bullet and case neck is blurred, and you have a joint.
Diffusion bonding, the usual name, has some industrial applications too, especially when you want to weld really delicate or thin materials where any heating, such as is used in conventional welding, would be impracticable. I remember doing some work on just such an application many years ago, involving a miniature heat exchanger made by diffusion bonding of very thin corrugated sheets together to make a honeycomb.
If there's something between the two surfaces, it can prevent the wandering atoms from crossing the line. Something like wax for example would have such an effect. Or, as MD and I have independently found, not bothering to tumble cases and relying on grottiness seems to work too.
Thanks for that explanation, best one I've heard yet. I'd always assumed it was a galvanic thing, but diffusion makes perfect sense. That's something that also takes place between copper and tin/lead alloys when electronic components are soldered onto circuit boards.
I've been using Imperial Application Media with dry neck lube for a while with the thought in mind that whatever causes cold weld, maybe the neck lube would help interfere with the reaction. I'm not sure any of the ammo I've ever reloaded developed cold weld in the first place but prevention is a good thing, I think.
