Just like regular fusion. The "cold" in cold fusion is a desciptor indicating a fusion reaction that can be sustained at reasonable temperatures, and thus commercially useful, rather than the many, many thousands of degrees centigrade required for natural fusion reactions (a la the center of a star).

Cold Fusion supposedly works by creating pressure between layers of molecules in a battery-like device. When electricity is passed between two electrodes submersed in deuterium, pressure builds up between the molecules causing them to fuse and release heat. The scientists who claimed to have created cold fusion said that more heat energy was released by their aperture then they put in. However, their results did not live up to scrutiny by other scientists.

Well, since it doesn't, really, this is sort of moot. But this is the best I heard when it was the rage: The hydrogens and heavy hydrogens would be introduced into a crystalline lattice, and somehow the electrostatic forces in the lattice helped overcome the repulsion of the protons in the hydrogen nuclei, causing them to fuse into helium, and generating heat. The whole flap started because the experimenters made some mistakes, and thought they saw more heat coming out of the experiment than they put in. They assumed it was fusion (since that's what they were looking for) and published. Fortunately, science works, and since no one could do what they claimed to have done, including them, it was recognized as a mistake in expermentation. Right now, the only places where fusion happens on a large scale is in fusion reactors, which contain the hot plasma with magnetic fields. The reaction happens, but since it is so hot, it can't contact any material, and is isolated with magnetic fields. They haven't "broken even" yet, which is to say, more energy goes in than comes out. Fusion, or thermonuclear, bombs also have some fusion going on, but they aren't practical for energy production for constructive uses, just destructive. The natural fusion reactors are the stars, our sun one of them, and they do a really good job of converting the gravitational potential of their mass into fusion energy. The sun looses much mass every second, not just in the mass ejected as solar wind, but from conversion of mass to energy, E=mc^2.

My answer pertains more to why cold fusion, would be improbable, near be it impossible to obtain. That simply put, the only fusion we have been able to create has been a bomb because the only force we have been able to use to induce fusion is that of fission. So, if we must use fission to create fusion, how could it ever be cold?

Let's start with fusion. The big problem is that nuclei really don't want anything to do with each other. There is a certain minimum amount of energy required to make them stick, no matter where that energy comes from. Energy comes in many forms: thermal, kinetic, electromagnetic, etc. The easiest way to make fusion work is a combination of heat and magnetic fields. Cold fusion, then, would be a method that uses no heat. In theory, the required energy could be produced solely using magnetic fields, or by some combination of sources other than heat. The phenomenon known as sonoluminescence, while NOT cold fusion, is a step in the right direction. Serious progress has been made in recent years regarding laser-induced fusion. The use of particle accelerators has also been explored. One interesting method involves spiking the deuterium pellet with small amounts of antimatter, producing the necessary energy on the spot. Currently, no method produces more energy than it consumes. Some methods need technology that doesn't exist yet, while others are too difficult to get energy from. In short, usable fusion, hot or cold, is quite a ways off yet. Even if it wasn't, the energy released in a fusion reaction ensures that it won't stay cold for long... Consider Sandia's Z-Machine... http://www.sandia.gov/media/z290.htm http://www.sandia.gov/media/NewsRel/NR1999/thermo.htm Sonoluminescence (Lawrence Livermore) http://www-phys.llnl.gov/N_Div/sonolum/