They don't have access to a strong magnetic field (except the earth). When you assemble ferromagnetic atoms together one at a time they tend to line up with ea

They don't have access to a strong magnetic field (except the earth). When you assemble ferromagnetic atoms together one at a time they tend to line up with each other, even if the applied field is weak to nonexistent. This is why the orientation of the atoms in an iron bar aren't random from one atom to the next: instead they form "domains" where all the atoms point in one direction. If you are building a tiny enough iron bar it's probably pretty easy to have the whole thing come out as one domain. And bacteria aren't very big, so the bars of magnetite that they're building inside themselves are pretty small. Are you suggesting they are aligning the magnetic moments of the iron atom by atom? Is that really possible? Absolutely. This is how all crystal growth works. One atom at a time. You may not be used to thinking about chemistry in this way, but that is what chemistry is. Of course, you shouldn't think of the bacteria the way you think of human model-makers, consciously assembling tinkertoys. That kind of tedious fiddling is for people like us, who are long on understanding but short on design time. The well-designed bacterium uses chemistry instead: Metaphorically, you throw properly shaped tinkertoys into a bucket and shake it around and they assemble themselves. ;) In fact, if you take a liquid and chill it below the freezing point, what usually happens is that crystals start to form in more than one place, and then those crystals expand atom by atom until they bump into each other and fuse. For example, this is the reason why there is more than one magnetic domain in the typical iron bar. If you grew your iron bar under perfect temperature-controlled conditions from a single seed crystal, you might be able to grow an entire bar with one domain. That is what the bacteria are doing, but they get to cheat because the bars they are growing are so small, so they get to quit while they are ahead. :) If you try to grow a meter-long, centimeter-thick bar like this, you'll find that you've got trouble with temperature control, trouble with impurities, trouble with rust (as I said), and that because you're operating at a temperature above absolute zero there's always some randomness going on, trying to scramble your work. Having said that, it might be possible for humans to do this. We do similar things with silicon every day. The chips inside your computer were cut from a single crystal of silicon 25 cm in diameter and perhaps 1 meter long: http://en.wikipedia.org/wiki/Boule_(crystal) http://en.wikipedia.org/wiki/Boule_(crystal) http://en.wikipedia.org/wiki/Single_crystal http://en.wikipedia.org/wiki/Single_crystal But understanding the detailed atomic behavior of ferromagnetic materials is actually quite a hard problem. Lots of difficult thermodynamics. Amusing chaotic systems problems. I basically hit the wall in my graduate course on the subject. ;)