A novel way of making computer memories, using bacteria FOR half a century, the       1     of progress in the computer industry has been to do more with less. Moore’s law famously observes that the number of transistors which can be crammed into a given space     2     every 18 months. The amount of data that can be stored has grown at a similar rate. Yet as     3     get smaller, making them gets harder and more expensive. On May 10th Paul Otellini, the boss of Intel, a big American chipmaker, put the price of a new chip factory at around $10 billion. Happily for those that lack Intel’s resources, there may be a cheaper option—namely to mimic Mother Nature, who has been building tiny     4    , in the form of living cells and their components, for billions of years, and has thus got rather good at it. A paper published in Small, a nanotechnology journal, sets out the latest example of the       5    . In it, a group of researchers led by Sarah Staniland at the University of Leeds, in Britain, describe using naturally occurring proteins to make arrays of tiny magnets, similar to those employed to store information in disk drives. The researchers took their     6     from Magnetospirillum magneticum, a bacterium that is sensitive to the Earth’s magnetic field thanks to the presence within its cells of flecks of magnetite, a form of iron oxide. Previous work has isolated the protein that makes these miniature compasses. Using genetic engineering, the team managed to persuade a different bacterium—Escherichia coli, a ubiquitous critter that is a workhorse of biotechnology—to     7     this protein in bulk. Next, they imprinted a block of gold with a microscopic chessboard pattern of chemicals. Half the squares contained anchoring points for the protein. The other half were left untreated as controls. They then dipped the gold into a solution containing the protein, allowing it to bind to the treated squares, and dunked the whole lot into a heated     8     of iron salts. After that, they examined the results with an electron microscope. Sure enough, groups of magnetite grains had materialised on the treated squares, shepherded into place by the bacterial protein. In principle, each of these magnetic domains could store the one or the zero of a bit of information, according to how it was polarised. Getting from there to a real computer memory would be a long road. For a start, the grains of magnetite are not strong enough magnets to make a useful memory, and the size of each domain is huge by modern computing     9    . But Dr Staniland reckons that, with enough tweaking, both of these objections could be dealt with. The     10     of this approach is that it might not be so capital-intensive as building a fab. Growing things does not need as much kit as making them. If the tweaking could be done, therefore, the result might give the word biotechnology a whole new meaning.