 Genedrive systems could solve ecological problems by altering the traits of wild organisms. Here we describe how we can build daisy chain drive systems to alter local populations. One form of gene drive uses CRISPR, which is a pair of molecular scissors that can be programmed with a guide molecule to precisely cut any DNA sequence and replace it with an edited version. To make a drive system to spread that change, instead program the new DNA sequence to do genome editing on its own. Introduce a single piece of DNA that encodes both the altered gene and the CRISPR components into the reproductive cells of an organism. CRISPR will cut the original sequence, causing the drive system DNA to be copied in its place. Once one copy is stably inserted, it will cut and replace the other copy. When this organism finds a wild mate, the offspring are guaranteed to inherit a copy of the drive system. And in those offspring, editing happens again. The original version is replaced with a new one, which ensures that the next generation will also inherit the change. Down through the generations, replacement will happen again and again and again. We first described how CRISPR could do this in 2014, before we ran any experiments, because technologies like this should never be developed behind closed doors. Almost two years later, it's been proven in yeast, fruit flies, and two species of mosquitoes. If used wisely, gene drive systems could benefit both human health and the environment. It's a way to solve ecological problems with biology, not bulldozers. But CRISPR-based gene drives are global. They can spread from a single organism to every population of that species in the world. How do we test them in the field to see if they work, and whether there are side effects, when introduction anywhere likely amounts to introduction everywhere? What if one country decides it wants to prevent Lyme disease, or eliminate malaria, but its neighbors don't agree? What we need is a local drive system that can only spread for a fixed number of generations. To do that, we need to teach DNA to count. Here in the Sculpting Evolution Group at the MIT Media Lab, we've devised a new kind of drive system, built like a daisy chain. Instead of a single piece of DNA that has everything it needs to copy itself, a daisy drive separates the components and scatters them around the organism's genome. None of them can drive on its own, but they're connected in a linear daisy chain. Element C helps B to drive, and B helps A to drive. The resulting daisy drive system will spread the DNA payload carried by A through the local population. Because there is nothing helping C to drive, it will never increase. When it's around, it can help B to drive, so B will steadily increase. And since A can drive whenever B is present, it increases the fastest of all. Of course, in reality, our changes may be good for us, but they usually harm the organism, so natural selection will drag each of the elements back to Earth. That means daisy drives are very like multi-stage rockets. Each element in the daisy chain is a genetic booster that propels the payload element, making it increasingly common in the local population. But as daisy elements are progressively lost, the payload runs out of fuel and eventually disappears. That means changes made by daisy drives are temporary and confined to the local area. To make the changes spread further and faster, simply add more elements to the daisy chain. Releasing one organism with a five-element daisy drive is like releasing hundreds of organisms with just the payload. But we do need to be careful, because a particular rare DNA rearrangement could turn a linear daisy chain into a daisy necklace that could spread globally. That can only happen if the elements have very similar DNA sequences. To avoid this, we design CRISPR components to be as different from one another as possible. Using these in daisy drives should prevent any such rearrangements. But should isn't good enough. We need a way to safely test drive systems in the laboratory that can predict how they will evolve in the wild. That's why we're studying them in huge populations of nematode worms, which evolve incredibly swiftly and that we can keep safely confined in the lab. Only if they behave as predicted should we consider using drive systems in the wild. In short, gene drive systems may let us solve ecological problems with biology rather than bulldozers. By limiting changes to local populations, we hope daisy drives will give that power to local communities who best know whether we should, or should not, use them.