While the world population is steadily growing, cultivable land is disappearing due to urban expansion and desertification. The challenge of the future will be to produce more edible crops, and one way to better use Earth's resources for this purpose is to shift from our current strategy (cutting down forests to free up farmland) to a new strategy (creating new farmland on the vast ocean surface through floating greenhouses. This presents some engineering challenges, but also has a number of advantages. For example, farming on water means that irrigation is a simple matter of desalinating the water under the plants and routing it into their growth medium. And marine farms would have mobility, potentially migrating across the year to stay in season for their crop. The aim of my project is to create a working miniature prototype of a marine greenhouse, which must house lettuce and radish plants, float on saltwater without capsizing from the effects of simulated waves, and passively desalinate the water and irrigate the plants. This greenhouse is intended to produce a crop of lettuce and radishes suitable for human consumption. The design of the greenhouse is a cube-shaped container that holds the plants, with air-filled outrigger floats extending from either side and a solar still on the roof. The walls of the greenhouse are made of high-density PET on a framework of galvanized steel wire. Vinyl tubes pick up saltwater from under the greenhouse, transport it to the roof through capillary action and evaporation-condensation, and drop it into the saltwater reservoirs of the solar still. The water evaporates, condenses on the ceiling of the still and drops into the freshwater reservoir; from there it flows through vinyl tubes into the peat. Excess water drains from the peat through newspaper and a plastic screen into a third reservoir at the base of the cube, where it falls through a check valve back into the pool of saltwater. The greenhouse is tested in an aquarium full of simulated seawater, I intended to simulate constant waves by placing the aquarium on a see-saw device connected to a motor, causing the aquarium to rock slowly back and forth. I had a number of unforeseen difficulties with this project, mainly pertaining to materials: for example, most adhesives designed to keep out water are not designed to adhere to non-porous, flexible surfaces such as the plastic I used. For this reason the project has been plagued by leaks, the patching of which has further delayed testing. The greenhouse is still undergoing tests, which so far have revealed that a) it floats and b) water evaporates and condenses in the solar still. The aquarium is not yet rigged for wave simulation, and the plants have been sown but have not yet germinated. (Lack of germination is not a result, as the normal gestation period for these seeds - as determined in separate germination trials - has not yet passed.)