Some novel concepts have invaded the arts: ideas of art as experiment, and associated understandings of art as research or research-creation conducted under laboratory conditions, and, beyond that, an idea of a hybrid art-science. It is as if art and science, two fields long regarded as distinct, and even polar opposites, are fusing into one. It is hard not to see politics at work in this. In the pursuit of academic standing, art allies itself with the field with the most prestige and funding, science. Conversely, the sciences, always in danger of social disconnection, try to plug themselves into the everyday world through art. I am not concerned here with politics, but I want to explore some of the ways in which art, science and experiment actually engage with one another. We can start by recognizing that ‘experiment’ has at least two meanings. In science, it often has the narrow and precise sense of determining some well defined parameter. I am not exactly sure what the acceleration due to gravity in Exeter is, so I set up the appropriate apparatus, do an experiment and measure it. Scientific experimentation is also often associated with hypothesis-testing. Theory X predicts the value of some parameter to be x, theory Y predicts y. We do an experiment, the value turns out to be y; theory X is falsified, theory Y confirmed. But at the other extreme from these examples, one can find another, much more open-ended and exploratory sense of ‘experiment’: experimentation as brute finding out. Try and see; what happens if . . .? In the history of physics, one could think about attempts to cool matter down closer and closer to absolute zero. It just turned out that some substances lose all electrical resistance and become superconducting when very cold. I cannot think of any artwork that conform to the narrow hypothesistesting sense of experiment. But much contemporary art belongs to the finding-out mode, and that is what I want to focus on here. The phrase ‘new media’ comes to mind. New-media art began and to some extent continues to be experimental in the sense of plunging into the unknown and finding out what can be achieved using, say, digital media instead of paints and brushes. Circuit-bending would be a striking example of this impulse, randomly reconfiguring electronic circuits to find out what sorts of sounds they can generate.1 And this is one way in which art and science/engineering are productively engaged. The artist Simon Penny (2008) has written about a historical genealogy of artist-inventors, in which artists have been central to important technological developments. This aspect of art as research-creation goes back at least to Leonardo da Vinci, but became more systematic with, say, the development of computer-art from the 1960s onwards (Reichardt 1971a, b). Though the connection to engineering is more evident here, science can also be important. Advanced microbiological techniques are crucial to bio-art, for example.2 The works of Evelina Domnitch and Dmitry Gelfand take off from esoteric physical effects like the implosion of bubbles induced by high-frequency sound (http://www.portablepalace.com). In works like these, science and engineering figure as a substrate for artworks that may (or may not) spark more developments in the former. This is the sense of art as a productive detour away from and possibly back to techno-science; techno-science as a surface of emergence and return, in Foucault’s (1972) sense, for art.3 But relations between art and science are often more interesting and oblique than simple mutual support. In their studies of science/art interdisciplinarity, Georgina Born, Andrew Barry and Gisa Weszkalnys (2008) documented important frictions around methods and goals between artists and engineers collaborating in