THE disadvantage of the ‘solid-image’ microscope described by Gregory and Donaldson1, as they point out, is that objects are seen through a relatively intense haze of background light, rather as if they were immersed in milk. There is a fairly straightforward way around this difficulty, though it entails a loss of some of the elegant simplicity of Gregory and Donaldson's instrument. During 1948–49 I devised and built at King's College a 3-dimensional scanning microscope on the same basic principle2, which was described briefly in a report on university research in physics in 19493 and exhibited at the College conversazione in that year. The slide carrier was vibrated in depth between 10 and 20 times a second with a saw-tooth motion, by means of a linear electromagnetic servo. The object, however, was scanned laterally by a flying spot on a cathode-ray tube, and the signal detected by a photomultiplier, exactly as in the 2-dimensional ‘flying spot microscope’ later (and apparently independently) described by Roberts and Young4 of University College, London. The three deflecting voltages and the photomultiplier signal were supplied to a 3-dimensional protective cathode-ray tube display of the type devised for use in radar and electronic computing5, so that a ‘solid’ representation of the object appeared on the screen and could be rotated at will by calibrated controls which enabled angles to be read directly.