Are there two length scales at phase transitions?

The development of high resolution X-ray scattering techniques, particularly using synchrotron sources enables very detailed measurements of the properties of the fluctuations associated with continuous phase transitions. Surprisingly, the results of almost all high resolution studies of structural and magnetic phase transitions in crystals show that two length scales are needed to describe the experimental results in striking contradiction to the basic hypothesis of the scaling theory of continuous phase transitions. The measurements further suggest that the longer length scale arises from a region of the crystal close (50 mu m) to the surface. Possible explanations for the two length scale behaviour will be discussed including the effects of surface damage, the effects of surface strains, or intrinsic behaviour and some experiments outlined which might distinguish between these different possibilities.