In Anthony Johnson's excellent book Solving Stonehenge he discusses the position of Stone 56:
There is no certainty that this stone, which had been leaning for centuries had been put back where the Neolithic surveyors intended it to be, but to have the confidence to assert this needed proof beyond that indicated by the computer model alone. The account of the work carried out before Gowland set the stone upright in 1901 needed to be re-examined, and the surveys made by John Wood (1740), the Ordnance Survey (1867) and Petrie (1877) had to be compared with Gowland’s careful drawing [122a] of the stonehole before the engineering work began. Rather than trust survey dimensions that may always have been measured relative to the fallen, broken and displaced stones, I struck two circles from the centre of Stonehenge on each of the historic maps: the first was drawn to touch the innermost part of the leaning trilithon, the second the outermost. The object was to compare its various plotted positions in relation to the undisturbed eastern trilithon (Stones 51- 52 and 152). These circles showed a good agreement on the three pre-1901 maps, hut all mismatched the modern survey. There was now little doubt that the standing stone of the Great Trilithon had been restored in the wrong place. But there was more to this exercise than identifying discrepancies in resetting the stone; if a displacement could he confirmed, then it demonstrated that the prehistoric survey methods and geometric principles which had been used to mark out all the other stones in the sarsen arrays had also been used to fix the original position of the Great Trilithon.
In 1902 Gowland published the plans and section drawings of his restoration work. Amongst his drawings are records of the deposits that lay against the south west face of the stone where three successive blocks of chalk were cut away to reach the bottom of the Neolithic foundation hole [122a]. As each section was removed it was replaced with concrete to form a wall against which the outer face of the stone was to be subsequently set. The section drawings show that, deep below the turf, the leverage exerted by the massive stone as it tilted into the centre of the monument had pushed its base outwards, causing it to ‘bite’ into the chalk on the (southwest) side of its foundation cut. When the stone was re-erected it was winched upright not to where it had been originally set, but against a vertical face representing the maximum displacement of its toe [122b]. This is perfectly understandable: Gowland’s task was to correct the tilt; the relatively small dislocation of the 45-ton trilithon was not important in the scheme. The result is that today the stone stands perhaps as much as 60 cm southwest of the position in which it was first erected, and is rotated approximately nine degrees anticlockwise from the perpendicular of the axis. These are small but very significant details.
The book really is very good, the first two thirds of it a great bringing together of the history of the stones, the latter part his impressive demonstration of how the stones may have been set out. There is more on his blog as well - http://sarsen56.wordpress.com/.
But I think he errs with his analysis of the position of 56. My arguments are set out at exhaustive length elsewhere on this blog, for instance at http://www.sarsen.org/2012/07/stone-56-proof-its-position-is-correct.html, I will leave this post as his argument that I am wrong.