Letter to New Scientist from Martin Whittle, with a reply from Bob Forrest

Martin Whittle’s letter was published in New Scientist, 27 January 1983, page 260.
Bob Forrest’s reply was sent to New Scientist but not published.

Ley off

Although I would be one of the first to congratulate anyone who can demonstrate the reality of ley lines, the recent article by Paul Devereux and Robert Forrest (“Straight lines on an ancient landscape,” 23 December, p 822) contains ill-founded statistical arguments to whet the sceptical appetite of the most enthusiastic ley hunter.

The authors consider the Saintbury ley “in the context of a 10 km square”, and apply Behrend’s statistical formaul to the sites within this square in order to demonstrate its significance. However, this ley was originally discovered by scouring the whole of the British Isles and a statistical evaluation on the basis of a small square already containing the alignment is therefore not convincing and indeed erroneous.

The correct approach is of course to examine many 10 km squares and compare the total number of leys of a particular order with the statistical prediction.

On a related topic, can anyone reconcile the two conflicting scales on the ordinate of the first graph in the article on the Dragon project (“The talking stones”, by Don Robins, October 1982, p 166)? If the intensity and pulse rate are really so perfectly correlated as the graph would suggest then surely this fact in itself is highly significant and worthy of comment. Such poorly presented results do nothing to support the credibility (and hence the funding!) of an interesting field which deserves more critical attention.

Not So Easi-ley dismissed

Martin Whittle’s letter¹ about my joint article² with Paul Devereux merits some correction, I’m afraid.

Any statistical model must stick as closely as possible to the process it is designed to model, and the fact of the matter is that the ley hunter cannot “scour the whole of the British Isles” without being constrained by the confines of the various O.S. map sheets he uses. Since one can study the whole of Britain only as a series of ‘packages’, a realistic model of ley hunting must reflect this.

Statistically we assume that ley hunting is a game of chance alignment played with randomised map symbols within specified map boundaries. Each potential ley is thus related to a specific context, that context being the area of map sheet on which the alignment was first discovered, rather than the total area of the British Isles.

An effect like that proposed by Mr Whittle can arise as, for example, when someone discovers a line on a 1:50,000 sheet, but conducts his statistical analysis only over the part of the sheet containing the line (ie instead of over the full sheet). An artificially enhanced statistical significance can certainly then arise. We are fully aware of this phenonenon – it is affectionately known to us as the homing-in effect – and it is because of it that we place emphasis in our article on the context in which a ley is first discovered.

The only way a “scouring the whole country” argument can work is when the whole country is seen as a piecemeal collection of map sheets from which collection the ley bearing sheets are reported but the non ley bearing sheets ignored. This argument we discussed briefly in our article (p. 825), and though we believe it not to be true, it was on its account that we were careful to say that “the statistical wheel is still in spin”.

To finish on a lighter note – I don’t know the David Austin who did the cartoon on the letters page of the 27th January issue, and nor, so far as I know, does he know me, but the duffle coated character in that cartoon bears a striking resemblance to myself. So much so that both my co-author and my wife fell about laughing when they saw it. The caption, however, is totally without foundation. Having tramped some of the alleged leys at Lands End in the company of their discoverer John Michell, I felt nothing but the November cold.