No safety in numbers

Bernard Dixon

It is, I suppose, some part of MY brief in this column to provide relief from the magazine's political content and flavour. That's a difficult task these days, when science itself — and scientists — are being kicked around like so many political footballs. But I take the point, especially in election week,. So instead of doing as I am tempted, and grousing about all those scientific and quasi-scientific issues (population, energy, environment,

and the perilous spread of nuclear technology) which should have been at the centre of the election Campaign but have not been, let me tell you a little story. What I shall do is to cast such doubt on the entire scientific enterprise as to suggest that it could come crashing down over night as a tissue of lies. My story

Concerns references yes, references, those tiresome details listed in profusion at the bottom of every scientific paper published in the

learned journals, which cite the precise whereabouts of earlier Papers quoted and used by the current author. If there is one thing scientists and journal editors are fanatical about, it is getting the references right. Many journals, not trusting either authors or Printers, even have a multi-stage

screening system, staffed by obsessional ladies who check every reference umpteen times and der

!Ye psychological kicks in correcting error.

Enough said. To get your reference wrong is a terrible sin, because it means that untruth has crept into the literature." The edifice of science rests on mutual trust and strict accuracy. In a very real sense, therefore, it is threatened by the tiniest footnote fault. And, of Course, there's also the bother and fury when you come across an

erroneous one and want to track it down.

That's what happened to Dr Duncan Blanchard, a scientist at the State University of New York, who has just thrown a bombshell into the scientific arena by announcing publicly his discovery of some very serious mistakes of this sort indeed. Writing in the current issue of the American journal Science (I will try to give You the reference later), he provides clear evidence of a veritable epidemic of referencing errors. What happened was this. Com13°sing his own latest addition to the world's scientific literature, Dr Blanchard wanted to quote a paper he had read in the Swedish geoPhysical journal Tellus in 1970. He couldn't find the reference in his card index, so he turned instead to a Copy of the Journal of Geophysical Research which he had just been reading and which carried several Papers each citing the one he was after.

Turning up one of those papers, he found the reference he wanted to the Tellus paper. It was: vol 21, 1970, PP 451-461. Just to be sure, he had a look at another of the papers. This time, the reference was to vol 22, So he tried a third author, who aISO gave vol 22 but a quite difference page reference.

Dr Blanchard began to feel deep'Y uneasy at this point. But then he noticed that a fourth paper was by

the roan who had written the original Tellus article. As we all K.now, authors are most punctilious about references to their own labours, so Dr Blanchard assumed that this time he would discover the correct answer. What he actually found was totally different t° anY of the other versions. _ In despair, poor Dr Blanchard ie.ventually had to trudge over to his 11'rarY to learn the truth of the

matter. With the correct answer revealed at last, he now realised that of his final haul of six authors citing the original paper, four (all different, and including the author himself) were wrong, and only two were correct. A sad, sad tale, which must be sending shudders through every researcher, editor, librarian, and abstractor who reads it.

Dr Blanchard, by the way, gives just one reference at the bottom of his own Science contribution. It is to Thurber's Father of Our Time, and to the moral of one of its stories: "There's no safety in numbers, or in anything else." What I cannot give you with absolute confidence is a reference to Dr Blanchard's own unhappy report. His contribution has no page number on it.

Dr Bernard Dixon, who writes fortnightly in The Spectator, is editor of New Scientist