On Paul McKenna's appalling Paranormal World on television the other week, they showed a chap in Nevada who has been doing research into precognition. Subjects are wired up to a device that gauges emotional reaction by measuring electrical activity in the skin, then they are shown randomly selected photographs, some of which are idyllic and pastoral and the others are so filthy that they had to blot out the rude bits on our screens. Anyway, what the results claim to show is that subjects reacted to the emotionally charged pictures even before they had been selected. The experiments were cited as evidence of precognition.

This, of course, merely adds a human dimension to the celebrated work of the parapsychologist R L Morris who, in 1967, showed that if you put three goldfish in a tank, then use a random number generator to decide which one to fish out, there is a tendency for the about-to-be-caught goldfish to swim about a bit more frantically before its number comes up.

We mention all this only because of some work we have been doing in our newspaper database. Since the murder of Gianni Versace, we have been monitoring occurrences in the press of the word "frock". We felt that it had come into relative prominence since his death, and were wondering whether this was purely temporary or the start of a new linguistic vogue. Since we needed a base level for purposes of comparison, we monitored the occurrences of "frock" over each of the past few months. here are the monthly totals since December:

Dec Jan Feb Mar Apr May Jun

Frock 124 126 104 102 90 135 167

Throughout 1996, frocks had been running at about 120 a month, and as these figures show, that rate was maintained in January. But the next three months showed a marked decline, followed by sudden and dramatic rises in May and June. The June 1997 figure is, in fact, the highest ever recorded for a calender month. And if we were to plot these figures on a graph, it would look just like the read-out from the device used on the Paul McKenna show. It is difficult to avoid the conclusion that precognition is an exact science - with hindsight.