Sheep BSE research is hampered by the lack of a highly specific diagnostic test suitable for field use. Conventional strain typing, pioneered by Moira Bruce at the BBSRC's Institute for Animal Health in Edinburgh, discrimates prion strains well, but involves inoculating legions of mice with sheep extracts, and analysing incubation times and lesion profiles in the brains of mice that succumb. Experiments take up to two years, and each costs £30,000.

John Collinge, head of the Prion Disease Group at Imperial College School of Medicine in London, and a member of SEAC, thinks he has a faster and cheaper method, but complains that the UK agriculture ministry has so far failed to commit itself to the refinement and scaling-up of the technique needed if large numbers of sheep are to be tested.

The technique distinguishes prion strains on the basis of Western blot patterns of conformation and glycosylation of prion proteins, and Collinge says pilot studies demonstrate its feasibility in screening for BSE in sheep.

But many scientists, including Stan Prusiner, who received the Nobel prize last year for his work on prions, remain sceptical about the theoretical basis of the method. Others question its validity in sheep where interpretation of patterns is complicated by the large variety of prion strains and host prion genotypes. The electrophoretic profiles of the many sheep prion genotypes needed to establish a baseline for comparison has also not yet been established.

Such uncertainties have divided SEAC members over the use of the technique. “It is not coming out clean in a way that makes it easy for us on SEAC to make decisions,” says one SEAC scientist.

These arguments miss the point, says Collinge, who argues that, even if the theoretical basis for his technique is not clear, it is an empirical marker that works. He admits the technique needs refinement: “But that's MAFF's job not mine.”

Opinion on SEAC now seems to be swinging in Collinge's favour. Pattison says: “Glycotyping is likely to be one of the useful techniques in surveying the strain characteristics of the agents recoverable from a large(ish) number of animals. I think we've reached the point where accumulation of further data is the way to inform ourselves about the usefulness of the technique rather than further theorizing.”

One compromise strategy that seems to be emerging from SEAC discussions is that, in the absence of a rapid specific test, a targeted search could use Collinge's technique to screen several thousand scrapie cases and identify suspicious signatures which could then be double-checked using the traditional strain-typing technique.

Some of the resistance to Collinge's technique seems, however, to be linked to fears that a preliminary diagnosis of BSE in sheep using this method may be interpreted by the public as firm evidence that BSE has passed to sheep. “The consequences of identifying a first case of BSE in sheep would be catastrophic so we need to be really sure [about the reliability of the identification],” says one member of the European Commission's independent Scientific Steering Committee.