Celsis tests the water with a faster bug-detection kit

SPOTTING contaminated drinking water is a bit like closing a barn door after the horses have bolted. The time it takes water to flow from reservoir intake to kitchen tap is typically about a day, but tests for microbes, using techniques unchanged for more than a century, can run for up to four days. Celsis International, a Cambridge-based diagnostics company, hopes to reduce the time-lag to just 24 hours.

The release of its new Colorcount for Coliforms test also demonstrates the company's new commitment to cutting the time taken in bringing products to market. The test was originally scheduled to appear in early 1996 but was rushed out after a boardroom shake-up.

Last April, Celsis ousted its chief executive, Dr Tony Martin, after its share price tumbled almost 30 per cent in the year following its flotation, which raised £12.4m and valued the company at £60m. Dr Martin's replacement, Arthur Holden, vowed that in future Celsis would be driven more by sales than research.

In his previous job as marketing vice-president of the renal division of Baxter International, a biomedical company, Mr Holden helped to boost sales from $200m (£140m) to $1.2bn. Colorcount is Mr Holden's first attempt to reassure investors that he can do the same at Celsis.

"What Arthur has been able to do is prioritise our focus," said Mark Clement, the company's finance director.

The new diagnostic kit is designed to solve a chronic problem for water company and government inspectors. E coli and coliforms, including salmonella, are key markers for organic pollution from faeces and dead animals. But the tests are so slow that warnings for customers to boil their water usually arrive after people have been exposed and become ill.

The traditional method of testing for microbes has changed little since it was developed in 1878. Water samples are placed on a glass dish coated with agar, a nutrient, and incubated for days until colonies of bugs grow big enough to be spotted by the naked eye. Trained technicians then have to identify and count the tiny spots on the dishes. The whole process is slow, labour intensive and prone to human error.

Rather than wait for the cultures to grow that large, Celsis looked for ways to make them visible earlier. Biochemist Dr Peter Grant and enzymologist Dr Nick Foote came up with two reagents each composed of a carbohydrate and a dye - either pink or blue. The microbes eat the reagents, breaking them in two to get at the carbohydrates and excreting the dyes.

The tests are to be marketed in a patented kit that allows laboratory technicians to process dozens of sub-samples at once. The company claims that dividing samples up in separate dimples on plastic trays makes the process of counting the concentration of microbes in the source water easier and more accurate.

Colorcount is derived from Celsis's first line of products, bioluminescence tests that reveal whether there are any organic substances at all in a sample. Those tests require special monitoring equipment to detect tiny flashes of light given off by chemical reactions between a different type of reagent, derived from fireflies, and a protein called ATP that is common to all life that uses oxygen.

Celsis estimates there is a £200m-a-year market in the developed world for tests that identify contaminated drinking water.

Mr Clement said the company is close to signing distribution agreements for Colorcount with several companies. It is now testing the system for use on salt water, where it could be useful in determining whether beaches are safe for swimming.

Success would go a long way towards turning the company's profit-and- loss account around. Celsis lost £2.2m in the six months before Mr Holden took over. Panmure Gordon, the broking house, has predicted it will break even by 1997 and generate pre-tax profits as high as £30m a year by the turn of the century.