For Roy and Neva Gordon, 10 August1937 was a bad day. Six days earlier, they had been in a three-car pile-up in Marion county, Indiana. Now they were in court. Neva faced charges of being drunk, Roy of drunk driving. The couple had picked the wrongtime and the wrong state to have one too many. The state police were in the middle of a road-safety campaign. And on that particular summer day, they unleashed their latest weapon in the war on drink-driving: the drunkometer. The Gordons were arrested and charged on the say-so of a brightly coloured balloon and a box of bottles and tubes. For Rolla Harger, a quiet professor at Indiana University medical school, it was a good day. After years of experiment, his device was about to face the ultimate test. If it passed, the roads would soon be safer.
IN 1920, the US gave up drink. After a long crusade by the temperance movement, alcohol was outlawed. Americans had always been enthusiastic drinkers. According to the Anti-Saloon League, at the end of the 19th century there was one saloon for every 150 Americans - including those who didn't drink. The league argued that alcohol was responsible for the moral decline of American society, and that prohibition was the way to stop it.
For a few years it worked. Alcohol consumption fell to a third of its former level. Whether society benefited is arguable, but the ban did help to put the brakes on the appearance of a new phenomenon: the drunk driver. Prohibition coincided with the first big boom in car sales. By the late 1920s, however, alcohol consumption was beginning to pick up and the number of road accidents was rising. When it became obvious that prohibition would be overturned, the police braced themselves for a big increase in drunk drivers.
One of their problems was the difficulty of getting convictions. Suspects might look drunk; they might smell of alcohol; but proving they were too intoxicated to drive was tricky. There were sobriety tests. Police could ask a suspect to walk a white line, stand on one leg or put finger to nose, but the results were hardly scientific. What the police needed was a chemical test that measured how much alcohol a suspect had consumed - a measure that would persuade a judge, leave defence lawyers no room for argument, and deter drivers from drinking.
The obvious solution was to measure how much alcohol drivers had in their blood. Physiologists had established that the concentration of alcohol in blood was a good reflection of how much was in the brain, and the amount in the brain was a good measure of intoxication. But to measure blood alcohol, the police needed both a doctor and the suspect's cooperation. "It's often very hard to get a blood sample from someone who's intoxicated. They might want to sing. They might want to dance with you or play. Or they might be belligerent," says James Klaunig, director of toxicology at Indiana University and recently retired as Indiana's state toxicologist.
Urine wasn't much easier. Persuading a suspect to pee into a bottle at the side of the road was impractical, and the drunker they were the harder it was. Even if the police succeeded, the test was less reliable than a blood test. And with either, it could be hours before the police could get a sample, by which time the alcohol content might have fallen. Then it could take days to get the results back from the lab.
The police needed a test they could do themselves, something portable that would give almost instant results. Rolla Harger, a biochemist at Indiana University medical school, decided the answer was a breath test. Earlier research had shown that there was a fixed ratio between the concentration of alcohol in the blood in the lining of the lung and the concentration in the air above it. Harger found the ratio was 2000 to 1: the amount of alcohol in 2 litres of air from deep in the lungs was equal to the amount in 1 millilitre of blood.
Harger developed his first breath tester in 1931, two years before prohibition ended. It was just about portable, fitting into a box the size of a small suitcase. It looked like a chemistry set, but apart from some messy mixing of chemicals it was simple to use and gave rapid results. The suspect blew into a rubber balloon, which was then attached to a tube of purple liquid - a weak solution of potassium permanganate in sulphuric acid - and the air allowed to bubble through it. If there was alcohol on the driver's breath, it reacted with the solution and began to bleach the colour. When the liquid was a faint yellowish brown, the reaction was complete and the test stopped. A fixed amount of solution absorbed a fixed amount of alcohol - so all that was needed to calculate the amount of alcohol in the blood was a measure of how much breath had passed through the solution before it changed colour. This could be gauged from what was left in the balloon, or measured by connecting the apparatus to a gas meter.
Harger deliberately made the device as simple as possible so that judges and juries would understand how it worked and police officers could easily be trained to perform the tests. He also made it immune to most of the tricks defence lawyers might try. Experiments showed that there was no illness that affected the result, and nothing anyone might eat - garlic, cloves, strong onions - would make any difference to the drunkometer. And it had another advantage. The dramatic colour change was often enough to make people admit how much they had drunk.
Test-driving the drunkometer, Harger went out on the road with the police, randomly stopping drivers and asking them to blow into a balloon. The drunkometer worked well, but the results were disturbing. "It showed that the percentage of people who had been drinking was much higher than expected," says Klaunig.
After a series of such trials, the Indiana police adopted the drunkometer as the centrepiece of their road-safety campaign in 1937. Roy and Neva Gordon were the first Americans arrested after failing a breath test. According to the police, they had crashed head-on with another car while overtaking, rebounding to hit the car they had been trying to pass. Suspecting the Gordons of drinking, the police took them to the nearest police station, and within the hour had them blowing into a drunkometer.
In court the following week, Roy Gordon claimed he had drunk a glass of wine and three beers. His wife said she'd had two beers. Harger appeared as an expert witness: he demonstrated how the drunkometer worked then explained the results. "Gordon had the alcoholic content of about 10 ounces of whiskey," he told the judge. "That's about seven shots of 80-proof whiskey or seven glasses of wine," says Klaunig.
In a valiant attempt to get his clients off, the Gordons' lawyer demanded to be tested. Harger handed him a red balloon. He blew into it, but not hard enough. He blew again. The balloon burst. Harger tried again with a blue balloon. This time, he succeeded in attaching the breath-filled balloon to the drunkometer and began to release the air into the purple liquid. The purple remained stubbornly purple.
"If a man drank three shots of French cognac and a bottle of beer at 12 o'clock today, wouldn't that show?" demanded the lawyer. He had, he declared triumphantly, drunk exactly that before coming to court. In fact, the drunkometer was not sensitive enough to pick up that quantity of alcohol. So all the lawyer had done was prove his clients had drunk more than he had, and both were convicted of being drunk. But the drink-driving charge was thrown out: none of the witnesses could be certain which of the pair had been at the wheel.
The invention of the drunkometer meant it was now possible to lay down laws ruling how much alcohol drivers could have in their blood before they were over the limit. In January 1939, Indiana became the first US state to pass a law specifying what it meant to be drunk. Below 50 milligrams per 100 millilitres of blood was legal. Up to 150 mg was a grey zone: some people might be able to drink that much and drive safely but others couldn't, so the court needed supporting evidence to convict. Over 150 mg and there was no doubt you were too drunk to drive, and a drunkometer reading was enough to get a conviction.
Other states followed suit, and police forces across the US acquired drunkometers. They stayed in use until 1954, when Robert Borkenstein, once a trooper in the Indiana state police, invented a device that gave an instant result at the roadside: the Breathalyzer.