Breath Alcohol Testing and the Two Beer Defense in 1937

Early breath alcohol testing instrument

For those of you keeping up with the wintery weather hitting Toronto and the eastern US seaboard, you will recognize this post as my first “snowday” post since beginning to write this blog not quite one year ago.

This blog is based on the excellent article by Stephanie Pain cited as: Pain, S., “Catch’em on the rye”, New Scientist, 10 July 2004, 46-47, which is about the first trial on breath alcohol testing and driving in the US in 1937. The instrument used at the time was the Drunkometer.

Rolla Harger and the Drunkometer

In 1931 Professor Rolla Neil Harger (1890-1983) of the Indiana University medical school developed the first portable and practical breath-testing instrument that could be used by the police to determine the blood alcohol concentration of suspected drinking drivers (Science, 73:10, 1931). When asked what he called his instrument he jokingly replied “the Drunkometer”, but the name stuck. He continued to develop and test the instrument until the Indiana Police decided to use the Drunkometer in their road-safety campaign in 1937.

The Drunkometer worked on the principle of bubbling breath from the suspected driver into a solution of potassium permanganate and sulphuric acid, which was captured in a balloon. If there was alcohol in the breath, the solution would change in color from purple to yellowish brown color. Since the Drunkometer analysed mixed expired breath rather than the current end-expired breath used in modern breath testing instruments, the results had to be converted assuming either mixed expired breath was 2/3 that of end expired breath or adjusting for a carbon dioxide concentration of 5.5% (Borkenstein, R.F., “The Evolution of Modern Instruments for Breath Alcohol Analysis”, Journal of Forensic Sciences, 5: 395-411, 1960)

Court Case

On August 4th 1937, Roy and Neva Gordon became the first Americans arrested and tried after failing a breath test. Their car had crashed head-on after attempting to pass another car and rebounded and hit the car they had been attempting to pass. The police suspected alcohol was involved and the Gordons were breath tested with a Drunkometer at the nearest police station and failed.

During the courtroom trial conducted the next week, Neva Gordon claimed she had consumed “two beers” and her husband said he consumed “a glass of wine and three beers”. This was the first time the “two beer defense” was used to discredit the breath test results. Professor Harger attended court with a Drunkometer and explained to the court how it worked and that the Gordons must have had a lot more to drink than they testified to.

Don’t Try This in Court!

After lunch the defense lawyer asked to be tested. He blew into one balloon but couldn’t inflate it.  He blew again and the balloon burst. He finally inflated another balloon and his breath was bubbled through the potassium permanganate solution which did not change color, indicating no alcohol could be measured by the Drunkometer. The defense lawyer, not understanding the results, said “If a man drank three shots of French cognacs and a bottle of beer at 12 o’clock, wouldn’t that show?” This was exactly what the defense lawyer drank at lunch, to show the error of the Drunkometer, but what he showed was that the Drunkometer was not sensitive enough to detect such low BACs and confirmed that the Gordons must have consumed a lot more alcohol. Although the prosecution was successful with the Drunkometer, it was not shown conclusively which of the Gordons were driving and so the drinking-driving charge was dismissed but they were convicted of being drunk in a public place.

Two Beer Defense Redux

Since 1937, breath alcohol testing has improved tremendously but the two beer defense hasn’t change and has remained the most common defense used against breath alcohol testing in Canada and probably in other countries as well. When I was a forensic alcohol toxicologist at the Centre of Forensic Sciences in Toronto, I always felt the two beer defense was a sterile and corrosive one.  It didn’t matter how well the police breath technicians were trained, how well they conducted the breath test, and all the other scientific controls such as blank tests, duplicate breath tests, and alcohol standards, all the defense had to do was get his client onto the stand and testify that only “two beers” or another small amount of alcohol was consumed and thus could be acquitted.

Science Wins This Round

On July 2nd 2008, Parliament in Ottawa enacted a new law that restricted the use of the two beer defense. Now the defense has to show that there was an error in the operation of the evidential breath alcohol testing instrument. This law was confirmed as constitutional by the Supreme Court of Canada (SCC) in its decision of R vs St Onge Lamoureaux on November 2nd 2012.

In a perceptive blog about the SCC, posted by Paul Bird at the York University website (, he states:

“One potential benefit of the new legislative regime is that successful defenses under the first requirement could spur more stringent police control and handling of breathalyzer devices. There was no similar incentive under the old regime that allowed Carter (i.e. “two beer”) defenses. Raising a reasonable doubt by means of a Carter defense shone a light on the accused’s testimony about his or her consumption, which implicated the breathalyzer results only indirectly.

By contrast, the new legislative regime requires that the accused raise a reasonable doubt about the test results specifically, without reference to his or her consumption. Successful defenses under this regime will necessarily point out concrete problems with the device or how it was used. This could be useful information that could inform police practices and even technological refinement of the devices themselves. Whether this feedback materializes is yet to be seen, but it is clear that science has won this round at the SCC.”

Plus ça Change, Plus C’est la Même Chose

If you find yourself wondering why I chose to write this blog, it’s because I’ve decided to add yet another new feature to my blog. I’m calling it Plus ça Change, Plus C’est la Même Chose which translates roughly into “The More Things Change, the More they Stay the Same.”

In this feature, I will highlight older historical studies about forensic alcohol toxicology and discuss how in some areas, things have not changed much in 50 years or more.


Breath Alcohol Testing and the Two Beer Defense in 1937 — 5 Comments

  1. Another great blog entry, Jim. In my experience, though, the “two beer” defense is not taken seriously by most juries, at least not here in Florida. Florida juries rarely find it convincing if records show the instrument was properly calibrated, maintained and tested. Evidence is easily presented to support the instrumental results, whereas “two beer” claims by defendants and associates of the defendants are unsubstantiated by any material evidence. It’s their verbal word against hard scientific data and here, at least, juries believe the scientific data. For those of us involved in DUI/DWI cases on a regular basis, the claim of “I only had two beers” is so common (and so commonly refuted by blood or breath testing), as to be an almost laughable cliché. On one occasion, when I was asked by a prominent defense attorney to reconcile such a claim by a defendant with my results of analysis of that defendant’s blood (which demonstrated a BAC of more than twice the “legal limit”), I responded that there were three possibilities:

    1. He was not being truthful
    2. He honestly only remembered the first two (or perhaps the last two) beers.
    3. They were REALLY BIG beers (hands spread wide for emphasis).

    After the laughter subsided, there were no further questions. The defendant was ultimately convicted. As I recall, that particular attorney never asked me about the “two beer” defense again.

  2. The first class I took in Toxicology was at the Indiana University Medical Campus in Indianapolis. It was in the summer of 1968. The Professor for the class was Rolla Harger. He was an amazing man. I saw he operate the “Drunkometer” on a number of occasions during training classes for law enforcement officers. During the drinking/testing sessions blood was drawn from some subjects and comparisons made between the test results (blood analysis by GC vs Drunkometer). Dr. Harger after reviewing the blood results remarked that “the blood results were close”. One other comment he often made was that the best bags for rebreathing samples were Sears brand. Dr. Harger truly was an amazing man.

    • Dan, thanks so much for this addition to my blog. These first hand stories really help to personalize the men and women who have contributed so much to the field of forensic alcohol toxicology. Thanks for sharing!

  3. Great post. It’s been awhile since I researched the subject, but one of the things that I noted in my reading a decade ago was the that readings on the machines were predicated on the ridiculous notion that everyone taking the test was able to exhale a Tidal Volume of 2500ml as a standard each and every time just like stroke volume when measuring cardiac output. The formula for measuring CO = HR X SV with HR being the variable. Blood loss is not the only thing that effects cardiac output causing the heart rate to increase. Anything preventing blood from circulating normally will increase the heart rate to keep the cardiac output up. So let’s apply this to the breathalyzer. All kinds of things can effect tidal volume. Smoking is number one and their is a close nexus between smokers and alcoholics. So to my way of thinking attainment of a tidal volume of 2500ml of every time someone blows into a machine would be impossible given the plethora of physiologic conditions effecting an individuals tidal volume. My recommendation would be to check the tidal volume and set the machine accordingly.

Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>