Presumptive Specificity: Examining the Duquenois-Levine Field Test for Marijuana

Practitioners who have handled cases involving marijuana are likely to be familiar with the Duquenois-Levine field test.  As of 2008, the NIK NarcoPouch 908 was the most commonly used marijuana field testing kit utilizing the Duquenois-Levine Reagent.  However, despite its wide use, forensic drug expert John Kelly stated in a 2008 report to the California Attorneys for Criminal Justice (CACJ) that there were no published studies examining the validity of the field test.  Kelly’s report, which was written pursuant to one such study, criticizes the Duquenois-Levine field test as being non-specific and rendering false positives, which he asserts violate Supreme Court rulings and undermines the integrity of tens of thousands of marijuana convictions.  

This is not the only article criticizing the Duquenois-Levine field test.  Indeed, a quick Google search shows a few results among the various outlets where law enforcement officers can purchase the kits.  However, most of these articles are written by Kelly himself.  The only other criticisms found were a page on a criminal defense law firm’s website and a post from a blog about substance abuse.  The latter criticism even cites Kelly’s AlterNet article.

While there isn’t a dearth of criticism on the test itself, some of the criticism that does exist warrants consideration—namely, a 2012 article in The Open Forensic Science Journal publishing the findings of the scientific experiments which gave rise to Kelly’s 2008 report.

The Duquenois-Levine field test is a variation of drug testing techniques that have been around since the 1930s.  The test itself is named after the chemical reagent used within the field kits that reacts to the presence of cannabinoids, turning the reagent “an intense violet blue color.”  One website which sells the NarcoPouch 908 writes that the NarcoPouch is “[a] Presumptive Drug Testing [Kit]. . . [which] consists of hermetically sealed ampoules prefilled with precise amounts of testing chemicals inside a plastic outer pouch.”  

In order to perform the test, a law enforcement official places small amounts of the suspected substance into the pouch, breaks the ampoules, and “agitates” the pouch.  A positive test will show a reaction in front of the ampoules.  It should be noted that a two page instruction sheet for the NarcoPouch 908 includes a note written in all capital letters instructing users to “always retain sufficient sample of suspect material for evidential analysis by the forensic laboratory or toxicologist.”  The webpage containing the PDF instructions also notes that a NarcoPouch only identifies the presence and identity of controlled substances “with reasonable certainty,” not absolute proof.

In 2012, John Kelly, along with Dr. Omar Bagasra and Krishna Addanki, two Claflin University researchers, published “The Non-Specificity of the Duquenois-Levine Field Test for Marijuana” in the Open Forensic Science Journal,  a peer reviewed Open Access journal published by Bentham Science.  The article published the findings from a 2008 experiment which aimed to determine the specificity of the Duquenois-Levine field test in identifying marijuana.  These appear to be the findings to which Kelly referred to in his 2008 CACJ report.

In the introduction, the article cites a lack of published studies on the specificity of the Duquenois-Levine field test as one of the primary reasons for running the experiment, coupling it with the assertion that manufactures, until recently, claimed that the test can render false positives.  The authors additionally provide anecdotal evidence from a Philadelphia police official and the testing kit’s manufacturer claiming that the presumptive test kits are 99% reliable.  The article also includes anecdotal evidence from two drug analysts who stated that they had never received a false positive in thousands of tests.

The testing itself was performed with NIK NarcoPouch 908 Duquenois-Levine field testing kits which were used to test non-marijuana substances, such as chocolate, plant extracts, and medication.  After testing forty-two samples, the authors found that patchouli, cypress, and eucalyptus tested positive while lavender, spearmint, oregano, and thyme gave inconclusive results.  All other non-marijuana substances involved tested negative.

In their discussion of the testing and results, the authors immediately noted that they found the Duquenois-Levine field test to be nonspecific and subjective, arguing that the positive results for three types of non-marijuana substances indicates an ability to return false positives.  The authors also take issue with what they perceive to be the inherent subjective nature of the test, arguing that the “proper blue-violet or purple” color yielding a positive result is different for each tester, because what is blue or purple enough for one may not be true for others.  From here, the authors argue that the lack of specificity combined with the potential for false positives, subjectiveness of the test, and now-disproved statements from law enforcement officials and drug analysts could jeopardize the integrity of “tens of thousands” of marijuana convictions.

While Kelly, Bagasra, and Addanki’s article presents readers with one of the few direct experiments performed on the Duquenois-Levine field test, one cannot help but ask how their results may affect criminal cases.  While Kelly has written articles for multiple news outlets to present his arguments for six years, the test results themselves were not released until just two years ago.  Even then, the publication that accepted the article, The Open Forensic Science Journal is relatively unheard of—there is no impact factor rating listed on two independent websites.

Even outside of the relative unknown status of the journal, the entirety of Kelly’s assertions are based upon tests performed on a presumptive field test which, by its very nature, is not meant to be asserted as conclusive.  Kelly even admits to this in his article published in The Guardian, in which he states that when the test “turns a certain color—purple, in the case of marijuana—it’s deemed likely to be the real thing.”  “Likely” is hardly a conclusory standard.

The field testing kits presently assert themselves as presumptive tests.  A quick glance at five forensic supplier sites (1, 2, 3, 4, 5) selling three different manufacturers’ Duquenois-Levine field test kits all credit them as being presumptive.  Additionally, as stated before, the instructions for the NarcoPouch 908, the very product tested in the Open Forensic Science Journal article, remind users to “always retain sufficient sample of suspect material for evidential analysis by the forensic laboratory or toxicologist” (emphasis added).

There are very few indications, if any, that Duquenois-Levine field test is meant to be conclusive in definitively determining that the substance tested is or is not marijuana.  While a positive presumptive test may be used to assert probable cause in some jurisdictions—Minnesota, for example—it would be careless  to rest proof beyond a reasonable doubt on a positive presumptive field test without a more specific, confirmatory test in the crime lab.  This goes for many field testing kits, including those for blood, cocaine, and methamphetamines.

While Kelly’s criticisms of the Duquenois-Levine field test may not have gained traction within mainstream media, his research has resulted in one of the few scientific studies specifically aimed at addressing the test’s specificity.  The results themselves indicate that the Duquenois-Levine field test may not be as reliable or specific as proponents of the test once claimed, though the implications may be limited.

The experiment’s results are perhaps most effective when the defense seeks to attack the field test’s use to establish probable cause.  Given that the test itself is presumptive, the added level of non-specificity given by the three types of false positives (patchouli, cypress, and eucalyptus), can be used to attack testing reliability.  However, this could be shaky, at best.  Unless someone is in possession of the plant form of these substances, they will mostly be found in products which are impossible to mistake for marijuana:  perfume and beauty products, teas, and wood-based products, to name a few.

Trevor Addie
Staffer, Criminal Law Practitioner



Image by US DEA, via Wikimedia Commons.