In this series of posts we are going to examine this seemly simple question:

  1. What is the goal and the purpose of testing of unknowns generally? How do we best design a test for marijuana?
  2. How is most marijuana testing conducted in the United States?
  3. What is microscopic morphological examination? Is it a “good” test?
  4. What is the modified Duquenois-Levine test? Is it a “good” test?
  5. What is Thin Layer Chromatography? Is it a “good” test?
  6. Is the combination of all three tests create a “good” testing scheme?
  7. Is there a better way to test for marijuana?

Part 4: What is the modified Duquenois-Levine test? Is it a “good” test?

Modified Duquenois-Levine testing

What is it?

It is technically referred to as a colorimetric test. In short, it is a color test. A totally different selection of the unknown is sampled and subjected to this test. A reagent is added to the unknown. The reagent is made up of a combination of vanillin, acetaldehyde, and ethanol in a specific ratio of these component materials.

Mechanically how is it preformed?

This totally separate sample from the unknown is placed into typically a test tube. A certain amount of the Duquenois reagent is added (typically about 10 drops). The tube is closed. The tube is agitated (shaken) for an unspecified period. The tube is reopened. Concentrated hydrochloric acid is then added (usually about 20 drops). The tube is closed. It is agitated (shaken) again. Any color change is then noted. The tube is reopened. Chloroform is added. The tube is closed. It is again agitated (shaken) or vortexed (mixed). The analyst is looking for a color change (thought to be violet or purple) and a separation into two layers.

This is the end result of the modified Duquenois-Levine test. This picture is after the HCl and the chloroform are added. There is a deep purple color at the top and then a pink color at the bottom.

This is the end result of the modified Duquenois-Levine test. This picture is after the HCl and the chloroform are added. There is a deep purple color at the top and then a pink color at the bottom.

How is the typical crime laboratory analyst trained to conduct this test?

They are trained in the process and the steps in the mechanical process. No crime laboratory analyst is ever instructed by a doctoral level analytical chemist as to why this process results in any sort of color change or the way the physical separation occurs in the colors with the addition of chloroform or why it these changes happen at all.

In essence, it is a subjective test as it is based upon perception of color by the analyst and the perception of this developing of a separate layer once chloroform is added.

It is a second sample that is tested. Prior to the testing by modified Duquenois-Levine technique, there is no microscopic morphological examination conducted on this sample. There is no TLC analysis on this sample.

It is tested one time and then discarded. No other testing is performed on this sample.

The analysts are not academically trained in the theory of the reagent use, and are not taught about cross-reactivity and false positives or other sources of errors.

Is this a verifiable test?

Potentially, yes. Digital cameras exist. Photographs can be taken just like the one above. Heck, even video can be taken to show how it is conducted on a particular sample from the unknown. However, the modern practice is to not take photos. They videotape nothing.  The laboratory produces nothing verifiable in court that the test was even conducted or that the analyst’s perception of the change in color did happen and was correct or that there was this separation once the chloroform is added.

Again, there are no crime laboratories that I am aware of that use the ACE-V (Analysis, Comparison, Evaluation, and Verification) technique that one would find in fingerprint identification with a double check in real time by a fellow bench analyst. In essence, it is checked one time, by one person with no double check by another, and nothing produced that proves that the analysis was done or that the features that are reported as present were in fact objectively present.

Is there empirical validity studies that prove that this is a specific and confirmatory test?

Nope. In fact the empirical studies clearly show the opposite. It is not specific for THC. The reaction is not unique to THC. In fact, the studies clearly show that it is not even meaningfully selective for THC. Chinese motherwort if tested, by this method will turn violet. But that is not all. There are a great many plants that yield similar color results when Duquenois-Levine testing is applied. Yet analysts are taught that is color change is diagnostic of THC. Nakamura himself published and acknowledges that M. J. de Faubert Maunder listed 25 species of plants which exhibited violet to purple colors in the Duqenois test, and were extractable in chloroform, which is the distinguishing features of the Duquenois Levine test. (502) Nakamura tested 23 of those species noted by M. J. de Faubert Maunder. Using the Duquenois-Leine test, he found that the violet to purple color when testing the leaves of the following (502):

  • coffee
  • a species of gum copal called Caplafer conjugata
  • gum Kawri
  • wood sage
  • Thuja occidentalis
  • Sandarac

He published that he tested and found violet to purple color reactions using the Duquenois-Levine test in other plants as well, and not just the leafy portions of these plants. (502). That list includes:

  • calamus
  • culver root
  • ginger
  • gum animi
  • gum copal
  • gum myrrh
  • henna
  • lettuce opium
  • sandal wood
  • tolu
  • wood betony
  • liquorice
  • nutmeg
  • poison flag

The United Nations study on marijuana concludes that Duquenois-Levine test on substances completely unrelated to marijuana will give false positives too. This includes Arthemisia Grancunulus, Eucalyptus Glabulus Labill and several other herbs that are very common such as rosemary, thyme, and sage.

None of these items above contain THC. Yet, they feature these color changes and some even have the separation when chloroform is added that the analysts are trained to look for when applying these reagents.

While we have examined the propriety and suitability of this modified Duquenois-Levine testing in the identification of THC, we cannot judge this Duquenois-Levine test in isolation. We must remember and in fairness this is a 3 test process and technique. What compounds the issue of the possibility of error is that as the analyst is not using the same sample throughout each of these 3 tests. It is in fact, three different samples are taken of the unknown and each test gets its own sample with no sample receiving the benefit of all three tests. The analyst presumes that the sampling and the sample selection of the unknown is homogeneous meaning that each sample will be exactly identical. This is not a justified scientific assumption (An assumption is not drawn from evidence; it is a hypothesis {my assumption can be tested by looking at the dictionary}. A presumption implies a basis in evidence {the legal presumption of innocence}) Just because the sought after features of the microscopic morphological examination were present in the first sample, that doesn’t necessarily mean that it will be present in the next sample that the analyst now subjects to modified Duquenois-Levine testing.

Nakamura tested and found that mace and nutmeg, which yielded colors similar to that obtained with marijuana with the Duquenois-Levine test, could be “credibly confused” with marijuana based on microscopic appearance. (502)

2 Responses to “The myth of specific identification of Marijuana in criminal court Part 4: What is the modified Duquenois-Levine test? Is it a “good” test?”

  1. Arizona says:

    Can you please let us know which article of Nakamura’s that you are referring to when you site “(502)”? Thanks so much!

  2. J.I. Thornton and G. R. Nakamura, The Identification of Marijuana, Journal of Forensic Science Society , V 12, 461 (1972) and G. R. Nakamura,
    Forensic Aspects of Cystolith Hairs of Cannabis and Other Plants, Journal of the AOAC V52, 5, (1969)

Leave a Reply