Cannabis Aroma Not a Reliable Indicator of Potency

New strains of cannabis are continuously being developed and modified to express different combinations of properties. Sometimes this leads to new medicinal cannabis strains with improved therapeutic properties, but equally this sort of selective breeding is increasingly attracting the interest of the retail cannabis market with the aim of creating cannabis strains that are more attractive to consumers.

The potency of a cannabis strain can be measured by its Δ9-tetrahydrocannabinol (THC) content; THC is widely regarded as the main chemical responsible for the intoxicating ‘high’ that many recreational users seek. There are numerous ways in which scientists can gauge the THC content of a sample, but until recently, there was no similar method for scientifically gauging the aroma profile of a sample from a consumer’s perspective.

Similarly, to different types of wines and coffees, the aroma of cannabis can vary hugely between strains. Both wine and coffee have sets of established terminology used to describe their aromas. These concise yet descriptive terms give consumers a simple way to understand the product they are consuming, as well as giving experts a well-defined lexicon for discussing the relative attributes of multiple products.

The study of cannabis aroma has mostly centered on the identification of the chemical compounds responsible for the strong aroma of cannabis, but this offers extremely limited insight into the related human sensory experience.

Creating a Sensory Lexicon for Cannabis

Researchers from Headspace Sensory and XTR Systems in Colorado set out to fill this sensory gap. Firstly, they compiled a set of existing descriptors from multiple online cannabis resource pages and used them to create a ballot paper consisting of 48 different descriptors. Test participants were hired to use this ballot paper to describe the aromas of 13 samples from 11 distinct strains, all of which were unlabeled. One of the thirteen samples was an exact duplicate of another sample, functioning as an internal control for participant consistency. An additional one of the thirteen samples was the same strain as another but purchased from two different dispensaries to act as a control for retail consistency. Their research was published in PLOS One.

Although recruitment to the study was also directed at non-users, a survey of the participants revealed that most of the participants had purchased or used cannabis since it was made legal in Colorado. As a result of this, the focus of the study became much more orientated around customer perception.

In addition to assigning aroma labels to the samples, the participants were also asked to rate the potency of the strain and separately their hypothetical interest in using it on a scale from 0 to 10. A rating of 0 indicated extremely weak potency/interest and a rating of 10 indicated an extremely strong potency/interest. Finally, participants were also asked to estimate the cost of in dollars per gram of each sample, within the limits of $2/g to $22/g. 

The actual potency, in terms of THC content, of each sample, was evaluated experimentally using a gas chromatograph fitted with a flame ionization detector (GC-FID). In Colorado, the THC content of a strain is required to be present on the retail label, but it is expressed as a range of the highest and lowest values obtained from recent testing of the strain’s source, so the experimentally derived THC values are a more accurate descriptor of potency. 

Consumer Perceptions and Preconceptions of Cannabis Aromas

It was found that the most common aroma descriptor used was “earthy”, and the least used was “tar”. In general, the more frequently used descriptors were those that people would interact with most often in day-to-day life, such as spices and florals, as well as more noticeably strong odors such as diesel and cheese. 

Despite this slight disposition towards everyday categories, there was a noticeable difference in the aroma profiles for differing strains. For example, the strain OG Kush was most frequently labeled as earthy and herbal, in contrast, the strain Durban Poison was commonly labeled as citrus and lemon, with undertones of diesel. 

The frequency that the 48 descriptors were used to describe each cannabis strain was analyzed using hierarchical cluster analysis, and it was found that the 13 samples could be grouped into two clusters. Cluster A contained eight samples, all commonly described as woody and herbal. Cluster B contained the other five, which were most frequently described as citrus and sweet. Importantly, Cluster A contained both of the identical samples that were included as an internal control and Cluster B contained both samples of the same strain that had been purchased from different dispensaries, indicating that consumers may be able to discriminate between samples from different strains and do so consistently. 

Overall, participants tended to rate strains from Cluster B as more potent, more expensive, and more of interest than those from Cluster A. Contrary to the estimations of the participants, the experimentally derived THC contents showed no correlation between the aroma clusters and potency. In fact, the strain Lemon Diesel, which was experimentally found to be the second-most potent sample in the study, was ranked 11th out of the 13 samples by participants. There was also no observable links between THC content and the interest ratings or the estimated prices given by participants.

Consequences for Cannabis Retailers

THC is a non-volatile and odorless chemical, so it is unsurprising that consumers cannot accurately evaluate potency by smell alone. Nevertheless, among this group of participants, there was a clear tendency to view the more citrus and sweet-smelling strains as more valuable. The creation of the effective aroma lexicon presented here may also be helpful to the retail market in terms of branding, as well as giving researchers and consumers a universal language to discuss the perceptions of different strains.

While it fell outside the scope of this study, the authors note that there is already a widely held belief that the Cannabis sativa and Cannabis indica types have a distinct aromatic difference. Further study could follow where genetics and taxonomy of the samples are also investigated alongside their aromas to investigate this belief. A similar study with a greater variety of strains may also confirm the presence of any additional aroma clusters other than the two discovered here.