In a recent study conducted by the University of Alberta, they determined that a dog could detect fire accelerants such as gasoline in quantities as small as one billionth of a teaspoon.
The study was conducted to determine the effectiveness of arson detection dogs. To test this the researchers used two dog-and-handler teams. One dog was trained to detect a variety of ignitable liquids, and the other was trained with gasoline. Interestingly, the dog trained on a variety of liquids performed well detecting all accelerants, but the dog trained on gasoline could not generalize to other accelerants at extremely low concentrations.
The reason for the tests was that accelerant dogs will give an indication at a site and when the material is tested in a lab, no trace of accelerant is found.
Those of us who use dogs for various scent work will find this information helpful. The study concluded that a dog could detect odor at one billionth of a teaspoon. In previous studies by Johnson and Johnson, they determined that a dog could detect any scent at one part per trillion.
The variables that must be considered in research such as this are:
The dog’s physical ability to detect odor. Even a good dog will have off days.
The trainer’s ability to train the dog.
The air flow at the testing site.
The age of the material, in other words, how old is the scent source?
And my last comment, if the dog who was trained to detect only gasoline would not generalize to other accelerants at a low level, he was doing what he was trained to do. It does not mean that the dog could not detect the low-level scents. The question that needs to be addressed is do the non-gasoline accelerants have a gasoline chemical that evaporates quickly that the dog would identify at a higher concentration, but at low levels are there other scents or chemicals in the accelerant that are residual which the dog is not trained to find? For example, a dog taught to follow human scent will never follow animal scent. Unless the other accelerants had the chemical composition of gasoline, the dog should not indicate it.
But this is a good study and does give us an idea of the concentration of scent that a dog can detect.