Sniffer dogs at Airports or Metro stations do not bark at you even if you carry things like plastic grocery bags that smell similar to explosives.
The sniffer dogs just do not respond to odours 'similar' to explosives and would detect only the real explosive, finds new research.
A research team at the Indiana University - Purdue University Indianapolis (IUPUI) has helped determine the science behind how canines locate explosives such as Composition C-4 - a plastic explosive used by the US military.
The study found that the dogs react best to the actual explosives - calling into question the use of products designed to mimic the odour of C-4 for training purposes.
"Appropriately, dogs that are trained to find real explosives are going to find real explosives and not much else," said John Goodpaster, director for the forensic and investigative sciences programme at the university.
The effectiveness of trained detector dogs is well established, but the study sought to determine which chemical compounds cause a dog to recognize a particular explosive and alert to it.
In the first phase of the study, the researchers discovered that the non-explosive chemicals given off by C-4 mimics also are present in a variety of everyday plastic objects.
Objects tested included PVC pipes, electrical tape, movie tickets, a plastic grocery bag and plastic food wrapping.
Several of the tested items emitted appreciable levels of a mimic compound recommended by some vendors for training the canines.
The second phase exposed 33 trained canines to these vapours to see if the dogs would respond.
The field trials demonstrated that the dogs failed to respond in any significant way to specific odour compounds found in C-4.
The results indicated that if the dogs are trained on the full scent, they will only detect real explosives.
"The canines are not easily fooled. You cannot pick and choose components of explosive odour and expect the dog to respond," Goodpaster said.
Canines remain the best option because of their speed, sensitivity and ability to search large numbers of items, the research said.