Different methods for volatile sampling in mammals.

Previous studies showed that olfactory cues are important for mammalian communication. However, many specific compounds that convey information between conspecifics are still unknown. To understand mechanisms and functions of olfactory cues, olfactory signals such as volatile compounds emitted from individuals need to be assessed. Sampling of animals with and without scent glands was typically conducted using cotton swabs rubbed over the skin or fur and analysed by gas chromatography-mass spectrometry (GC-MS). However, this method has various drawbacks, including a high level of contaminations. Thus, we adapted two methods of volatile sampling from other research fields and compared them to sampling with cotton swabs. To do so we assessed the body odor of common marmosets (Callithrix jacchus) using cotton swabs, thermal desorption (TD) tubes and, alternatively, a mobile GC-MS device containing a thermal desorption trap. Overall, TD tubes comprised most compounds (N = 113), with half of those compounds being volatile (N = 52). The mobile GC-MS captured the fewest compounds (N = 35), of which all were volatile. Cotton swabs contained an intermediate number of compounds (N = 55), but very few volatiles (N = 10). Almost all compounds found with the mobile GC-MS were also captured with TD tubes (94%). Hence, we recommend TD tubes for state of the art sampling of body odor of mammals or other vertebrates, particularly for field studies, as they can be easily transported, stored and analysed with high performance instruments in the lab. Nevertheless, cotton swabs capture compounds which still may contribute to the body odor, e.g. after bacterial fermentation, while profiles from mobile GC-MS include only the most abundant volatiles of the body odor.