Evidence for signal coding for attributes such as age, sex or population of origin, was sought in odour profiles derived from natural secretions of the preorbital gland of red deer Cervus elaphus (n = 26), sika deer C. nippon (15), Chinese muntjac Muntiacus reevesi (23) and Chinese water deer Hydropotes inermis (53); from metatarsal secretions from red deer (n = 35), sika (30), fallow (193) and roe deer Capreolus capreolus (26); and from roe deer interdigital glands (n = 48). Volatiles were eluted from sample materials at body temperature, to restrict analysis to those elements which would occur within the natural odour profile; the different volatile elements were then separated by gas chromatography. Initial results of these analyses were strongly suggestive of some level of signal-coding for attributes of the signaller. Few secretions (species × gland) contained many volatiles which differed significantly between deer from different populations of origin and, except in the muntjac preorbital and sika deer metatarsal secretions, separation of populations was generally rather poor in formal discriminant function analyses (DFA). However, where no constraint is imposed on the number of volatiles allowed to enter the analysis, secretions from at least one gland accurately discriminated between the sexes in the original sample for most species; separation of sexes was only ineffective for Chinese water deer and fallow deer. Discrimination between deer of different age-classes was similarly effective in the original sample for red deer preorbital and roe deer metatarsal secretions and fair for preobital secretions of Chinese water deer; unbalanced age-structure of samples did not permit analysis for muntjac or sika deer. Sometimes the actual predictive power of discrimination was poor; it is notable that this was characteristically associated with small sample sizes and the evidence indicates that scent secretions from most species sampled here offer the potential for coding for at least the sex and age of the signaller. Although chromatographic analysis of odours from some glands reveals consistent differences between odour profiles characteristic of sex, age or population, this does not imply that the signal is deliberately coding for such information, or that the deer use the information available. It is possible that differences in the composition of secretions in relation to age or sex of the signaller, for example, is coincidental or of secondary significance.

Evidence for individuality of odour profile, coding for individual identity, was sought in scent profiles derived from natural secretions of the preorbital gland of red deer Cervus elaphus (n = 26), sika deer C. nippon (15), fallow deer Dama dama (50), Chinese muntjac Muntiacus reevesi (23) and Chinese water deer Hydropotes inermis (53); from metatarsal secretions from red deer (n = 35), sika (30), fallow (193) and roe deer Capreolus capreolus (26), and from roe deer interdigital glands (n = 48). Volatiles were eluted from sample materials at body temperature, to restrict analysis to those elements that would occur within the natural odour signal; the different volatile elements were then separated by gas chromatography. For each species considered, secretions from at least one scent gland were individually distinct and sufficiently complex to code for identity. Within our samples no two individuals produced identical odour profiles; yet in analysis of metatarsal gland secretions of individual fallow deer sampled in successive years, odour signatures of individuals remained consistent over time. The wider potential for individual coding was assessed through calculation of the number of different possible combinations of all volatiles recovered in any species×gland system. Every secretion considered seems to have the potential to provide individually characteristic signals (with the exception of secretions from the fallow deer preorbital gland, which notably contained no odour information under any analytical system). Complexity of signals differed markedly between species as did the gland-type responsible for production of the most complex or distinctive signal. No simple evolutionary patterns are apparent to account for this variation and we suggest that the form of gland selected for production of a signal carrying information about individual identity is a function of habitat type and sociality.