Chapter 6 Web Topics

6.1 How Are Pheromones Identified?


Identification and analysis of chemical signals has lagged behind the analysis of visual and auditory signals because our own sense of smell is rather poor. We generally cannot smell most animal pheromones, so we are not aware of most chemical senders. In addition, chemical signals are complex blends in many species, especially in insects and mammals. To compound matters, the amount of chemical released is often small, which makes it difficult to collect a sufficiently large sample for chemical analysis. Nevertheless, new techniques have improved our ability to collect the secretions and analyze the chemical components. Although it is not possible to conduct “playback” experiments in digital form from a computer as we do with acoustic and video signals, one can release controlled amounts of specific chemical components to assess differential responses. In this Web Topics unit we shall briefly outline the typical steps taken to identify chemical signals, drawing in large part from Chapter 2 of Wyatt (2014).

Step 1: Finding a bioassay

The first step in the process of identifying a chemical signal is observing the animals and discovering behaviors or interactions that appear to be influenced by chemicals. As we mentioned in the main text, ritualized marking behaviors and release postures are clear indicators of the sending of chemical signals, and these behaviors also provide clues to the likely location of secretory glands. However, many chemical signals are released without such visual correlates. It is more important to identify receiver responses to putative chemical signals because these will provide the bioassay needed to test the salience of chemical components in all of the subsequent steps below. Olfactory investigation behaviors such as increased sniffing (mammals) or tongue flicking (lizards and snakes), and very obvious responses such as flehmen in male mammals, aggregation in bark beetles, and wing fluttering in male moths, make excellent bioassays. Some chemicals signals produce specific emotional responses, such as the agitation, mobbing, and retreat responses from alarm pheromones. Chemical signals with a priming function have more subtle effects on receivers, such as changes in hormonal profiles and reproductive condition. The most useful bioassay for further analysis of a putative pheromone or chemical signal is a reliable behavior or other response associated with receipt of the natural chemical product.

Simple bioassays that are easily detected in a laboratory situation will be most useful. For instance, the wing fanning behavior of male moths is a far simpler test of responses to female pheromone products than flight tests in a wind tunnel. However, there are tradeoffs here, because males will readily wing fan in response to partial blend components, whereas upwind anemotactic flight depends on recognition of the appropriate mixture (Charles Linn, personal communication). The ability to discriminate among odors can also be tested with various types of laboratory tasks, such as reward-based operant conditioning discrimination tasks, habituation tasks, and free-field natural presentation of stimuli. Care must be taken to ensure that all other test conditions are controlled for, such as time of day, animal age, reproductive state, and sex. Nevertheless, the results may differ depending on the methodology used (Schellinck et al. 1994, 1995; Brown et al. 2000; Schellinck et al. 2001). Finally, because many chemical signals are blends that may contain redundant information, mere removal of a specific component may not diminish the response, even though it is an important component. Thus care must also be taken to present different combinations of the blend components to look for synergistic effects, using a bioassay that is highly sensitive to the quality of the mixture.

If the olfactory receptors of a species are sufficiently well understood, one can also use the animal’s own sense organs as direct bioassay detectors. This procedure has been well developed in the study of moth pheromones, where the investigator implants electrodes in the base of the antennae to record the electrical responses (electroantennogram, EAG) to various airborne odors (van der Pers and Denotter 1978; Groot et al. 2005). Simple portable units with moth antennae as the sensor have been developed to monitor presence of pheromone in the wild (van der Pers and Minks 1993, 1998). This method will not detect pheromone components for which there are only a few receptor cells, so a more accurate method is single-sensillum recording (Baker et al. 2004; Linn et al. 2007; Inoshita et al. 2011). Neurophysiological studies are also widely used to understand pheromone detection in vertebrates (Su et al. 2009; Tirindelli et al. 2009; Touhara and Vosshall 2009).

Step 2: Collecting the chemicals

Pheromone chemicals are often collected by squeezing or swabbing potential glands or by excising the whole gland. Unfortunately, such crude extracts of gland products risks gathering many non-pheromone components, including carrier compounds and chemical precursors of the pheromone. For volatile pheromones, a better method is to trap and concentrate the chemicals actually released by the animal (or plant) into what is called the headspace area above it. With this method, clean air (or water) is passed over the animal and the chemicals in the exhaust are then trapped by special chemical “sponges” or “wicks”. These are porous polymer resins that adsorb organic molecules (e.g., Porapak-Q and Tenax). They release the chemicals when heated or washed with organic solvents. Exhaust air can also be cooled to -195° C with liquid nitrogen. Washing a gland or polymer resin with solvent results in a relatively diluted solution with low resolution for subsequent analyses. An alternative method is Solid Phase MicroExtraction (SPME). This technique avoids the use of solvents by employing an inert fiber coated with an absorbing polymer, which can then be directly exposed to the input air or water directed into the gas chromatography analyzer (see below) (Pawliszyn 1997, 1999; Ouyang et al. 2011).

Step 3: Separating and identifying the chemicals

Once the chemicals have been collected, the goal is to find the active components. To make some initial inroads into figuring out the pheromone compound, a traditional chemical technique called fractionation may be undertaken. This process separates the components based on their physical characteristics, such as solubility in different solvents. The sample is typically mixed in water and a non-water-soluble organic solvent and shaken well. Once the polar water phase and non-polar solvent phase separate out, one can then use the bioassay test to determine which fraction contains the active component. This process narrows down the general chemical class of the active component into lipophilic alkanes, oils, and fats, versus hydrophilic acids, alcohols, aldehydes, and esters, and also purifies and concentrates the pheromone.

Next, the sample is submitted to analysis by gas-liquid chromatography (GC). The development of this technique has revolutionized the study of chemical signals, and it is especially useful when there is a mixture of components and small amounts of sample. It can only be used for chemicals that vaporize upon heating without decomposing. The gas chromatograph (Figure 1) separates the chemicals on the basis on their physical characteristics. The instrument inputs a steady flow of inert gas, usually nitrogen or helium, into a long thin tube called the column (although it is coiled for space efficiency). The column is then heated, and a small sample of gas or solution containing the pheromone is then injected into the tube with a syringe. The inside of the tube is coated with a thin layer of liquid or polymer that interacts with the chemicals as they pass through. The compounds in the sample are separated based on their vapor pressures (boiling points) and their relative solubility in the inert gas. Smaller, more volatile components travel faster through the tube and exit first, followed by larger and less volatile components. The retention time in the tube is the key output variable that distinguished the components. A known reference chemical is included in the sample so that the unknown components can be assessed relative to the reference peak.

Figure 1. Gas chromatograph coupled with flame ionization detector and electroantennogram recorder. Inert gas from blue canister flows into the chromatograph at a controlled flow rate. Gas flows into the chromatograph (tan) through the column, a thin tube coated with polymer. The column coil is heated to a controlled temperature. A sample of insect pheromone is injected into the injector input dock. After sample has traversed the column, it is split into two streams: one branch flows into the detector to produce a record of the retention times of the chemical mix (FID), and the other branch flows over a real insect antenna, which is wired with electrodes to record neural spiking activity (EAD). (After Leal et al. 2001; Wyatt 2014.)

Some type of detector is needed to record the spikes as each chemical component exits the tube. A common detector is the flame ionization detector (FID). The FID ionizes the chemicals as they emerge from the column using a small hydrogen and air flame. It burns the chemicals, producing positively charged ions and negatively charged electrons. Two electrodes are used to measure the potential difference between them. The measured current is proportional to the number of reduced carbon atoms in the flame, so it is sensitive to the mass of the chemical compound. Another detector is the mass spectrometer (MS). This instrument also ionizes the sample, but with an electron beam, and the ions are separated according to their mass-to-charge ratio based on the electromagnetic fields they generate. This method is especially useful for analyzing larger molecules such as peptides.

The GC analysis can identify the chemical components of a mixture, but it does not reveal which ones are the active ones, and certainly does not allow the investigator to collect aliquots of the components for further testing with a bioassay. However, an ingenious method has been devised to obtain this information at least from insect subjects. It involves coupling the CD analysis with electroantennogram detection (EAD) or single-sensillum recording, as also shown in Figure 1. Basically the gas sample is split once it has passed through the column, with parallel streams entering the FID or MS detector and the neurophysiological set-up with the insect’s antenna. The timing of the spikes in the neural responses can then be linked to specific peaks in the detector output. Some examples of studies on a variety of insects using this technique can be found in the following articles: Toth et al. 1991; Leal et al. 1997; Gunawardena et al. 1998; Kalinova et al. 2006; Leal et al. 2006; Burger et al. 2008; Kim et al. 2009; Vitta et al. 2009; Zhang et al. 2011.

When it is not feasible to conduct parallel neurophysiological and gas chromatograph recordings, especially for vertebrates, there are two alternatives. First, one can pursue the fractionation process to separate and purify chemical components using the chemist’s tool kit. This strategy requires obtaining a sufficiently large sample of the pheromone-containing substance from the gland or resin washing. At each step, the separated aliquots are tested for bioactivity using the bioassay. Second, if gas chromatography has resulted in identification of the chemical structures of the components, these compounds can then be synthesized. Again, each synthesized component must be tested with the bioassay to make sure it is the correct chemical. This process can be tedious if there are several active components acting synergistically along with a variety of inert gland components.

A good example of the use of these techniques is the discovery of the sex attractant of female red-sided garter snakes (Thamnophis sirtalis parietalis). Early field studies showed that males were attracted to the trails left by moving females and they required an intact vomeronasal organ to accomplish this task. This indicated that a non-volatile component of the female’s skin was the likely source of a chemical signal. Once males approach a gravid female, they begin to court them by rubbing their chin along the female’s dorsal side; this chin-rubbing behavior provided the reliable bioassay. Organic solvent washes of the female’s skin revealed a large amount of lipid material, but there are no obvious skin glands. Serum and liver extracts from receptive females applied to unreceptive females made them attractive to males, so it was initially hypothesized that the pheromone source was the egg-yolk protein vitellogenin, which is produced by females under the control of estrogen (Garstka and Crews 1981). However, field trials and other evidence did not support this hypothesis. Gas chromatograph analysis of the skin lipids (coupled with MS and nuclear magnetic resonance spectroscopy) then revealed the presence of a series of 12 non-volatile saturated and monounsaturated long-chain (29- to 39-carbon) methyl ketones. These compounds were then successfully synthesized, and the mixture of synthetic compounds produced the full courtship bioassay response when applied to towels (Mason et al. 1989; Mason et al. 1990). This was the first pheromone identified and synthesized in reptiles (Mason and Parker 2010). Several amphibian and many rodent pheromones have now been identified (Beynon and Hurst 2004; Houck et al. 2007; Houck 2009; Woodley 2010). Many of these pheromones are proteins, requiring additional techniques to identify them such as polymerase chain reaction (PCR) amplification of DNA and amino acid sequencing.

Further reading

Millar, J.G. and K.F. Haynes. 1998. Methods in Chemical Ecology, Vol. 1. London: Chapman and Hall.

Haynes, K.F. and J.G. Millar. 1998. Methods in Chemical Ecology, Vol. 2. London: Chapman and Hall.

Wyatt, T.D. 2014. Pheromones and Animal Behaviour: Chemical Signals and Signatures. Cambridge: Cambridge University Press.

Web resources

Solid phase microextraction:

Gas chromatography:

Flame ionization detector:

Mass spectrometry:

UC Davis Chemical Ecology and Olfaction group:

Oregon State University pheromone labs: and

Literature cited

Baker, T.C., S.A. Ochieng, A.A. Cosse, S.G. Lee, J.L. Todd, C. Quero and N.J. Vickers. 2004. A comparison of responses from olfactory receptor neurons of Heliothis subflexa and Heliothis virescens to components of their sex pheromone. Journal of Comparative Physiology a-Neuroethology Sensory Neural and Behavioral Physiology 190: 155–165.

Beynon, R.J. & J.L. Hurst. 2004. Urinary proteins and the modulation of chemical scents in mice and rats. Peptides 25: 1553–1563.

Brown, R.E., L. Stanford and H.M. Schellinck. 2000. Developing standardized behavioral tests for knockout and mutant mice. ILAR Journal 41: 163–174.

Burger, B.V., W.G.B. Petersen, B.T. Ewig, J. Neuhaus, G.D. Tribe, H.S.C. Spies and W.J.G. Burger. 2008. Semiochemicals of the Scarabaeinae—VIII. Identification of active constituents of the abdominal sex-attracting secretion of the male dung beetle, Kheper bonellil, using gas chromatography with flame ionization and electroantennographic detection in parallel. Journal of Chromatography A 1186: 245–253.

Garstka, W.R. and D. Crews. 1981. Female sex pheromone in the skin and circulation of a garter snake. Science 214: 681–683.

Groot, A., C. Gemeno, C. Brownie, F. Gould and C. Schal. 2005. Male and female antennal responses in Heliothis virescens and H. subflexa to conspecific and heterospecific sex pheromone compounds. Environmental Entomology 34: 256–263.

Gunawardena, N.E., F. Kern, E. Janssen, C. Meegoda, D. Schafer, O. Vostrowsky and H.J. Bestmann. 1998. Host attractants for red weevil, Rhynchophorus ferrugineus: Identification, electrophysiological activity, and laboratory bioassay. Journal of Chemical Ecology 24: 425–437.

Houck, L.D. 2009. Pheromone communication in amphibians and reptiles. Annual Review of Physiology 71: 161–176.

Houck, L.D., C.A. Palmer, R.A. Watts, S.J. Arnold, P.W. Feldhofff and R.C. Feldhofft. 2007. A new vertebrate courtship pheromone, PMF, affects female receptivity in a terrestrial salamander. Animal Behaviour 73: 315–320.

Inoshita, T., J.R. Martin, F. Marion-Poll and J.F. Ferverur. 2011. Peripheral, central and behavioral Responses to the cuticular pheromone bouquet in Drosophila melanogaster males. Plos One 6: 9.

Kalinova, B., P. Jiros, J. Zd’arek, X.J. Wen and M. Hoskovec. 2006. GC x GC/TOF MS technique — A new tool in identification of insect pheromones: Analysis of the persimmon bark borer sex pheromone gland. Talanta 69: 542–547.

Kim, J., S.G. Lee, S.C. Shin, Y.D. Kwon and I.K. Park. 2009. Male-produced aggregation pheromone blend in Platypus koryoensis. Journal of Agricultural and Food Chemistry 57: 1406–1412.

Leal, W.S., J.M.S. Bento, Y. Murata, M. Ono, J.R.P. Parra and E.F. Vilela. 2001. Identification, synthesis, and field evaluation of the sex pheromone of the citrus fruit borer Ecdytolopha aurantiana. Journal of Chemical Ecology 27: 2041–2051.

Leal, W.S., A.L. Parra-Pedrazzoli, A.A. Cosse, Y. Murata, J.M.S. Bento and E.F. Vilela. 2006. Identification, synthesis, and field evaluation of the sex pheromone from the citrus leafminer, Phyllocnistis citrella. Journal of Chemical Ecology 32: 155–168.

Leal, W.S., P.H.G. Zarbin, H. Wojtasek, S. Kuwahara, M. Hasegawa and Y. Ueda. 1997. Medicinal alkaloid as a sex pheromone. Nature 385: 213–213.

Linn, C.E., M.J. Domingue, C.J. Musto, T.C. Baker and W.L. Roelofs. 2007. Support for (Z)-11-hexadecanal as a pheromone antagonist in Ostrinia nubilalis: Flight tunnel and single sensillum studies with a New York population. Journal of Chemical Ecology 33: 909–921.

Mason, R. and M. Parker. 2010. Social behavior and pheromonal communication in reptiles. Journal of Comparative Physiology A 196: 729–749.

Mason, R.T., H.M. Fales, T.H. Jones, L.K. Pannell, J.W. Chinn and D. Crews. 1989. Sex pheromones in snakes. Science 245: 290–293.

Mason, R.T., T.H. Jones, H.M. Fales, L.K. Pannell and D. Crews. 1990. Characterization, synthesis, and behavioral responses to sex attractiveness pheromones of red-sided garter snakes (Thamnophis sirtalis parietalis). Journal of Chemical Ecology 16: 2353–2369.

Ouyang, G.F., D. Vuckovic and J. Pawliszyn. 2011. Nondestructive sampling of living systems using in vivo solid-phase microextraction. Chemical Reviews 111: 2784–2814.

Pawliszyn, J. 1997. Solid Phase Microextraction: Theory and Practice. New York: Wiley-VCH.

Pawliszyn, J. 1999. Solid Phase Microextraction: a Practical Guide. New York: Marcel Dekker.

Schellinck, H.M., C.A. Forestell and V.M. LoLordo. 2001. A simple and reliable test of olfactory learning and memory in mice. Chemical Senses 26: 663–672.

Schellinck, H.M., E. Rooney and R.E. Brown. 1994. Methodological questions in the study of the rat’s ability to discriminate between the odours of individual conspecifics. Advances in the Biosciences 93: 427–436.

Schellinck, H.M., E. Rooney and R.E. Brown. 1995. Odors of individuality of germ-free mice are not disciminated by rats in a habituation–dishabituation procedure. Physiology and Behavior 57: 1005–1008.

Su, C.Y., K. Menuz and J.R. Carlson. 2009. Olfactory perception: receptors, cells, and circuits. Cell 139: 45–59.

Tirindelli, R., M. Dibattista, S. Pifferi and A. Menini. 2009. From pheromones to behavior. Physiological Reviews 89: 921–956.

Toth, M., G. Szocs, C. Lofstedt, B.S. Hansson, F. Schmidt and W. Francke. 1991. Epoxyheptadecadienes identified as sex-pheromone components of Tephrina arenacearia HBN (Lepidoptera, Geometridae) Zeitschrift Fur Naturforschung C–a Journal of Biosciences 46: 257–263.

Touhara, K. and L.B. Vosshall. 2009. Sensing odorants and pheromones with chemosensory receptors. Annual Review of Physiology 71: 307–332.

van der Pers, J.N.C. and C.J. Denotter. 1978. Single cell responses from olfactory receptors of small ermine moths to sex attractants. Journal of Insect Physiology 24: 337–343.

van der Pers, J.N.C. and A.K. Minks. 1993. Pheromone monitoring in the field using single sensillum recording. Entomologia Experimentalis Et Applicata 68: 237–245.

van der Pers, J.N.C. and A.K. Minks. 1998. A portable electroantennogram sensor for routine measurements of pheromone concentrations in greenhouses. Entomologia Experimentalis Et Applicata 87: 209–215.

Vitta, A.C.R., B. Bohman, C.R. Unelius and M.G. Lorenzo. 2009. Behavioral and electrophysiological responses of Triatoma brasiliensis males to volatiles produced in the metasternal glands of females. Journal of Chemical Ecology 35: 1212–1221.

Woodley, S.K. 2010. Pheromonal communication in amphibians. Journal of Comparative Physiology A 196: 713–727.

Wyatt, T.D. 2014. Pheromones and Animal Behaviour: Chemical Signals and Signatures. Cambridge: Cambridge University Press.

Zhang, Q.H., J.H. Ma, F.Y. Zhao, L.W. Song, J.H. Sun and A.I. Cognato. 2011. Aggregation pheromone of the Oriental spruce engraver Pseudips orientalis. Agricultural and Forest Entomology 13: 67–75.

6.2 Guide to Organic Compounds and Biosynthetic Pathways

Basic compounds and nomenclature

Organic molecules are based on a chain of carbon atoms with 1–3 attached hydrogen atoms. The carbon backbone usually forms a zigzag shape because of the tetrahedral arrangement of the four possible bonds around each carbon atom. The hydrogen atoms attached to the backbone lie in two planes, one above the plane depicted below by the solid wedges, and one below the plane depicted below by the dashed wedges.

The structure is often simplified to show just the carbon backbone.

When other atoms such as oxygen (O), nitrogen (N), sulfur (S), or other functional groups are present, their letter symbols are inserted in the chain. The naming of organic compounds indicates the length of the carbon chain and what and where the important functional groups are located.

Functional Groups

The table below shows the formulae and names of the main functional groups.

Alcohol –OH Hydroxy- -ol
Aldehyde –CH=O Formyl- -al
Ketone >C=O Oxo- -one
Carboxylic acid –COOH Carboxy- -oic acid
Ester –COOR R-oxycarbonyl- -R-oate
Amine –NH2 Amino- -amine
Thiol –SH sulfanyl- -thiol

Wikipedia Nomenclature of Organic Chemistry.


Hydrocarbons contain only carbon and hydrogen. They vary in length and the presence, number, and position of double or triple carbon bonds. The basic classes are shown in the table below.

Alkane CnH2n+2 -ane
Alkene CnH2n -ene
Alkyne CnH2n-2 -yne

The longest continuous chain of carbons (n) is the parent structure and specifies the prefix: 1 = meth-, 2 = eth-, 3 = prop-, 4 = but-, 5 = pent-, 6 = hex-, 7 = hept-, 8 = oct-, 9 = non-, 10 = dec-, 12 = dodec-, 15 = pentadec-, and so on. The position of a functional group or double/triple bond is specified by counting the carbons from whichever end gives the lowest possible position number. For example, the compound shown below, which is the pheromone of the pink bollworm (Pectinophora gossypiella), is (Z,E)-7,11-hexadecandien-1-ol acetate. The “dien” specified that there are two double bonds, and the numbers specify that they occur at carbons 7 and 11 counting from the functional group (acetic acid) on the right (from Wyatt 2014).


Molecules with the same formula can be assembled in different ways. The arrangement of atoms affects the shape of the molecule and therefore affects its binding (ligand) properties with receptor and binding proteins. There are several types of isomers. Positional isomers have functional groups or double bonds in different positions, as in the examples shown below (from Wyatt 2014).

Stereoisomers have the same connectivity but differ in the arrangement of atoms in space. Geometrical isomers have a spatial twist around a double bond and are indicated by E and Z (formerly, trans and cis, respectively). E stands for entgegen (opposite) and Z stands for zusammen (together), and refers to the location of the “high-priority” substituents on the same or different sides of the double bond.

Optical isomers, also called enantiomers, are mirror images of each other. Solutions of pure enantiomers rotate the plane of polarized light in opposite direction. Variants are indicated by L and D (or left and right, or – and +).

Classes of larger organic compounds

Amino acids, peptides and proteins

Amino acids contains both amine and carboxyl functional groups. Different amino acids have a variety of other functional groups and side chains attached, including cyclic hydrocarbon units. There are 20 standard amino acids used by cells in protein biosynthesis that are specified by the DNA genetic code. These twenty amino acids can be biosynthesised from simpler molecules, but organisms differ in how many they are able to produce and essential amino acids must be obtained in their diet. A few are shown below, and more can be found at

There are two naming conventions for amino acids—single-letter abbreviations and three-letter abbreviations, as follows:

Amino acid
1-letter name
3-letter name
Alanine A Ala
Arginine R Arg
Asparagine N Asn
Aspartic acid D Asp
Cysteine C Cys
Glutamic acid E Glu
Glutamine Q Gln
Glycine G Gly
Histadine H His
Isoleucine I Ile
Leucine L Leu
Lysine K Lys
Methionine M Met
Phenylalanine F Phe
Proline P Pro
Pyrrolysine O Pyl
Selenocysteine U Sec
Serine S Ser
Threonine T Thr
Tryptophan W Trp
Tyrosine Y Tyr
Valine V Val

Peptides are formed by linking various amino acids together. The link between one amino acid residue and the next is an amide bond, also called a peptide bond. Polypeptides are a single linear chain of amino acids. Figure 1 shows the structure of oxytocin, a neuropeptide with nine amino acids.

Figure 1: Oxytocin. A peptide of nine amino acids linked in the sequence cysteine-tyrosine-isoleucine-glutamine-asparagine-cysteine-proline-leucine-glycine. The cysteine residues form a sulfur bridge (yellow). Black circles are carbon, blue circles are amine units, and red circles are carboxyl units.

Proteins are basically very large peptides. A general convention is that proteins are molecules with more than 50 amino acids. Although the amino acid sequence is still read out by the DNA genetic code as a linear chain, it is folded and twisted as a consequence of salt bridges, hydrogen bonds, and disulfide bonds. Proteins have complex three-dimensional structures that affect their function. Many proteins are enzymes that catalyze biochemical reactions. Others have structural or mechanical functions, such as the proteins in the cytoskeleton that form a scaffolding system to maintain cell shape. Proteins are also important in cell signaling, immune responses, cell adhesion, and the cell cycle. Visual and olfactory sensory cells possess G protein coupled receptor proteins responsible for the critical first step in sensory transduction. Finally, binding proteins play an essential role in chemical communication, both at the production stage (transporting pheromone components from internal organs to external release sites) and at the reception stage (transporting incoming pheromone molecules through the mucus to the receptor).

Figure 2: Amino acid sequence and proposed structural model for a rat olfactory G protein coupled receptor protein anchored in a ciliary cell membrane. The 7 trans-membrane helices have been stretched out. Circles with letters indicate the different amino acids. The gray circles highlight the domains that are highly variable among different receptor cell types that bind different odorants. (From McClintock 2003.)

Lipids: fatty acids and glycerides

Lipids are carboxylic acids or esters with a long hydrocarbon tail. This tail makes them generally non-polar and thus insoluble in water (hydrophobic). Fatty acids have a single carboxylic acid at the end, and tail chain lengths between 12–24 carbons with the general formula CH3(CH2)nCOOH. Very small fatty acids are volatile and can serve as pheromones. Glycerides are esters formed from glycerol and fatty acids. They have three hydroxyl functional groups, which can be esterified with one, two, or three fatty acids to form monoglycerides, diglycerides, and triglycerides. The hydrocarbon tails may be saturated or unsaturated. Lipids are used for energy storage and some are important hormones; they form the lipid bilayer of cell membranes. The tails of fatty acids may be saturated or unsaturated. A few examples are illustrated below.

Wikipedia fatty acid.


Carbohydrates are the most common type of biomolecule. They are based on a ring unit containing 4 or 5 carbons and one oxygen. Carbohydrate molecules vary enormously in size, from one to several thousand of these units. Monosaccharides are the smallest. They contain carbon, hydrogen, and oxygen in a ratio of 1:2:1, with the general formula CnH2nOn (where n = at least 3). Disaccharides contain two units. These small carbohydrate molecules are commonly called sugars. They have a sweet taste and provide readily accessible fuel for cellular metabolism.

Oligosaccharides contain 3–10 ring units. Smaller molecules consisting of short chains of fructose units are found in fruit and can only be partially digested by humans, hence their use as sugar substitutes. Oligosaccharides are often combined with proteins to form glycoproteins. Mucins are the glycoproteins secreted in the mucus of the respiratory and digestive tracts. The sugars attached to mucins give them considerable water-holding capacity and also make them resistant to proteolysis by digestive enzymes. Some hormones are glycoproteins, such as follicle stimulating hormone, luteinizing hormone, thyroid stimulating hormone, and human chorionic gonadotropin.

Polysaccharides are increasingly larger molecules often referred to as complex carbohydrates. They contain less carbon relative to oxygen and hydrogen, depending on how they are connected. They have no taste. Cellulose is a polymer built from glucose units. Plants use it as the main structural component of their cell walls. Animals can neither manufacture it nor digest it (unless they obtain the aid of microbes). Glycogen is an animal polysaccharide used for energy storage.

Wikipedia carbohydrates.


Terpenes are a large and varied class of hydrocarbons based on the 5-carbon isoprene unit. They were once believed to be produced primarily by plants, hence the name terpene (from the word turpentine, a product derived from plant resins that contain these compounds). However, as the biosynthetic pathways became better understood, it was discovered that animals can synthesize some plantlike terpenes. Moreover, isoprenoid building blocks give rise to a wide range of metabolically important compounds. Isoprene units can be connected head-to-tail, tail-to-tail, or assembled into rings, and they can be oxygenated with different functional groups. As chains of isoprene units are built up, the resulting terpenes are classified sequentially by size as hemiterpenes, monoterpenes, sesquiterpenes, diterpenes, sesterterpenes, triterpenes, etc. The all-inclusive class of terpenoids thus includes the longer terpene hydrocarbons along with isoprene-containing compounds modified by oxidation, methylation, and cyclization. The table below shows examples of terpenoids from each of the size classes.

Wikipedia terpene.


A steroid is a terpenoid lipid containing a carbon skeleton with four fused rings. Different steroids vary in the functional groups attached to these rings. Hundreds of distinct steroids have been identified in plants, animals, and fungi. All steroids are derived either from the sterol lanosterol (in animals and fungi) or the sterol cycloartenol (in plants). Both sterols are derived from the cyclization of the triterpene squalene. The figure below shows squalene laid out in way that reveals its similarity to the steroid skeleton.

The vertebrate sex hormones are all derived from cholesterol.

Pathways for pheromone biosynthesis

A few animal species ingest and modify plant compounds to produce their pheromones. In the main text we described the use of secondary plant compounds such as pyrollizidine alkaloids to synthesize the male sex attractant in Utetheisa moths. Other moths in the Geometridae, Arctiidae, and Noctuidae families use plant-specific linoleic or linolenic acids to produce their pheromones, which characteristically have desaturated carbon chains with two or three double bonds, respectively. However, the majority of animals produce their pheromones de novo using biosynthetic pathways of normal metabolism. These products are then altered to yield species-specific pheromone components. There are three main sources of pheromones: isoprenoid hydrocarbons, fatty acids, and amino acids. We briefly review the biosynthetic pathways for the first two sources, but for further details the reader should consult Tillman (1999) and Jurenka (2004).

Isoprene pathways

Isoprene itself does not undergo the building process. Instead, activated forms, isopentenyl diphosphate (IPP) and dimethylallyl diphosphate (DMAPP), are the key building blocks (Figure 3). IPP is formed from acetyl-CoA via the the MEVALONATE pathway, which involves production of the intermediate compound mevalonic acid using HMG-CoA reductase enzymes. An alternative, totally unrelated biosynthesis pathway of IPP, called MEP/DOXP pathway, has been discovered recently in some some eubacteria, green algae, and the plastids of plants. This route is initiated from pyruvate and glyceraldehydes (C5 sugars), with 2-methyl-D-erythritol-4-phosphate (MEP) and deoxyxylulose posphate as intermediates. In both pathways, IPP is isomerized to DMAPP by the enzyme isopentenyl pyrophosphate isomerase.

Synthesis of all higher terpenoids proceeds via formation of geranyl pyrophosphate (GPP), farnesyl pyrophosphate (FPP), and geranylgeranyl pyrophosphate (GGPP). Organisms differ in the possession of these enzymes and which pathways they employ. Although plants routinely produce hemi- and monoterpenes, the latter comprising the fragrant essential oils, some insects have evolved enzymes to synthesize similar compounds. Two key pheromones of bark beetles, ipsdienol and frontalin, are derived from geranyl diphosphate. The addition of an IPP unit to GPP leads to farnesene and other C15 sesquiterpenoids in both plants and animals. Insects derive their main developmental hormone, juvenile hormone, from this compound (Dipterans are an exception, producing a steroid as in vertebrates). Most organisms, but not insects, combine two FPP molecules by their tails to produce a C30 triterpenoid, often squalene. Squalene then undergoes cyclization to yield phytosterol in plants and cholesterol (and its many important derivatives) in animals. Only plants add an IPP unit to produce a diterpenoid, and then combine two of these molecules to yield C40 tetraterpenoids such as carotene.

Figure 3: Isoprene pathway. Key building block components are shown in red, enzymes in blue, and major terpenoid classes comprising increasing numbers of isoprene units in green. Plants and animals can both produce the key double isoprene compound geranyl-PP, but do so via different pathways: animals through the mevalonate pathway on the left, and plants through the MEP/DOXP pathway on the right. Major signaling compounds produced via this pathway include juvenile hormone, the bark beetle pheromones ipsienol and frontalin, farnesene pheromones in rodents and many insects, steroids in animals and plants, and carotenoids in plants. (After Dewick 2001; Seybold and Tittiger 2003.)

Fatty acid pathways

The pheromones of many Lepidopterans are based on fatty acids and make use of the enzymes that produce normal fatty acids found in all organisms. The basic fatty acids—palmitic acid (16-carbon) and stearic acid (18-carbon)—are constructed from the base substrate acetyl coenzyme A (CoA). Malonyl-CoA then provides 2-carbon units to elongate the chain in steps. Fatty acids always have an even number of carbons. These two basic saturated fatty acids then undergo a series of modifications to generate the species-specific pheromone (Morse and Meighen 1986). The steps may occur in any order. The saturated fatty acid acquires a double bond at some point along the chain under the control of a desaturase enzyme. The chain is often reduced in length in 2-carbon steps under the control of Acyl-CoA oxidase. The compound then typically is converted to an acyl alcohol via a reductase step. The alcohol may then be converted to an aldehyde via an oxidase enzyme or to an acetate ester via an acetyl transferase enzyme. A large variety of mono-unsaturated compounds with different double bond positions, chain lengths, functional groups, and stereoisomers can be generated from this basic process (Figure 4).

Figure 4: Fatty acid pathway. (A) Chain elongation with acetyl-CoA and concatenated additions of two carbons from Malonyl CoA yields palmitic and stearic acid, both saturated fatty acids. (B) Palmitic and stearic acid undergo chain shortening, desaturation, and reduction to an alcohol, and the functional group may be modified to aldehyde or acetate ester. (C) Production of the components of the cabbage looper moth (Trichoplusia ni) pheromone blend. Z8–12:CoA is illustrated. (After Jurenka 2004.)

Literature cited

Dewick, P. M. 2001. Medicinal Natural Products. West Sessex, UK: John Wiley and Sons.

Jurenka, R. 2004. Insect pheromone biosynthesis. Topics in Current Chemistry 239: 97–132.

McClintock, T. S. 2003. Molecular biology of olfaction. In The Neurobiology of Taste and Smell (Finger, T. E., W. L. Silver, and D. Restrepo, eds.), pp. 179–199. New York: Wiley-Liss, Inc.

Morse, D. and E. Meighen. 1986. Pheromone biosynthesis and role of functional groups in pheromone specificity. Journal of Chemical Ecology 12: 335–351.

Seybold, S. J. and C. Tittiger. 2003. Biochemistry and molecular biology of De Novo Isoprenoid pheromone production in the Scolytidae. Annual Review of Entomology 48: 425–453.

Tillman, J. A., S. J. Seybold, R. A. Jurenka, and G. J. Blomquist. 1999. Insect pheromones—an overview of biosynthesis and endocrine regulation. Insect Biochemistry and Molecular Biology 29: 481–514.

Wyatt, T.D. 2014. Pheromones and Animal Behaviour: Chemical Signals and Signatures. Cambridge: Cambridge University Press.

6.3 Marking Behaviors and Displays


Overt scent marking behaviors are much more conspicuous in terrestrial than in aquatic animals. Aquatic animals do release pheromones, but they are more likely to exude these over much of their external body surface instead of from a dedicated organ (e.g., many non-arthropods), or pump them into excurrent respiration channels (e.g., crustaceans). As noted in the text, sea hares (Aplysia spp.) scent mark their fertilized egg massses while laying them, and barnacle larvae leave pheromonal “footprints” while seeking a site to settle. Many fish release pheromones in their urine, which being liquid like the surrounding water, is not visually conspicuous. As a result, most videographers have paid little attention to the pheromone dispersing actions of aquatic animals and we were unable to find many examples for this module.

In contrast, terrestrial species often have specialized organs and behaviors that they employ to disperse odorants into air. Many male moths have eversible brushes and hair-pencils that they can wave to increase pheromonal dispersal (Birch and Poppy 1990). Most wasps, ants, and bees have specialized abdominal, foot, or head glands that they use to mark substrates. Some lizards rub femoral glands inside each thigh on substrates to defend territories and attract females. Urine and feces are often used by mammals as vehicles for pheromonal compounds; they also have pheromonal glands on the head (antelopes), wrists (lemurs), chests (koalas), and near the anus or genitals (most carnivores). Nearly all of these species have specific behaviors that they use to deposit urine, feces, and/or gland secretions on specific sites as territorial markers or during courtship to females. These behaviors are quite conspicuous, and it was thus much easier to find good video examples of pheromone dispersal for terrestrial taxa.

(Note: Windows users may need Quicktime installed to see some examples; Mac users may need to install Flip4Mac [free version] to view some Windows-based formats).

Examples of pheromonal dispersal behaviors and displays

Below, we provide short descriptions of marking behaviors in a variety of taxa and links to corresponding videos. See Chapters 6, 12-13 for more details on the contexts in which these behaviors are performed.


  • Combtooth blenny (Ecscenius bicolor): Nesting males of the blenny Scartella cristata have a large gland near their anal fin (Neat et al. 2003). It is not clear how the pheromone is applied, but the gland is enlarged only in nesting and hole-defending males; sneaker males have only a rudimentary gland. This video shows a nesting male of a related species repeatedly entering and exiting its nesting hole in a way that could apply pheromones to the substrate, in addition to fanning the eggs.


  • Indian luna moth (Actias selene): In this species, the female releases pheromones from the tip of her abdomen that attract males. Whereas this female lacks the scent-pencils or brushes of males in other species, the source gland is under pressure and the video shows the forced extrusion of pheromone.
  • Leaf-cutter ants (Atta cephalotes): Colonies of this species cut fresh leaves and return them to their nests as nutrient for their cultivated fungus gardens. Ants follow pheromone traits to and from current leaf sources. Based on numbers and quality of leaves being brought back, the trails are reinforced with pheromones by the tiny workers seen here not carrying leaves.
  • Army ants (Eciton burchellii): Raiding army ants stream out in all directions from their current bivouac, and then kill, dismember, and return parts of encountered prey (largely other arthropods for neotropical species) to the colony. Streams follow pheromone marked trails with outgoing ants in lanes on each side of the stream, and returning ants in the center lane (see Figure 14.30, Chapter 14). Incoming ants do most of the trail marking. Having three lanes instead of two minimizes the tendency for outgoing ants to keep veering towards the incoming lane (where trail pheromone is strongest), and thus pushing the incoming lane further and further to one side.
  • Euglossine bees scrape fragrant oils from flowers and other plant material and store it in their enlarged hind tibia. Males establish small territories and display to females in a hovering dance while dispersing the scent from their legs. This website from the laboratory of Thomas Eltz shows how males pull the scent out of their tibial pockets and disperse it by fanning with their wings.




Chiroptera (bats)

  • White-lined bat (Saccopteryx bilineata): Each afternoon, adult males refill the sac in each wing with urine and secretions from several glands. Later, males hover in front of females in the male’s territory, snap open the sacs to release odorant that is blown into the female’s face by the draft of the wings, and exchange vocalizations with the female. Males will also open the gland on one wing when roosting and shake it at either other males or females.








  • African elephant (Loxodonta africana): adult male in full musth; notice wet inside of rear legs from dribbling urine and dark streaks from temporal glands on head.

Literature cited

Birch, M.C. and G.M. Poppy. 1990. Scents and eversible scent structures of male moths. Annual Review of Entomology 35: 25–58.

Neat, F.C., L. Locatello and M.B. Rasotto. 2003. Reproductive morphology in relation to alternative male reproductive tactics in Scartella cristata. Journal of Fish Biology 62: 1381–1391.

6.4 Chemical Transmission Models

This unit provides the quantitative details of the chemical transmission models presented graphically in the text. It includes the classic work by Wilson and Bossert from the 1960s on diffusion principles and design rules for chemical signals, which we now know is limited in its applications to very small organisms. For most animals communicating over even moderate spatial scales, current flows overwhelm the action of diffusion. The unit also briefly describes the history of attempts to model the spread of chemicals under current flow conditions.

Reynolds number

The Reynolds number, named after Osborne Reynolds (1842–1912) who first proposed it, is a dimensionless number that estimates the tendency of a flowing medium to develop a turbulent versus laminar flow pattern, based on the medium’s viscosity and density. More specifically, it is the ratio of inertial forces to viscous forces in the flowing medium, so Reynolds number quantifies the relative importance of these two forces for given flow conditions. Quantitatively:

whereU = mean flow velocity
L = characteristic length
μ = absolute dynamic fluid viscosity
ρ = fluid density
ν = kinematic viscosity = μ/ρ

Inertial forces can be envisioned as the tendency of an object to continue moving when given a push, or its momentum. In fluid mechanics terminology, this is expressed as (mean flow velocity times the medium’s density). Note that the term fluid can refer to either air or water. Viscosity is the stickiness of the medium, or how difficult it is for one molecule to slide past its neighbors, and is expressed as μ/L (the force required to push an amount of medium over a specified distance in a specified period of time). The ratio of viscosity to density, μ/ρ, is an important variable in fluid dynamics called kinematic viscosity, ν. The Reynolds number is essentially the inverse of v times the mean or “bulk” flow velocity of the medium over some specified distance or area. Laminar flow occurs at low Reynolds numbers (Re < 2000), in which viscous forces dominate, and is characterized by a smooth, even flow at all layers. Turbulent flow occurs at high Reynolds numbers (Re > 3000), in which inertial forces dominate, producing eddies, vortices, and other random flow variations. Water is 70 times as viscous as air, but about 830 times as dense, so its kinematic viscosity is lower by a factor of 8–15. Thus water currents will develop turbulence at lower velocities and shorter distances compared to wind flow in air (Denny 1993). Figure 1 illustrates the boundary conditions between laminar and turbulent flow in air and water and the tradeoff between distance and speed at which this switch occurs.

Figure 1: Conditions for laminar and turbulent flow in air versus water. The relative distances and flow speeds at which flowing currents in air and water will become turbulent. The broad band represents the uncertainty over whether turbulence will develop at intermediate Reynolds numbers. Characteristic length on the y-axis is shown in both log units (left) and untransformed metric units (right). (After Dusenbery 1992.)


Molecules move down their concentration gradients in a process called diffusion. The rate of diffusion depends upon: (1) the steepness of the concentration gradient, and (2) the ease with which a particular type of molecule moves in a particular medium. The slope of the concentration gradient can be described as the change in concentration of a molecule over a given distance. The diffusion constant is a measure of the ease of movement of a particular molecule type in a particular medium; it depends on the size of the molecule, how the molecule interacts with the medium, and how the medium molecules interact with each other. Fick’s first law provides a quantitative description of the rate of movement of molecules diffusing through a small window per unit time. Armed with Fick’s Law, we can predict the concentration of diffusing molecules at any time and at any distance away from the source of the molecules for a variety of emission strategies and ecological conditions. Quantitative expressions for these processes are described below. In the following discussion of chemical signal transmission strategies, we shall refer frequently to the following quantitative variables:

C = concentration of odorant molecules per unit of volume (molecules/cm3)
K = minimum concentration of odorant molecules required for receiver detection
Q = total number of molecules released
D = medium-specific diffusion constant for a given type of molecule (cm2/sec)
r = distance from the chemical release source for circular transmission (cm)
x = distance from the chemical source for longitudinal transmission (cm)
t = time from the onset of emission (sec)

Law of Diffusion

Fick’s first law describes the rate of movement of molecules down a concentration gradient. This rate depends on the diffusion constant D and the slope of the concentration gradient. If the concentration of the molecules at any point x is C(x), then the slope of the gradient is the derivative of C(x) with respect to x, or

The number of molecules diffusing through a small window per unit time, J, is therefore given by:

where J is in units of molecules/cm2 × sec, D is in units of cm2/sec, C is in units of molecules/cm3, and x is in cm. The sign of the concentration is negative since molecules are moving from a higher to a lower concentration region as x increases. Adding the minus sign to the expression therefore makes J a positive number of molecules moving down the gradient.

Single Puff in Still Air

If a single instantaneous puff of odorant is released quickly from a point well away from any boundaries, the odorant will diffuse outward in all directions. Using Fick’s Law, it can be shown that the concentration C (r,t) at any distance r and time t is:

where Q is the number of molecules released and D is the diffusion constant. The variable d is a dimensionality constant: d = 1 for diffusion occurring in one dimension as in a pipe, d = 2 for diffusion occurring in two dimensions away from a point, and d = 3 for diffusion in three dimensions from a point source (the usual case for animal signals). If the animal is on the ground, the same number of molecules must diffuse into half as much space, so C is then twice that predicted above for each r and t (Sutton 1953; Wilson and Bossert 1963; Dusenbery 1992). The active space is the region in which the concentration C is equal to or larger than the detection threshold concentration K. The size (radius) of the active space, rA, can be computed at any time t as follows:

If we plot the profile of odorant concentration with distance from the source in a series of time snapshots, as shown in Figure 2, we would see a spherical (or hemispherical) cloud that spreads outward in time, and then shrinks back down as the odorant diffuses away.

Figure 2: Spread of a single olfactory puff. Each graph shows a snapshot of odorant concentration (C) versus distance from the source (r) at different times since emission (t). At the instant the puff is produced (t0), all of the odorant is highly concentrated at the point of release (r0). At each subsequent time interval, the odorant molecules diffuse out from the source and the local concentration (molecules per unit volume) is reduced. The active space, rA, is the enclosed region in which the concentration of the odorant is above the threshold detection concentration, K. The active space first increases, then decreases to zero.

The maximum radius of the active space when the puff is released near the ground can be calculated as:

The time required for the single puff to expand to rmax is:

Finally, the time to fadeout of the signal is:

The relative change in rmax and the time points are illustrated in Figure 3.

Figure 3: Change in the radius of the active space of a single puff over time. The active space increases and reaches maximum size at the time trmax (dashed line), and then decreases to zero. The time to reach rmax is always 0.37 times the time to fadeout, or tfadeout = 2.72 trmax. (After Bossert and Wilson 1963.)

Continuous Emission in Still Air

If the sender continuously emits Q molecules/sec from a surface position, the concentration is given by:

where erfc(x), the error function complement, is the area under the normal curve out to infinity. Here, rmax increases and levels off at a value of approximately (Figure 4):

where D is measured in units of distance moved per second. The time to reach 95% of rmax is:

Figure 4: Time to achieve maximum radius with continuous emission of an odorant. Curve shows radius of the threshold sphere as a function of time. (After Wilson and Bossert 1963.)

Continuous Release From a Moving Source

The case of an animal releasing a trail of odorant while moving can be modeled as a linear series of single puffs (Figure 5). If Q is the number of molecules of pheromone released per second and u is the animal’s velocity, the total length of the active space, L, will be

The location of the maximum radius of the active space occurs 0.37 along the active space axis from the point near the animal:

The maximum diameter of the active space at this location is

where e is the base of the natural logs. The time it takes before the trail at any given location has dropped below K is

Figure 5: Active space of an ant trail. A trail is essentially a sequential series of small single puffs along a linear transect. L is the length of the active space, rmax is the width of the active space, and Lrmax is the distance to the maximum radius point from the ant. (After Bossert and Wilson 1963.)

Advection in a current flow

In the discussion below we will use the Cartesian coordinate system to reference the spatial dimensions of the flow. The source of chemical release occurs at the origin: downstream distance is specified by x, horizontal spreading perpendicular to the flow direction by z, and vertical distance by y.

Laminar Flow

When there is laminar current flow, the spread of a continuously emitted chemical can be modeled with the same logic as the trail above with a moving source, but in this case the source is stationary and the center of the active space moves with the flow. The concentration of the substance at any point in space is given by

where U is current velocity, r is the straight-line distance from the source to the position of interest, θ is the angle between this line and the downstream direction, and D is the diffusion constant. In this model, the maximum distance of odorant transmission is not actually increased in the downstream direction compared to a flow velocity of zero, but transmission is faster and the active space is narrower (Figure 6). This type of model would only be relevant for small animals living in the viscous boundary layer next to a surface and communicating over short distances—several centimeters in air and even less in water. Most animals must cope with turbulent flow conditions.

Figure 6: The active space for a constantly emitted source in a laminar flow field. Concentration intensity decreases as 1/r with distance from the source. (After Dusenbery 1992.)

Turbulent Flow

In a turbulent flow, there is a range of scales of the fluid motion. A single packet of fluid moving with a bulk velocity is called an eddy. The size of the largest eddies is set by the overall geometry of the flow, and the size of the smallest eddies is set by the Reynolds number. As the Reynolds number increases, smaller and smaller eddies occur. The Reynolds number is therefore an indicator of the range of eddy size scales in the flow. Under large Re conditions, inertial forces predominate over viscous forces, and the smallest scales of fluid motion are undamped—there is not enough viscosity to dissipate their motions. The kinetic energy must “cascade” from these large scales to progressively smaller scales until a level is reached for which the scale is small enough for viscosity to become important (i.e., viscous and inertial forces become approximately equal). It is at these small scales where the dissipation of energy by viscous action finally takes place. Therefore, although the energy dissipation is produced by a viscous mechanism, the rate at which it occurs is dictated only by large-scale characteristics of the flow, while viscosity only determines the size of the smallest eddies at which the energy will be dissipated (from (

Figure 7: A dark-field photograph of smoke rising from a cigarette. This photo illustrates several important points. First, it shows the initial laminar flow over a short distance and the turbulent distribution of particles over larger distances. Second, it shows the wide size range of vortices and eddies. One can see large-scale meandering of the primary plume, as well as fine-scale eddies and filaments.

Sutton Model

To cope with the variable and random pattern of odorant density within turbulent plumes, modelers attempted to estimate a smoothed plume shape by averaging the patchiness over time. The first such time-average model was developed by Roberts 1923 and Sutton 1953. They used the same principle of laminar current flow described above but redefined the diffusion constant as the rate of molecular advection by the current flow. The model specifies the average concentration of a continuously released odorant at points downwind from the source as:

where Dy and Dz are diffusivity constants in the y and z planes measured empirically at a wind speed U, Q is release rate in molecules/sec, and n is a parameter ranging from 0 to 1 determined by wind speed, atmospheric conditions, and terrain slope (typically n = 0.25). The 2 in the numerator again reflects the fact that the odorant is released from the ground. An application of this model to animal signalers (moths) is shown in Figure 8, where a threshold concentration K was incorporated to determine the active space, and the maximum downwind detection distance was estimated as:

Figure 8: Active space for an olfactory signal released into wind of different velocities. Active space dimensions estimated from the Sutton are given. Transmission distance is better for low wind speeds compared to higher wind speeds where molecules are whisked away from the surface faster. (After Wilson and Bossert 1963.)

Gaussian Model

The Sutton model did not seem to capture the way odor plumes spread out in a wedge as distance from the source increases. The Gaussian time-average model was developed to remedy this short-coming by evaluating horizontal and vertical spreading as a function of distance and wind speed. The Gaussian model is largely the same as the Sutton model, except the diffusivity constants Dy and Dz are replaced by σy and σz, standard deviations of the cloud dimensions in the horizontal and vertical directions, respectively:

and the standard deviation terms are functions of downwind distance:

This model has also been modified to take into account the release of odorant from some specified height off the ground, and to include the absorption of molecules on the ground surface. The Gaussian model does yield a more broadly spreading plume (Figure 9), but is it still often not as wide as real measurements of active spaces in the field (Fares et al. 1980; Elkinton et al. 1984; Murlis et al. 1992).

Figure 9. A Gaussian plume. It has a wider spread at increasing distances from the source than a plume estimated via the Sutton model.

Meandering Plume Model

Another step toward reality involves incorporating the meandering path of the main plume. Such models use the same basic logic as the Gaussian model, but allow the center of the clouds to meander with distance from the source (Figure 10).

Figure 10. Increasing complexity of plume models. (A) The basic Gaussian plume model, in which the spread of the cloud increases with distance from the source, estimated with a Gaussian standard deviation. (B) A meandering plume model, where the clouds both increase in size and follow a meandering path as they move away from the source. (C) The filamentous structure of a real plume, showing the three-dimensional regions in which the plume may meander. (After Murlis et al. 1992.)

Computational Models

The most recent attempts to generate realistic models of turbulent plumes involve statistical models and simulations of the cascading eddy behavior of turbulent flows. Figure 11 shows an example of such a model. The model uses a direct numerical simulation (DNS) technique to compute momentum with continuity constraints, taking into account the conservation of energy and the incompressibility of the medium. The model is able to produce the plume meandering pattern observed in an unstably stratified open channel flow (Liu and Leung 2006a, 2006b).

Figure 11. A direct numerical simulation (DNS) model of turbulent flow in a neutrally stratified open channel. The scalar transport behaviors are shown for five emission heights (zs = 0, 0.25H, 0.5H, 0.75H, and H, where H is the channel height) at a Reynolds number of 3000, a Prandtl number and a Schmidt number of 0.72, and a Richardson number of 0.2. The vertically meandering mean plume heights and dispersion coefficients calculated by the DNS model agree well with laboratory results and field measurements. The plume meandering is caused by the movement of the positive and negative vertical turbulent scalar fluxes above and below the mean plume heights, respectively. (From Liu and Leung 2006a.)

There are several other approaches to modeling turbulent flows, including: the Reynolds-averaged Navier-Stokes (RANS) approach, which provides a time-averaged solution like the Gaussian model; large eddy simulation (LES) approach, which removes the smallest scales of the flow to focus on the major eddy forms; the detached eddy simulation (DES), which focus on the small-scale eddies; coherent vortex simulation, which decomposes the flow into the coherent vortex motion component and the incoherent background flow using wavelet filtering; and the Reynolds stress model (RSM), which attempts to solve transport equations for the Reynolds stresses. See recent books on this topic listed in Further Reading below.

Useful websites on turbulence

Further reading

Durbin, P. A. and B. A. Pettersson-Reif. 2011. Statistical Theory and Modeling for Turbulent Flows, 2nd ed. West Sussex, UK: Wiley.

Mathieu, J. and J. Scott. 2000. An Introduction to Turbulent Flow. Cambridge: Cambridge University Press.

Moore, P. and J. Crimaldi. 2004. Odor landscapes and animal behavior: tracking odor plumes in different physical worlds. Journal of Marine Systems 49: 55–64.

Pope, S. B. 2000. Turbulent Flows. Cambridge: Cambridge University Press.

Literature cited

Bossert, W. H. and E. O. Wilson. 1963. Analysis of olfactory communication among animals. Journal of Theoretical Biology 5: 443–469.

Denny, M. W. 1993. Air and Water: The Biology and Physics of Life’s Media. Princeton, NJ: Princeton University Press.

Dusenbery, D. B. 1992. Sensory Ecology. New York: W. H. Freeman and Company.

Elkinton, J. S., R. T. Carde, and C. J. Mason. 1984. Evaluation of time-average dispersion models for estimating pheromone concentration in a deciduous forest. Journal of Chemical Ecology 10: 1081–1108.

Fares, Y., P. J. H. Sharpe, and C. E. Magnuson. 1980. Pheromone dispersion in forests. Journal of Theoretical Biology 84: 335–359.

Liu, C. H. and D. Y. C. Leung. 2006a. Finite element solution to passive scalar transport behind line sources under neutral and unstable stratification. International Journal for Numerical Methods in Fluids 50: 623–648.

Liu, C. H. and D. Y. C. Leung. 2006b. Turbulent transport of passive scalar behind line sources in an unstably stratified open channel flow. International Journal of Heat and Mass Transfer 49: 4305–4324.

Murlis, J., J. S. Elkinton, and R. T. Carde. 1992. Odor plumes and how insects use them. Annual Review of Entomology 37: 505–532.

Roberts, O. F. T. 1923. The theoretical scattering of smoke in a turbulent atmosphere. Proceedings of the Royal Society A 104: 640–654.

Sutton, O. G. 1953. Micrometerology. New York: McGraw Hill.

Wilson, E. O. and W. H. Bossert. 1963. Chemical communication among animals. Recent Progress in Hormone Research 19: 673–716.

6.5 Vertebrate Dual Chemosensory System


Most amphibians, reptiles, and mammals have a dual olfactory system: the main olfactory system (MOS) with input through the main olfactory epithelium (MOE) and neural connections to the main olfactory bulb (MOB), and the vomeronasal system (VNS) with input through the voneronasal organ (VNO) and neural connections to a separate accessory olfactory bulb (AOB). Vertebrates have other chemical sensors as well, including taste bud receptors on the tongue, the septal organ of Masera, the Grueneberg ganglion, and the trigeminal nerve system (Table 1, Figure 1). It was once thought that the VNS’s function was restricted to reception of pheromones and the MOS’s function was mainly detection of food, predators, and other ambient odors, but this division is not at all so distinct. Some pheromones are detected only through the MOS, some species follow prey using VNS input, and there are many examples in which input from both systems is integrated. The VNS and MOS differ in three main ways: (1) the receptor proteins and their genes, (2) the specificity of these receptors to different stimuli, and (3) the connection patterns between sensory cells, glomeruli, olfactory bulbs, and brain processing centers. We will first describe these differences in more detail, and then review recent findings on the relative roles of the two systems. This overview draws from several recent reviews which should be consulted for more details (Dulac and Torello 2003; Wyatt 2014; Brennan and Zufall 2006; Müller-Schwarze 2006; Ma 2007; Touhara and Vosshall 2009; Su et al. 2009; Munger et al. 2009; Tirindelli et al. 2009; Mucignat-Caretta 2010; and Ferro and Liberles 2010).

Table 1. Chemoreception systems in vertebrates.

Name Location Presumed function
Main olfactory system (MOS) Nasal cavity Detection and discrimination of a large range of ambient odorants from food, predators, and conspecifics.
Vomeronasal system (VNS) Between nasal cavity and mouth Detection of water-borne (also urine and mucus) pheromones and proteinaceous prey/predator chemical cues.
Taste buds Tongue and mouth Discrimination of sweet, sour, bitter, salty, and savory (umami) chemicals in potential food items.
Trigeminal nerve (TN) Innervates the face and eyes Responsive to touch, pain, temperature, and irritating and noxious chemical stimuli
Nervus terminalis (NT) From hypothalamus to nasal epithelium Modulation of olfactory epithelium in response to hunger state
Septal organ of Masera (SOM) Base of nasal cavity near entrance to nasopalatine ducts Modulation or synchronization of responses of olfactory nerves with respect to respiratory air flow
Grueneberg ganglion (GG) Dorsal to nares Detection of alarm pheromone, functional in neonate pups.
Guanylyl cyclase-D cells
Dispersed ciliary cells in olfactory epithelium (mice) Responsive to uroguanylin, guanylin, cues in urine, CO2 connections to necklace ganglia.

Sources: Breer et al. (2006); Su et al. 2009; Munger et al. (2009); Fleischer et al. (2009).

Receptor proteins and their genes

The majority of receptor cells in the MOE have G-coupled receptor proteins and are therefore called G-protein-coupled receptors (GPCRs). The genes that encode these receptors belong to large multi-gene families. Most genes specific to the olfactory system are referred to as ORs (which stands for Olfactory Receptor gene). Vertebrate species often have up to 1,000 different OR gene variants, originating from gene duplication events, and sensitive to different medium-borne chemicals as a consequence of different amino acid variation in the ligand-binding regions (Buck and Axel 1991). A second family of receptor genes present in the MOE, called trace-amine-associated receptors (TAARs), are sensitive to biogenic amines (Liberles and Buck 2006; Hashiguchi and Nishida 2007). Both classes of genes are found throughout the vertebrates, including fish. A third class of receptors, sparsely distributed throughout the MOE of most vertebrates but not in primates, comprises the guanylyl-cyclace-D (GC-D) system; these cells project axons to the large necklace glomeruli in the MOB.

Figure 1: The mouse dual olfactory system (sagittal cross section through the nasal region of the head, lower jaw is not shown). The MOE and MOB are shown in green, and the two layers of the VNO and corresponding connection regions in the AOB are shown in yellow and red. SOM = septal organ of Masera, NG = necklace glomeruli, NC = nasal cavity, GG = Grueneberg ganglion, GCD = guanylyl cyclase-D system. (After Brennan and Zufall 2006; Tirindelli et al. 2009.)

There are two main classes of receptor genes in the VNO, referred to as V1R and V2R. Both are also multi-gene families from the GPCR superfamily. The sensory cells expressing these two types of genes are segregated into two layers of the vomeronasal organ, and they send their axon projections to two distinct regions within the AOB (Figure 1). V1R-expressing sensory cells are located in the apical half of the VNO (yellow area), and V2R-expressing cells occur in the basal half (red area) (Johnston 2000; Halpern and Martinez-Marcos 2003). A final type of receptor cell found in a restricted area of the basal VNO expresses a formyl peptide receptor (FPR), which mediates cell responses to disease as part of the general immune system and in the VNO may confer sensitivity to disease-related molecules (Liberles et al. 2009; Riviere et al. 2009). This cell type is absent in primates (Yang and Shi 2010).

Figure 2 illustrates some of the key differences among the receptor proteins. OR, TAAR, V1R, and V2R proteins are all 7-transmembrane G protein-coupled receptors embedded in the membranes of the cilia (OR) or the microvilli (VRs) of the bipolar sensory cells. The G protein for OR and TAAR receptor cells is distinguished as Gαolf. These cells employ cyclic nucleotides (cAMP) as a second messenger to open cyclic nucleotide gated ion channels (CNG) and facilitate the signaling cascade for nerve impulse generation. In contrast, VRs (and a few ORs) employ a unique gated cation channel to facilitate or modulate transduction, called a transient receptor potential (TRP) channel. The V1R protein superficially appears very similar to the OR protein, although they do not share any sequence motifs. Its G protein is called Gαi2. The V2R receptor molecule is quite distinct, with a very large hydrophobic N-terminal extracellular domain (Figure 2) and a different G protein, Gαo. Unlike the single-expressed-gene-per-cell rule in all other vertebrate chemical sensors, each basal VNO sensory cell usually expresses two different V2R genes. In rodents, some V2R receptor cells co-occur with a non-classical major histocompatibility complex molecule, either M10 or M1, and an associated molecule of β2-microglobulin (β2m), shown as a purple strand in the figure (Ishi et al. 2008; Leinders-Zufall et al. 2009). Classic MHC molecules are also expressed in association with β2m at the surface of most cells in the body, where they signal the presence of foreign invaders. The unique structures of V1R and V2R proteins and their associated molecules are believed to determine receptor specificity to certain types of ligands, as discussed below.

Figure 2: Five types of vertebrate olfactory receptor proteins. Illustrations show the basic shape of each type of receptor protein and some of the coupled transduction components: the ciliary or microvillar cell membrane in which the receptor is imbedded is shown in gray; the cell exterior is above, the cell interior below. All but GC-D receptors are 7-transmembrane G protein-coupled receptors, shown schematically in their bundled shape. All have a nitrogen (N) terminal on the cell exterior and a carboxylic acid (C) terminal extending into the cell interior. An FPR is not shown here, but is very similar to the V1R receptor in basic shape. (After Dulac and Torello 2003; Spehr and Munger 2009.)

The identification and sequencing of these olfactory receptor genes presents the opportunity to analyze the evolutionary relationships among chemical receptor systems. TAAR receptors belong to the same multigene family as the serotonin and dopamine receptors in the brain. The two VR gene families are completely unrelated to each other, and neither one shares any sequence homology to the OR family of genes except the regions specifying the helical structure (Ryba and Tirindelli 1997). The V1R genes share some sequence motifs with the T2R bitter taste receptors and with opsin genes (Adler et al. 2000). The V2R genes are closely related to glutamate and GABAB neurotransmitter receptors in the brain, as well as to the T1R sweet and umami taste receptors on the tongue, which all possess the large N-terminal (Alioto and Ngai 2006; Bjarndottir et al. 2005). Thus the four main vertebrate chemoreceptor gene classes arose and diverged independently, on different time scales (Grus and Zhang 2006, 2008, 2009).

All of these gene families are present in the olfactory epithelium of fish (Hashiguchi and Nishida 2006; Grus and Zhang 2006; Liberles and Buck 2006; Hino et al. 2009; Saraiva and Korshing 2011). Even though fish do not possess a vomeronasal organ, different cell types with their associated receptor gene types form layers within the olfactory epithelium. The ciliated receptor cells expressing OR-type receptor genes have cell bodies located in the basal or middle layers of the epithelium, whereas microvillous and crypt cells containing V1R or V2R receptor genes are located in middle or surface layers of the epithelium (Hansen et al. 2004; Pfister and Rodriguez 2005; Hamdani and Døving 2006). The move onto land by the amphibians is associated with a morphological subdivision between the microvillous VR receptor cells and ciliary OR cells into separate but interconnected chambers of the nasal organ. Access of odorants occurs through grooves on the sides of the external nares, even when the main air-breathing chamber is closed while underwater (Døving et al. 1993; Døving and Trotier 1998; Petti et al. 1999). Complete separation into two chambers arose with the fully terrestrial reptiles and mammals, and in most cases access of odorants shifted to the roof of the mouth (rodents re-evolved odorant access through the base of the nasal chamber). Birds, as well as marine mammals, Old World monkeys, and some bats and reptiles, have completely lost the vomeronasal organ and its associated genetic components (Figure 3).

Figure 3: Phylogenetic tree of major vertebrate groups and corresponding number of intact OR, V1R, and V2R genes (number of non-functional pseudogenes in parentheses). A blue cross indicates the loss of the entire V2R repertoire of genes, and a red circle indicates the loss of most or all of the V1R genes. (After Niimura and Nei 2006; Shi and Zhang 2007; Tirindelli et al. 2009.)

During these shifts between terrestrial and aquatic habitats, different sets of olfactory receptor genes either expanded or decreased. With the complete genomic sequencing of about a dozen vertebrate species, recent analyses of the receptor genes have revealed a fascinating history of olfactory system evolution. What is especially interesting is that the “lost” genes are still mostly present in the genome as pseudogenes, whose functionality is gone as a result of missing or disrupted sequences (pseudogenization). Figure 3 also shows the number of intact versus non-functional genes for OR, TAAR, V1R, and V2R receptor proteins in different vertebrate species. The rodents, opossum, platypus, and frog have substantial gene repertoires for all four receptor types. Many terrestrial mammals (dog and cow) have lost the V2R genes, humans and birds have essentially no VR genes and a reduced OR repertoire, and fish possess mainly OR and V2R genes (Niimura and Nei 2006; Shi and Zhang 2007; Liman 2006).

These shifts in the relative importance of different receptor types make sense in light of the different types chemicals each one is designed to detect. Finding the active ligands for specific receptor types is painstaking work, but numerous matches have now been made. OR receptors bind to a wide variety of small volatile molecules, mainly food and environmental chemicals, but also to a few species-specific pheromones. TAAR receptors respond selectively to biogenic amines, such as those found in urine that vary as a function of sex, status, and stress. V1R receptors respond narrowly to small volatile water-soluble pheromones in urine such as sulfated steroids, and in rodents to specific known male urinary pheromones. V2R receptors bind to nonvolatile peptide and protein pheromone molecules in urine and tears, and must be received by direct contact (Zhao et al. 1998; Boschat et al. 2002; Leinders-Zufall et al. 2004; Mombaerts 2004; Kimoto et al. 2005; Su et al. 2009; Tirindelli et al. 2009; Munger et al. 2009; Mucignat-Caretta 2010).

Figure 4 shows the results of an analysis of the relative gene-class repertoire sizes for some of the species shown in Figure 3. The ratio of the number of intact V1R genes to V2R genes for a given species has been plotted as a function of aquatic, transitional, and terrestrial environment (red bars). There is a preponderance of V2R genes in the two aquatic groups (fish and frogs) and a reduction or complete loss of V2R genes in favor of V1R genes in the terrestrial vertebrates (Shi and Zhang 2007).

Figure 4: Vertebrate evolutionary patterns during aquatic to terrestrial transition. Vertebrate species arrayed from aquatic (fish) to transitional (amphibians) to terrestrial (birds and mammals). Red bars show the ratio (R) of the number of vomeronasal gene types (V1R/V2R) for different species, no value is shown if the ratio is infinity (dog) or 0/0 (chicken, human). Black bars show ratio of the number of olfactory gene types (OR Class II / OR Class I) for the same species, Results suggest that V1R and Class II OR receptors may bind selectively to volatile airborne ligands, V2R and Class I OR receptors to water-soluble and potentially larger ligands. (After Shi and Zhang 2007.)

A similar analysis has been made for OR genes. Olfactory genes cluster into many related subfamilies, but a study of OR genes in the African clawed frog (Zenopus laevis) provided a critical insight (Freitag et al. 1995). Its genes clustered into two distinct groups, one closely related to fish OR genes (Class I) and the other more closely related to mammalian OR genes (Class II). This species lives in both aquatic and terrestrial environments as an adult, and its nasal chamber is divided into two pockets, one for breathing in air and the other for water (the VNO is a third pocket close to the water-breathing pocket). A toggle-like valve opens one or the other to ambient media. The air-pocket epithelium contains only cells with Class II (mammal-like) receptor genes, while the water-pocket epithelium contains only cells with Class I (fish-like) receptor genes. Figure 4 also presents the ratio of OR Class II / Class I genes for the different vertebrate species (black bars). Not surprisingly, Class I genes predominate in the fish and Class II genes in the terrestrial species, but the frog, which has both, in this case shows the terrestrial pattern (Shi and Zhang 2007). One difference in the structure of Class I and II receptors is a longer extracellular loop between several of the transmembrane helices in the Class I proteins. This structure may enable these receptors to bind with water-soluble ligands, while Class II proteins bind with hydrophobic volatile ligands (Freitag et al. 1995). A possible scenario for the VR and OR patterns is that fish use OR Class I receptors for detection of small water-soluble odorants such as food-derived amines and conspecific alarm pheromones, and V2R receptors for detection of large water-soluble molecules such as conjugated steroids, peptides, and proteins. Amphibians continued to use V2R receptors for large water-soluble molecules, but evolved new OR (Class II) receptors to detect volatile airborne odors. Terrestrial vertebrates subsequently expanded their V1R repertoire to detect small water-soluble odorants in urine.

Connection patterns

The wiring patterns of vomeronasal sensory neurons also differ significantly from that seen in the olfactory sensory neurons, further supporting important functional differences between the two systems. Differences are apparent at the first level of connection to glomeruli, the spherical neuropils in the corresponding olfactory bulbs (Figure 5). Olfactory neurons in the MOE that express a given OR receptor all project to the same one or two glomeruli in the MOB. Several thousand sensory cells converge on a given MOB glomerulus. The spatial pattern of glomeruli in the olfactory bulb is highly conserved between individuals, and adjacent glomeruli often respond to similar kinds of chemical stimuli. Each glomerulus synapses with a unique mitral cell, which projects its dendrite to the brain. Lateral inhibitory connections between mitral cells by interneurons may sharpen the tuning of stimulated glomeruli, but there is little integration of signals from different receptor types at the level of the MOB. In contrast, sensory neurons in the VNO project to multiple (10–30) glomeruli in the AOB, and each glomerulus receives sensory cell input from a few hundred neurons that express several different receptor types. However, the two zones within the VNO remain separate and send axons only to the corresponding zone within the accessory bulb, and for apical zone cells at least, neurons expressing closely related subfamilies of V1R genes converge their axons on spatially clustered glomeruli. VNO glomeruli in general are smaller and more variable in size, and their positions are variable within and among individuals, compared to MOB glomeruli. At the next level, AOB mitral cells send dendrites to multiple glomeruli as well. Clearly far more integration between sensory cell types is occurring at the primary input stages of the vomeronasal system compared to the main olfactory system (Dulac and Torello 2003; Mombaerts 2004; Dulac and Wagner 2006).

Figure 5: Glomerulus organization in the main olfactory and vomeronasal systems. (A) Each glomerulus in the main olfactory bulb (MOB) receives axons only from cells expressing the same OR receptor type, indicated by a different color, and connects to a dedicated mitral cell. (B) Sensory neurons in the epithelium of the VNO have their cell bodies segregated into separate zones (all have microvilli reaching the surface of the organ’s lumen as illustrated). Neurons with cell bodies located in the apical zone (shown in yellow) express members of the V1R family of receptors and project to multiple glomeruli in the anterior half of the accessory bulb (AOB). Neurons with cell bodies in the basal zone (shown in red) express V2R receptors and project to multiple glomeruli in the posterior half of the AOB. (After Dulac and Wagner 2006.)

Brain circuits also differ for the main olfactory and vomeronasal systems (Figure 6). Input from the MOE and MOB projects mainly into the higher cortical areas of the brain and the lateral amygdala. Input from the VNO and AOB projects primarily to the medial and anterior amygdala, which along with the hypothalamus comprises the limbic system. The limbic system controls reproductive, aggressive, and parental care behaviors. The hypothalamus is a critical control center, integrating internal and environmental cues, ensuring organismal homeostasis, and orchestrating long-lasting endocrine changes as well as short-term behavioral effects elicited by chemical signals (Yoon et al. 2005; Martinez-Marcos 2009; Mucignat-Caretta 2010).

Figure 6: Brain pathways for the vomeronasal and main olfactory systems. Major brain regions shown include amygdala (orange), cortex (green), and hypothalamus (dard violet). The vomeronasal system (VNO, AOB, and red pathways) projects to several areas of the amygdala, as well as to bed nuclei of the accessory olfactory tract and stria terminalis. The main olfactory system (MOE, MOB, and blue pathways) projects primarily to several cortical areas, including the piriform cortex, olfactory tubercle, enthorhinal cortex, anterior olfactory nucleus, and tenia tecta (not shown); these nuclei then project to higher levels in the brain, as well as back to the main olfactory bulb (not shown). The olfactory system also connects to several nuclei in the amygdala. Areas in light purple indicate nuclei that receive input from both the VNO and MOE. Importantly, both pathways connect to the hypothalamus (medial preoptic area and ventromedial hypothalamus). (After Dulac and Wagner 2006; Tirindelli et al. 2009.)

The main olfactory system and VNO system are clearly designed for different tasks. Olfactory receptors show broader selectivity over a large range of concentrations and there are a large number of different sensory receptor types in the MOE, so they can collectively perceive a huge range of odorants. The mitral cells are uniglomerular and receive inhibitory connections on lateral dendrites, so they can compare and analyze the pattern of inputs from many glomeruli and process the information in higher cognitive brain centers. The main olfactory system thus seems designed to detect and analyze a wide range of environmental chemical compounds. Nevertheless, some single-chemical pheromones shown to trigger specific behavioral responses operate through the MOE system, so the main olfactory system, with its uniglomerulus input pattern, can operate in a labeled line fashion (see discussion of labeled line versus across-fiber pattern strategies for neural coding in main text Chapter 6). Subsets of the MOE may possess specialized circuits for detecting critical odorants and pheromones, such as the TAAR receptors that detect stress products in urine, the GC-D receptors that detect several peptides and CO2, and the Gruenberg ganglion neurons that may respond to an alarm pheromone (Stowers and Logan 2010). The vomeronasal system possesses fewer, more narrowly tuned, and very sensitive sensory cells. The mitral cells are multiglomerular and interconnected in ways that lead to integrative processing at early stages. The VNO system projects to lower levels in the brain that regulate genetically programmed behaviors and autonomic responses. This system seems to be designed in part to analyze and compare blends of structurally-related compounds involved in individual, sex, status, and species recognition, and aspects of reproductive behavior (Dulac and Wagner 2006; Trindelli et al. 2009).

Relative roles of vomeronasal and main olfactory systems

As previously mentioned, the classical view of the VNS as a pheromone detection system and the MOS as a general odorant detection system is now regarded as far too simplistic. In the prior sections, we outlined a few new views on the functions of these two systems, one a phylogenetic view arguing that both systems had components designed to operate in terrestrial versus aquatic environments, and a neural structure view arguing that the MOS is designed to detect and analyze volatile environmental odors and single-component pheromones while the VNS is designed to compare chemical blends. In this section we examine the evidence for different behavioral effects mediated by the two systems. Over the past few decades, numerous studies have evaluated the consequences of “removing” one or the other system. A system can be blocked by plugging or stitching closed the input ducts, by severing the nerves, or more recently, by genetically “knocking out” a specific second messenger transduction component, such as the TRP or a CNG gated ion channels. The emerging picture is that some effects are strongly dependent on the VNS, others on the MOS, and quite a few either require both systems or can be compensated for by the other one. See Table 2 in Tirindelli et al. (2009) for a summary of VNS- and MOS-removal studies.

Main role by VMS

Modulation of estrous in rats and mice is strongly affected by input of urine signals from males and other females through the vomeronasal organ (Ma et al. 1999; Keverne 1999; Musignat-Caretta 2010). These priming effects include acceleration of puberty in young females, pregnancy block produced by the odor of a strange (non-sire) male, and suppression of estrous by the odors of group-living females. Table 2 below summarizes the probable pheromones involved in mediating these effects in mice. Most of the volatile pheromones are MUP ligands. In prairie voles and Mondelphis opossums, the VNS is required to induce estrous. Some aspects of maternal behavior in rats, mice, and voles are also impaired by VNO removal, including maternal aggression, recognition of offspring, retrieval of pups, licking, and lactation.

Normal reproductive behaviors in a large number of male vertebrates, such as sexual arousal and copulatory behavior, are dependent on VNO input, although VNO-knockout males will still mate and prior mating experience before VNO removal can sometimes compensate. Male ungulates, carnivores, rodents, and snakes detect estrous females via chemical signals in their urine or skin secretions. In rodents, especially mice, VNO removal causes reduced aggression towards other males and copulation attempts towards both males and females. The reason for these effects seems to be that males (as well as females) require VNO input to distinguish individuals, assess status, and identify maleness. Many of the pheromones responsible for these effects are the same as those that affect female estrous.

Table 2. The structure and function of mouse urinary pheromones

Name Chemical structure Origin Possible signal function in female mice Possible signal function in male mice
2,5-Demethyl-pyrazine Female urine Suppression of estrous cycle  
2-Sec-butyl-4,5-dihydrothiazole Male bladder urine Estrus synchrony
Puberty acceleration
Male aggression, female attraction
2,3-Dehydro-exo-brevicomin Male bladder urine Estrus synchrony
Puberty acceleration
Male aggression, female attraction
α- and β-Farnesenes Male preputial gland Puberty acceleration Male territorial status, female attraction
2-Heptanone Female or male urine Estrus extension  
6-Hydroxy-6-methyl-3-heptanone Male bladder urine Puberty acceleration  
n-Pentyl acetate Female or male urine Suppression of estrous cycle  
Isobutylamine Male urine    
Methylthio-methanethiol Male urine   Female attraction
Major urinary proteins   Female or male urine Puberty acceleration,
individual recognition
Indiv. recogn., heterozygosity
MHC class I peptides   Female or male urine   Attractiveness to different strains

Sources: Ma et al. 1999; Novotny et al. 1999, Leinders-Zufall et al. 2000; Hurst and Beynon 2004; Thom et al. 2008; Trindelli et al. 2009.

Snakes use VNO input to mediate conspecific interactions such as copulation and aggressive behavior, as well as for trailing of conspecifics, aggregation, and shelter selection. In addition, they also employ this modality for prey discrimination and location. Strike behavior is reduced in rattlesnakes without a functioning VNO, and they fail to ingest envenomated prey. They also fail to recognize and take defensive action against predatory king snakes. Lizards and salamanders probably also use the VNO for prey detection. These non-conspecific chemical cues are often proteinaceous compounds inadvertently deposited as footprints by the prey (Halpern and Kubie 1984).

Main Role by MOS

The primary function of the main olfactory system certainly is detection of general odorants, but there are several very clear examples of pheromones detected only through this channel and not the VNO. One example is the rabbit mammary pheromone that guides pups to the nipple. The young of the European rabbit (Oryctolagus cuniculus) are feed only once per day for a short period of time, and when the female appears the pups must respond quickly and find a nipple. A pheromone emitted from the female’s nipple area and detected only by the MOS guides and stimulates the pups to search for the nipple (Schaal et al. 2003; Luo 2004; Moncomble et al. 2005). The second example is the androsterone compounds release by the male boar in the presence of a female, which attracts and induces her to stand while the male mounts (Dorries et al. 1997). Similarly, in sheep the ewes are attracted to the smell of ram wool, and detection of this odor via the MOS elicits a surge of luteinizing hormone. In mice, a highly volatile component of male urine, (methylthio)methanethiol (MTMT), is attractive to females and appears to be detected by the main olfactory system, in contrast to most of the other components of male urine (Lin et al. 2005). A final example is the suppression of estrous and reproduction in subordinate marmosets (Callithrix jacchus), which is clearly not mediated by the VNS. Olfactory, visual, and behavioral cues and signals from the dominant female appear to mediate this effect (Barrett et al. 1993).

Interaction between VNS and MOS 

Both the vomeronasal and main olfactory systems may operate together to elicit some behavioral responses and primer effects. Several examples have been described in the golden hamster (Mesocricetus auratus) (Johnston 1998). Hamsters live solitarily, probably with the usual small mammal pattern of exclusive female territories and larger overlapping male home ranges. Females mark their territories with fluid vaginal secretions that are detected from a distance by the male via the main olfactory system. The volatile pheromone appears to be methyl disulfide. Other secretions from the flank gland are critical for individual recognition in both sexes and are also detected via the main olfactory system. Once a male has approached a female, a large protein component of the vaginal mark appears to be critical for a surge of androgen that induces further male sexual behavior, including copulation. Removal of the vomeronasal organ eliminated the androgen surge in males, but sexually experienced males were usually still able to copulate. Dual lesions eliminated male mating behavior, while lesions of either system alone had little influence on mating. Similarly, both male and female will call to each other after an initial encounter. The calling is stimulated by odor input to both the VNO and MOB, and is reduced if either system is removed.

In addition to the primary role of VNO input for normal reproductive behavior in mice, blocking or knocking out the main olfactory system greatly reduces reproductive performance in both sexes. MOE-knockout males fail to investigate females’ anogenital region, mount, copulate, and behave aggressively towards other males (Mandiyan et al. 2005). Neurophysiological responses in the MOE to two known pheromones, 2-heptanone and farnesene, are absent in knockout mice but present in wild type mice (Wang et al. 2006). Similarly, MOE-knockout females fail to retrieve pups, construct normal nests, or show maternal aggression, and lack EOG responsiveness to known urinary pheromones (Wang et al. 2011). This evidence indicates that both MOE and VNO input are required for normal reproductive behavior.

The mouse olfactory recognition system is another example of integration of both volatile and non-volatile chemical components (Hurst et al. 2001; Hurst and Beynon 2004). Male mice place numerous urine marks around their territory. Mouse urine contains a significant amount of protein, most of it lipocalin-binding proteins from the multi-gene MUP family. As mentioned above, MUPs bind most, if not all, of the pheromones listed in Table 2, as well as a variety of other compounds in urine. The ligands are gradually released over many hours and become airborne volatiles, while the proteins remain stable for weeks. The proteins provide fixed genomic information about species, sex, and individual identity and can only be detected by contact through the vomeronasal system. The volatiles provide variable metabolic information concerning owner’s recent social, reproductive, and health status and his food resources and are detected by the main olfactory system. Figure 7 shows a model by which males might integrate information from the volatile and nonvolatile sources to learn about their male neighbors and competitors. The volatile components are more readily detected at a distance and would not require close investigation if the information in the volatile mix was familiar. If unfamiliar, the receiver could approach and contact the mark to obtain the associated stable information. In this way an owner could update his knowledge about the health and status of known or new intruders. Females also respond to these male marks, but the estrous-modulating effects (Table 2) only operate when they contact fresh urine marks in which the ligands are still bound to the MUPs. These ligand-MUP complexes are detected by the vomeronasal system: the protein component is apparently detected by the V2R/basal layer and anterior AOB part, and the ligand component by the V1R/apical and posterior AOB part. Variation in the suite of MUP proteins themselves provide information on individual identity and male heterozygosity. There are also MHC proteins in the urine that may add further useful information for both male and female receivers, especially with regard to genetic relatedness and strain differences. Either the proteins themselves, or their peptide products, or MHC-based metabolites bound to MUPs, may affect mating decisions in females. But the MHC associated odors are neither required nor sufficient for producing scent ownership recognition.

Figure 7: Associative matching between the volatile and involatile components of a urine mark. (A) When a mouse encounters the volatile signature of an unfamiliar mouse (orange), it moves up the concentration gradient to the source to make contact with the scent mark. An association is built between the volatile (more variable) and involatile (genomically hard-wired) information. (B) If the same pattern of volatiles is encountered, the receiver knows the identity of the scent owner and no further investigation is needed. (C) If the scent donor’s volatile signature includes new components (green) because of a change in diet, social status or infection status) then the receiver must update its association for this donor. The receiver can thus retain an image of other individuals in the population even in the context of a shifting metabolic profile. (After Hurst and Beynon 2004.)

There is ample evidence that neural output from the VNS and MOS combine in certain regions of the brain to integrate information from the two sources (Martinez-Marcos 2009; Tirindelli et al. 2009). In the hamster, some outputs of the MOS and VNS both converge onto single neurons in the amygdala in hamsters (Licht and Meredith 1987), in the hypothalamus in rats (Han and Swanson 2010), and in the telencephalon in salamanders (Roth and Laberge 2011). New techniques for mapping active circuits in the brain also demonstrate how sexually experienced animals can use inputs from the MOS in place of the VNO, and where these link to outputs from the vomeronasal system (Martinez-Marcos 2009). Inexperienced hamster males show low levels of activation in the medial preoptic area (MPOA) after they have investigated vaginal fluid, whereas sexually experienced males show a high level of activation after equivalent exposure to vaginal fluid. Removal of the VNO in experienced males does not reduce this activation. Experience seems to sensitize the MPOA to chemosensory input and to re-route input from the main olfactory system so that it can substitute for vomeronasal input driving the MPOA (Meredith 1998).

What about humans? Although a vomeronasal organ is present in human embryos, in the adult it becomes a blind-ended diverticulum in the septal mucosa, and contains no functional sensory nerve cells that extend to the brain. Furthermore, despite the existence of 5 intact V1R genes in humans, several of these may not be functional. One of them is expressed in the main olfactory epithelium, but actual cells expressing this gene that make connections between the epithelium and the brain have not yet been found (Rodriguez et al. 2000; Mundy and Cook 2003; Young et al. 2005; Witt and Hummel 2006). Thus it appears that humans do not possess any semblance of a vomeronasal sensory system. Nevertheless, humans do show sex-specific behavioral and physiological responses to various odors received through the MOE that likely qualify as pheromones (Stern and McClintock 1998; Jacob et al. 2001). Androstadienone is a male-produced product that activates preoptic and ventromedial nuclei of the hypothalamus in women and affects their endocrine levels, physiological arousal, mood, and sexual orientation (Savic et al. 2001; Saxton et al. 2008; Maraziti et al. 2011). The OR receptor for this chemical has recently been discovered (Keller et al. 2007). Estrogen-like substances activate the paraventricular and dorsomedial nuclei of the hypothalamus in men (Savic et al. 2001). Apocrine sweat glands of the human auxilla are the likely source of these steroid pheromones (Beier et al. 2005). Such odors affect mate choice and preferences, as described in main text Chapter 16. As in other mammals, the olfactory system of humans is well-designed to detect some volatile pheromones.

Literature cited

Adler, E., M.A. Hoon, K.L. Mueller, J. Chandrashekar, N.J.P. Ryba and C.S. Zuker. 2000. A novel family of mammalian taste receptors. Cell 100: 693–702.

Alioto, T.S. and J. Ngai. 2006. The repertoire of olfactory C family G protein-coupled receptors in zebrafish: candidate chemosensory receptors for amino acids. BMC Genomics 7: 309.

Barrett, J., D.H. Abbott and L.M. George. 1993. Sensory cues and the suppression of reproduction in subordinate female marmoset monkeys, Callithrix jacchus. Journal of Reproduction and Fertility 97: 301–310.

Beier, K., I. Ginez and H. Schaller. 2005. Localization of steroid hormone receptors in the apocrine sweat glands of the human axilla. Histochemistry and Cell Biology 123: 61–65.

Bjarndóttir, T.K., R. Fredriksson and H.B. Schiöth. 2005. The gene repertoire and the common evolutionary history of glutamate, pheromone (V2R), taste(1) and other related G protein-coupled receptors. Gene 362: 70–84.

Boschat, C., C. Pélofi, O. Randin, D. Roppolo, C. Lüscher, M.C. Broillet and I. Rodriguez. 2002. Pheromone detection mediated by a V1r vomeronasal receptor. Nature Neuroscience 5: 1261–1262.

Breer, H., J. Fleischer and J. Strotmann. 2006. The sense of smell: multiple olfactory subsystems. Cellular and Molecular Life Sciences 63: 1465–1475.

Brennan, P.A. and F. Zufall. 2006. Pheromonal communication in vertebrates. Nature 444: 308–315.

Buck, L. and R. Axel. 1991. A novel multigene family may encode odorant receptors — a molecular basis for odor recognition. Cell 65: 175–187.

Dorries, K.M., E. AdkinsRegan and B.P. Halpern. 1997. Sensitivity and behavioral responses to the pheromone androstenone are not mediated by the vomeronasal organ in domestic pins. Brain Behavior and Evolution 49: 53–62.

Døving, K.B. and D. Trotier. 1998. Structure and function of the vomeronasal organ. Journal of Experimental Biology 201: 2913–2925.

Døving, K.B., D. Trotier, J.F. Rosin and A. Holley. 1993. Functional architecture of the vomeronasal organ of the frog (genus Rana). Acta Zoologica 74: 173–180.

Dulac, C. and A.T. Torello. 2003. Molecular detection of pheromone signals in mammals: From genes to behaviour. Nature Reviews Neuroscience 4: 551–562.

Dulac, C. and S. Wagner. 2006. Genetic analysis of brain circuits underlying pheromone signaling. Annual Review of Genetics 40: 449–467.

Fleischer, J., H. Breer and J. Strotmann. 2009. Mammalian olfactory receptors. Frontiers in Cellular Neuroscience 3: 10.

Freitag, J., J. Krieger, J. Strotmann and H. Breer. 1995. Two classes of olfactory receptors in Xenopus laevis. Neuron 15: 1383–1392.

Grus, W.E. and J.Z. Zhang. 2006. Origin and evolution of the vertebrate vomeronasal system viewed through system-specific genes. Bioessays 28: 709–718.

Grus, W.E. and J.Z. Zhang. 2008. Distinct evolutionary patterns between chemoreceptors of 2 vertebrate olfactory systems and the differential tuning hypothesis. Molecular Biology and Evolution 25: 1593–1601.

Grus, W.E. and J.Z. Zhang. 2009. Origin of the genetic components of the vomeronasal system in the common ancestor of all extant vertebrates. Molecular Biology and Evolution 26: 407–419.

Halpern, M. and J.L. Kubie. 1984. The role of the ophidian vomeronasal system in species typical behavior. Trends in Neurosciences 7: 472–477.

Halpern, M. and A. Martinez-Marcos. 2003. Structure and function of the vomeronasal system: an update. Progress in Neurobiology 70: 245–318.

Hamdani, E.H. and K.B. Døving. 2006. Functional organization of the olfactory system in fish. In Communication in Fishes (Ladich, F., S.P. Collin, P. Moller and B.G. Kapoor, eds). Enfield (NH): Science Publishers. pp. 223–257.

Hahn, J.D. and L.W. Swanson. 2010. Distinct patterns of neuronal inputs and outputs of the juxtaparaventricular and suprafornical regions of the lateral hypothalamic area in the male rat. Brain Research Reviews 64: 14–103.

Hansen, A., K.T. Anderson and T.E. Finger. 2004. Differential distribution of olfactory receptor neurons in goldfish: Structural and molecular correlates. Journal of Comparative Neurology 477: 347–359.

Hashiguchi, Y. and M. Nishida. 2006. Evolution and origin of vomeronasal-type odorant receptor gene repertoire in fishes. BMC Evolutionary Biology 6: 13.

Hashiguchi, Y. and M. Nishida. 2007. Evolution of trace amine-associated receptor (TAAR) gene family in vertebrates: Lineage-specific expansions and degradations of a second class of vertebrate chemosensory receptors expressed in the olfactory epithelium. Molecular Biology and Evolution 24: 2099–2107.

Hino, H., N.G. Miles, H. Bandoh and H. Ueda. 2009. Molecular biological research on olfactory chemoreception in fishes. Journal of Fish Biology 75: 945–959.

Hurst, J.L. and R.J. Beynon. 2004. Scent wars: the chemobiology of competitive signalling in mice. Bioessays 26: 1288–1298.

Hurst, J.L., C.E. Payne, C.M. Nevison, A.D. Marie, R.E. Humphries, D.H.L. Robertson, A. Cavaggioni and R.J. Beynon. 2001. Individual recognition in mice mediated by major urinary proteins. Nature 414: 631–634.

Ishii, T. and P. Mombaerts. 2008. Expression of nonclassical class I major histocompatibility genes defines a tripartite organization of the mouse vomeronasal system. Journal of Neuroscience 28: 2332–2341.

Jacob, S., L.H. Kinnunen, J. Metz, M. Cooper and M.K. McClintock. 2001. Sustained human chemosignal unconsciously alters brain function. Neuroreport 12: 2391–2394.

Johnston, R.E. 1998. Pheromones, the vomeronasal system, and communication — From hormonal responses to individual recognition. In Olfaction and Taste XII (Murphy, C. A., ed). New York: New York Acad Sciences. pp. 333–348.

Johnston, R.E. 2000. Chemical communication and pheromones: the types of signals and the role of the vomeronasal system. In The Neurobiology of Taste and Smell, 2nd edition (Finger, T.E., W.L. Silver and D. Restrepo, eds). New York: Wiley. pp. 99–123.

Keller, A., H.Y. Zhuang, Q.Y. Chi, L.B. Vosshall and H. Matsunami. 2007. Genetic variation in a human odorant receptor alters odour perception. Nature 449: 468–U6.

Keverne, E.B. 1999. The vomeronasal organ. Science 286: 716–720.

Kimoto, H., S. Haga, K. Sato and K. Touhara. 2005. Sex-specific peptides from exocrine glands stimulate mouse vomeronasal sensory neurons. Nature 437: 898–901.

Leinders-Zufall, T., P. Brennan, P. Widmayer, P. Chandramani, A. Maul-Pavicic, M. Jager, X.H. Li, H. Breer, F. Zufall and T. Boehm. 2004. MHC class I peptides as chemosensory signals in the vomeronasal organ. Science 306: 1033–1037.

Leinders-Zufall, T., T. Ishii, P. Mombaerts, F. Zufall and T. Boehm. 2009. Structural requirements for the activation of vomeronasal sensory neurons by MHC peptides. Nature Neuroscience 12: 1551–1598.

Liberles, S.D. and L.B. Buck. 2006. A second class of chemosensory receptors in the olfactory epithelium. Nature 442: 645–650.

Liberles, S.D., L.F. Horowitz, D.H. Kuang, J.J. Contos, K.L. Wilson, J. Siltberg-Liberles, D.A. Liberles and L.B. Buck. 2009. Formyl peptide receptors are candidate chemosensory receptors in the vomeronasal organ. Proceedings of the National Academy of Sciences of the United States of America 106: 9842–9847.

Licht, G. and M. Meredith. 1987. Convergence of main and accessory olfactory pathways onto single neurons in the hamster amygdala. Experimental Brain Research 69: 7–18.

Liman, E.R. 2006. Use it or lose it: molecular evolution of sensory signaling in primates. Pflugers Archiv-European Journal of Physiology 453: 125–131.

Lin, D.Y., S.Z. Zhang, E. Block and L.C. Katz. 2005. Encoding social signals in the mouse main olfactory bulb. Nature 434: 470–477.

Luo, M.M. 2004. Got milk? A pheromonal message for newborn rabbits. Bioessays 26: 6–9.

Ma, W.D., Z.S. Miao and M.V. Novotny. 1999. Induction of estrus in grouped female mice (Mus domesticus) by synthetic analogues of preputial gland constituents. Chemical Senses 24: 289–293.

Ma, M.H. 2007. Encoding olfactory signals via multiple chemosensory systems. Critical Reviews in Biochemistry and Molecular Biology 42: 463–480.

Mandiyan, V.S., J.K. Coats and N.M. Shah. 2005. Deficits in sexual and aggressive behaviors in Cnga2 mutant mice. Nature Neuroscience 8: 1660–1662.

Martinez-Marcos, A. 2009. On the organization of olfactory and vomeronasal cortices. Progress in Neurobiology 87: 21–30.

Marazziti, D., P. Torri, S. Baroni, M.C. Dell'Osso, G. Consoli and V. Boncinelli. 2011. Is androstadienone a putative human pheromone? Current Medicinal Chemistry 18: 1213–1219.

Meredith, M. 1998. Vomeronasal, olfactory, hormonal convergence in the brain — Cooperation or coincidence? Annals of the New York Academy of Sciences 855: 349–361.

Mombaerts, P. 2004. Genes and ligands for odorant, vomeronasal and taste receptors. Nature Reviews Neuroscience 5: 263–278.

Moncomble, A.S., G. Coureaud, B. Quennedey, D. Langlois, G. Perrier and B. Schaal. 2005. The mammary pheromone of the rabbit: from where does it come? Animal Behaviour 69: 29–38.

Müller-Schwarze, D. 2006. Chemical Ecology of Vertebrates. Cambridge: Cambridge University Press.

Mundy, N.I. and S. Cook. 2003. Positive selection during the diversification of class I vomeronasal receptor-like (V1RL) genes, putative pheromone receptor genes, in human and primate evolution. Molecular Biology and Evolution 20: 1805–1810.

Munger, S.D., T. Leinders-Zufall and F. Zufall. 2009. Subsystem organization of the mammalian sense of smell. Annual Review of Physiology 71: 115–140.

Mucignat-Caretta, C. 2010. The rodent accessory olfactory system. Journal of Comparative Physiology A 196: 767–777.

Niimura, Y. and M. Nei. 2006. Evolutionary dynamics of olfactory and other chemosensory receptor genes in vertebrates. Journal of Human Genetics 51: 505–517.

Novotny, M.V., W.D. Ma, D. Wiesler and L. Zidek. 1999. Positive identification of the puberty-accelerating pheromone of the house mouse: the volatile ligands associating with the major urinary protein. Proceedings of the Royal Society of London Series B-Biological Sciences 266: 2017–2022.

Petti, M.A., S.F. Matheson and G.D. Burd. 1999. Differential antigen expression during metamorphosis in the tripartite olfactory system of the African clawed frog, Xenopus laevis. Cell and Tissue Research 297: 383–396.

Pfister, P. and I. Rodriguez. 2005. Olfactory expression of a single and highly variable V1r pheromone receptor-like gene in fish species. Proceedings of the National Academy of Sciences of the United States of America 102: 5489–5494.

Riviere, S., L. Challet, D. Fluegge, M. Spehr and I. Rodriguez. 2009. Formyl peptide receptor-like proteins are a novel family of vomeronasal chemosensors. Nature 459: 574–577.

Rodriguez, I., C.A. Greer, M.Y. Mok and P. Mombaerts. 2000. A putative pheromone receptor gene expressed in human olfactory mucosa. Nature Genetics 26: 18–19.

Roth, F.C. and F. Laberge. 2011. High convergence of olfactory and vomeronasal influence in the telencephalon of the terrestrial salamander Plethodon shermani. Neuroscience 177: 148–158.

Ryba, N.J.P. and R. Tirindelli. 1997. A new multigene family of putative pheromone receptors. Neuron 19: 371–379.

Saraiva, L.R. and S.I. Korsching. 2007. A novel olfactory receptor gene family in teleost fish. Genome Research 17: 1448–1457.

Savic, I., H. Berglund, B. Gulyas and P. Roland. 2001. Smelling of odorous sex hormone-like compounds causes sex-differentiated hypothalamic activations in humans. Neuron 31: 661–668.

Saxton, T.K., A. Lyndon, A.C. Little and S.C. Roberts. 2008. Evidence that androstadienone, a putative human chemosignal, modulates women's attributions of men's attractiveness. Hormones and Behavior 54: 597–601.

Schaal, B., G. Coureaud, D. Langlois, C. Ginies, E. Semon and G. Perrier. 2003. Chemical and behavioural characterization of the rabbit mammary pheromone. Nature 424: 68–72.

Shi, P. and J.Z. Zhang. 2007. Comparative genomic analysis identifies an evoutionary shift of vomeronasal receptor gene repertoire in the vertebrate transition from water to land. Genome Research 17: 166–174.

Spehr, M. and S.D. Munger. 2009. Olfactory receptors: G protein-coupled receptors and beyond. Journal of Neurochemistry 109: 1570–1583.

Stern, K. and M.K. McClintock. 1998. Regulation of ovulation by human pheromones. Nature 392: 177–179.

Stowers, L. and D.W. Logan. 2010. Olfactory mechanisms of stereotyped behavior: on the scent of specialized circuits. Current Opinion in Neurobiology 20: 274–280.

Su, C.Y., K. Menuz and J.R. Carlson. 2009. Olfactory perception: Receptors, cells, and circuits. Cell 139: 45–59.

Thom, M.D., P. Stockley, F. Jury, W.E.R. Ollier, R.J. Beynon and J.L. Hurst. 2008. The direct assessment of genetic heterozygosity through scent in the mouse. Current Biology 18: 619–623.

Tirindelli, R., M. Dibattista, S. Pifferi, and A. Menini. 2009. From Pheromones to Behavior. Physiological Reviews 89: 921–956.

Touhara, K. and L.B. Vosshall. 2009. Sensing odorants and pheromones with chemosensory receptors. Annual Review of Physiology 71: 307–332.

Wang, Z.S., C.B. Sindreu, V. Li, A. Nudelman, G.C.K. Chan and D.R. Storm. 2006. Pheromone detection in male mice depends on signaling through the type 3 adenylyl cyclase in the main olfactory epithelium. Journal of Neuroscience 26: 7375–7379.

Wang, Z.S. and D.R. Storm. 2011. Maternal behavior is impaired in female mice lacking type 3 adenylyl cyclase. Neuropsychopharmacology 36: 772–781.

Witt, M. and T. Hummel. 2006. Vomeronasal versus olfactory epithelium: Is there a cellular basis for human vomeronasal perception? International Review of Cytology — a Survey of Cell Biology 248: 209–259.

Wyatt, T.D. 2014. Pheromones and Animal Behaviour: Chemical Signals and Signatures. Cambridge: Cambridge University Press.

Yang, H. and P. Shi. 2010. Molecular and evolutionary analyses of formyl peptide receptors suggest the absence of VNO-specific FPRs in primates. Journal of Genetics and Genomics 37: 771–778.

Yoon, H.Y., L.W. Enquist and C. Dulac. 2005. Olfactory inputs to hypothalamic neurons controlling reproduction and fertility. Cell 123: 669–682.

Young, J.M., M. Kambere, B.J. Trask and R.P. Lane. 2005. Divergent V1R repertoires in five species: Amplification in rodents, decimation in primates, and a surprisingly small repertoire in dogs. Genome Research 15: 231–240.

Zhao, H.Q., L. Ivic, J.M. Otaki, M. Hashimoto, K. Mikoshiba and S. Firestein. 1998. Functional expression of a mammalian odorant receptor. Science 279: 237–242.