Behavioral Genetics Research Paper

The writings of two Victorians, Charles Darwin and his cousin Francis Galton, influenced different paths in the field of behavior genetics. One is concerned with the causes of variation in human behaviors, especially cognition, personality, psychopathology, and addictions; this path is derived largely from the works of Galton. The other is concerned with the genetics of adaptive behaviors in animals and humans;this path is derived largely from the writings of Darwin; it is the main focus of this review. In each section of this research paper, the relevant genetics are considered first and then their application to some exemplar behaviors is described.

The structure of the genetic material called DNA(deoxyribonucleic acid) was first proposed by Watson and Crick (1953a). Implicit in the structure of DNAare the mechanisms for how genes are replicated, how they mutate, and what they do (Watson & Crick, 1953b). This discovery has had a profound effect on the biological sciences and is beginning to have one on psychology and the other behavioral sciences. The potential impact on the behavioral sciences was described in a seminal paper by Ginsburg (1958) that was published just 5 years after those by Watson and Crick on the structure and function of DNA. Ginsburg proposed that genetics was a tool for the study of behavior in four ways. First, it is a tool to dissect behavior into its natural units. Second, it is a tool to study the neural and other mechanisms of behavior. Third, it is a tool to study the effects of the environment on behavior. Fourth, it is a tool to study the evolution of behavior.

The Heritability of Behavior

Background

The methods of quantitative genetics (see Falconer & McKay, 1997) are used to assess the relative roles of genes and environment in individual differences in behavior. In animals, inbreeding, selective breeding, and crossbreeding are used, and in animals and humans, the resemblance among relatives and nonrelatives is assessed. Basically, these methods allow the phenotypic variance to be partitioned into genetic and environmental components. This can be expressed as the ratio of genotypic to phenotypic variance; this ratio is known as the heritability of the trait. This ratio can vary from zero to one. The value is specific to the population in which it is measured, and its estimate always has an error term dependent on sample size. The following are considered in this section as examples: mating speed in fruit flies, aggression in stickleback fish, nest building in mice, and body weight in humans.

Mating Speed in Fruit Flies

Manning (1961) selectively bred for fast and slow mating speeds based on the time from being introduced to the mating chamber to the start of copulation. From 50 pairs of flies, the 10 fastest and10 slowest pairs were selected. From these pairs, two fast-mating and two slow-mating lines were established, and a randomly bred control line was also maintained. After 25 generations of selective breeding, the mean mating speed was about 3 min in the fast line and 80 min in the slow line. The heritability computed from the response to selective breeding was .30, which demonstrates low to moderate genetic variability for this behavior. This study also demonstrated that the genes for variability in mating speed and general activity are not the same.

Aggression in Three-Spine Stickleback Fish

In stickleback fish, juveniles and adults of both sexes attack other species members, and attacks by adults of both sexes are usually territorial. Bakker (1994) selected for territorial attacks in males and females from a population of stickleback fish living in a fresh water stream in the Netherlands. High, low, and control lines were developed. For adult male attacks, there was no change in the high line across the three generations of selective breeding, but there was a decrease in the low line over the three generations. For adult female attacks, there was an increase in the high and a decrease in the low lines over the three generations. Where heritability could be calculated, it ranged from .29 to .64 (moderate to high genetic contribution to variation). It may be concluded from these results that (a) there is genetic variability for these behaviors in the natural population, (b) there may have been more intense selection on male than female attacks, and (c) some of the genes for variability in male and female attacks may not be the same.

Nest Building in Mice

Mice build nests as adaptations to the cold. In one study, mice were selectively bred for the size of their nests as measured by grams of cotton used to build a nest (Lynch, 1994). The foundation population was a randomly bred heterogeneous stock, and there were two high, two low, and two control lines. After 15 generations of selection, the heritability computed from the response to selection was .28, which is moderate. There have been more than 48 generations of selection. Thus, there is genetic variability for this trait in the foundation population. Studies with natural populations of mice show that there is less genetic variability in mice from Maine than from Florida; this finding is consistent with there being more selection in Maine for thermoregulatory nest size than in Florida.

Body Weight in Humans

The regulation of body weight is an adaptive trait; this regulation has behavioral components, such as eating and exercise. The correlation for body weight of identical twins raised together is .80 for identical twins reared together and .72 for those raised apart, whereas it is .43 for fraternal twins (Grilo & Pogue-Geille 1981). The correlation for biological parents and adopted siblings is .23, similar to that for nonadoptive parents and offspring (.26). Also, the correlations for adoptive parent and child or adoptive siblings are essentially zero. These findings are consistent with a heritability of .70 for body weight and with nonshared rather than shared environmental effects on individual differences in body weight. The shared environment is that which differs from one family to another; the nonshared environment is conceived of as the portion of trait variance not explained by genetics or shared environment. For many individual differences in behavioral or mental traits of humans, most of the variance is due to genetic and nonshared environmental effects (Plomin, DeFries, McClearn, & McGuffin, 2000).

Summary

It is now firmly established that individual differences in every studied behavioral trait in any animal species, including humans, is a function of both genetic and environmental variability. That is to say, for no trait studied is the heritability either zero or one in outbred populations. The interesting issues now are (a) what are the genes that do or can affect a trait’s variability, (b) how they interact with each other and the environment in the trait’s development, and (c) what the genetic mechanisms are of species diversity in and evolution of behavioral adaptations. These issues are beginning to be understood with the aid of molecular genetics.

The Genome Projects

An individual’s nuclear genome consists of the DNAfound on all the chromosomes in the nucleus of its cells; there is one molecule of DNA for each chromosome. The goal of the genome projects is to determine the sequence of the nucleotide bases—adenine, cytosine, guanine, or thymine (A, C, G, orT)—of the nuclear genome of one or more individuals in a species. This has been done for C. elegans with a genome size of 100 Mb (megabases), D. melanogaster with a genome size of 165 Mb, and Homo sapiens with a genome size of 3,300 Mb; it is nearing completion for M. musculus with a genome size 3,300 Mb.After the entire sequence is known for a species, it is possible to estimate the number of proteincoding genes in the genome; this now appears to be about 35,000 for humans. Also, the amino acid sequence in each protein can be deduced from the coding nucleotide triplets in the gene’s structural region. Other sequences of a proteincoding gene bind proteins known as transcription factors. These factors and sequences together are involved in controlling when and where a gene is transcribed as RNA (ribonucleic acid). The transcribed RNA is processed into a messenger RNA (mRNA), and the mRNA is then translated into the sequences of amino acids in its protein. Many DNA sequences, however, appear not to be transcribed nor to regulate transcription; in mice and humans, these make up about 98% of the nuclear genome.

There is also DNA in the mitochondria; this DNA codes for the amino acid sequence of some of the proteins involved in energy metabolism. This DNA has been sequenced for many organisms. There are neurological effects of variants of these mitochondrial genes (Wallace, 1999), and these genes may also have behavioral effects.

Identifying The Genes

Background

Geneswitheffectsonbehaviorinmicecanbedetectedbymutagenizing one allele of a gene (Takahashi, Pinto, &Vitaterna, 1994). There are then two copies (variants or alleles) of the gene, and the homozygotes for the two variants can be compared for differences in one or more behaviors. One mutagenesis approach targets a specific gene creating an allele that does not produce a functional protein. These are the so-called knockout mutants. To delete a specific gene, the DNA sequence of the gene must be already known, and the chromosomallocationofthegenemayormaynotbeknown.Knockouts with effects on mouse behavior are reviewed in Nelson and Young (1998) and Crawley (2000). This approach is illustrated in the following section for genes with effects on mouse aggression.Another approach exposes male mice of an inbred strain to a chemical mutagen, such as N-ethyl-N-nitrosourea (ENU), with the goal of finding most—if not all—the genes that can cause variation in a trait such as a behavior. Dominant mutants would be detected in the first generation of progeny, and recessive mutants would be detected in subsequent generations of progeny. The chromosomal location and the DNA sequenceofeachmutantgenewithabehavioraleffectarethen determined. There are large-scale behavioral mutagenesis projects at the Jackson Laboratory, Northwestern University, andThe University ofTennessee Health Sciences Center.This approach is illustrated later in this research paper with the circadian rhythm gene known as Clock. Also, both mutagenesis approaches have also been used with nematodes (Segalat, 1999) and fruit flies (Jallon, 1999).

Genes with effects on behavior can also be detected by chromosomally mapping genes with existing allelic variants (see Segalat, 1999, for nematodes; see Sokolowski, 1999, for fruit flies; see Belknap, Dubay, Crabbe, & Buck, 1997, for mice; and see Plomin & Crabbe, 2000, for humans). Single genevariantswithlargeeffectcanbemappedasillustratedfor the foraging gene in Drosophila. Polygenic variants can also be mapped; later in the paper, this is illustrated in mice for quantitative trait loci (QTLs) with effects on emotionality. A goal in mapping these is to eventually identify the actual protein coding genes with effects on a trait’s variation. The markersthatareusedforgenemappingtoaregionofachromosome are usually DNA variants (Plomin et al., 2000; Plomin & Crabbe, 2000). These include restriction fragment length polymorphisms (RFLPs), simple sequence length polymorphisms (SSLPs), and single nucleotide polymorphisms (SNPs). These markers have two advantages in mapping: Many of them are closely spaced on the linkage map of each chromosome, and they have no effect on the measured traits. Another approach is to associate DNA variants of candidate genes with behavior, as is illustrated for a human personality trait later in this research paper. The selection of candidate genes often depends on hypotheses about the neural and other mechanisms involved in a behavior.

Another approach to identifying genes with potential effects on behavior is to look for quantitative differences in gene expression in brains of genotypically or phenotypically different individuals. Genes that differ across these individuals in level of brain mRNA are candidates for ones with behavioral effects. DNAmicrochips can be used to determine what genes are being expressed in a tissue (e.g., as a region of brain) and how the level of expression differs between individuals of different genotype or phenotype (Nisenbaum, 2002). A single microchip can be used to assay for thousands of genes simultaneously. The assay is based on specific hybridization between some of the DNAof a gene and of that gene’s mirror image or complimentary DNA (cDNA), which is synthesized from the gene’s messenger RNA.

Knockout Mutants and Aggression in Mice

There is more than one type of aggression in mice and other animals (Maxson, 1992, 1999). The research with knockout mutantshasfocusedonaggressivebehaviorknownasoffense, which has the adaptive function of obtaining and retaining resources such as space, food, and mates. It is also characterized by specific motor patterns and attack targets. About 25 genes witheffectsonoffenseinmaleshavebeenidentified(Maxson, Roubetoux, Guillot, & Goldman, 2001; Miczek, Maxson, Fish, & Faccidomo, 2001); most of these have been identified with knockouts. Many of the genes act on either hormone or neurotransmitter systems, and it has recently been suggested that many such genes ultimately act on offense by affecting one or more of the serotonergic neurotransmitter systems (Nelson & Chiavegatto, 2001).

The effect of a knockout on a trait is determined by comparing mice homozygous for the knockout allele with those homozygous for the normal or functional allele. In evaluating results with knockouts, there are several methodological concerns (Nelson, 1997):

• To avoid maternal effects, the mother of the two genotypes must be the same, and the offspring should be the result of the mating of a heterozygous female to a heterozygous male.
• Some knockout strain pairs are coisogenic, differing only in the normal and mutant alleles of a single gene, but others are only congenic—they differ not only in the mutant and normal alleles of gene of interest, but also in alleles of genes linked to it. For congenic strains, any differences may be due to the genes linked to the knockout rather than due to the knockout itself, as discussed by Gerlai (1996).
• Often the knockout is made in one inbred strain, such as one of the 129 inbred strains and then transferred to another strain. Sometimes the effect of a knockout seen in one strain background is not detected in another. For example, the knockout for the NOS-1 (nitric oxide synthase-1) gene increases attacks is lost after many generations of backcrossing to C57BL6 inbred strain mice (LeRoy et al., 2000).
• For many knockouts, the mutant gene is present from the time of conception. Thus, it was not possible to tell when or where in the mouse the gene was expressed with consequent behavioral effect. Recently, techniques have been developed that permit tissue- and temporal-specific knockouts in mice (Tsien, 1999). Similar techniques are available for elegans (Seglant, 1999) and Drosophila (Jallon, 1999).

Saturation Mutagenesis and Circadian Rhythms in Mice

Knockouts mutagenize a specific gene. Exposure to a chemical mutagen, such as ENU, in theory can mutagenize all the genes that can cause a trait to vary (Takahashi et al., 1994). This approach was first used in Drosophila (see review by Benzer, 1971); mutants on the X chromosome were detected for circadian rhythms, courtship, and learning-memory. Saturation mutagenesis has also been used with nematodes (Seglant, 1999). This approach was first used in mice to screen for mutants with effects on circadian rhythms.

Circadian rhythms in mice can be measured by observing their running in a wheel. Mice normally run at night in a 12 light, 12 dark schedule, with precise onset of locomotor activity. Even when maintained in constant darkness, mice (and all other animals examined) display this cycle of activity that deviates only slightly from the 24-hour pattern observed during a light-dark cycle. Because these rhythms persist in constant conditions as well as other evidence, it is well accepted that these rhythms are generated from within the organism. Male mice were treated with ENU and their progeny were screened for dominant mutations affecting circadian rhythms (Vitaterna et al., 1994). One mouse out of 300 had a circadian period that was 1 hour longer than normal. This mouse had a semidominant mutation that was named Clock. When initially placed in constant darkness, homozygous Clock mice have long circadian periods of 27 to 28 hours. After about 2 weeks in constant darkness, the homozygous Clock mice have a complete loss of circadian rhythms. Intraspecic mapping crosses were used to show that the Clock gene is located on Chromosome 5 of the mouse. It was then possible to positionally clone the DNAand thereby to identify the protein of this gene (King et al., 1997). Clock encodes a member of the basic-helix-loop-helix (bHLH) PAS family of transcription factors, and it has a key role in the regulating the genetics mechanism of the biological clock (Allada, White, So, Hall, & Rosbash, 2001; King & Takahashi, 2000). The role of Clock in circadian rhythms was confirmed with transgenic rescue (Antoch et al., 1997); for this, the normal allele of Clock was inserted into the Clock mutants. These mice have normal circadian rhythms.

Natural Variants and Foraging in Fruit Flies

There is a polymorphism in the foraging behavior of larval Drosophila (Sokolowski & Riedl, 1999). Rovers have longer foraging trails on food than do sitters. This difference is not seen in the absence of food. These variants occur in natural populations of flies; in these, there are about 70% rovers and 30% sitters. Breeding experiments showed that this polymorphism is due primarily to allelic differences in a single gene with a large effect, but that this trait is also influenced by other genes that each contribute small effects. Chromosomal analyseslocalizedthegene,whichwascalledfor,tothesecondchromosome. Because minor genes and the environment affect the distribution of rover and sitter phenotypes, a lethal tagging technique was used to map for to Region 24 of the polytene chromosome map. Chromosome rearrangements were used to further map for to region 24A3-5 of the polytene chromosome map containing about 150 to 125 kb (kilobases) of DNA. It is now known that the for gene is the same as the gene dg2, which encodes a Drosophila cGMP/cGMP-dependent protein kinase (PKG). Insertion of transposon elements into dg2 caused a change from rover to sitter phenotypes; removal of the transposon elements caused the behavior to revert back to rover.

Genome-Wide Scan to Map QTLs for Emotionality in Mice

When confronted with a novel and unexpected situation, such as an open field, mice may freeze, defecate and urinate, or simply explore the new environment (Broadhurst, 1960). These behaviors, singly or in combination, are often used as measures of emotionality. In mice, negative correlations between defecation and ambulation are fairly general, although the association is affected somewhat by environmental variables such as light or noise (Archer, 1973). To some extent, the relation also depends on the strain, sex, and early experience of the subjects.

In open-field tests, C57BL6 inbred strain of mice are much more active than BALB/c inbred strain of mice. These strains were crossed to obtain an F3 generation, which was the base population for selective breeding for open-field activity. Replicate high and low lines were selectively bred over 30 generations for open-field ambulation (DeFries, Gervais, & Thomas, 1978). Two unselected control lines were also bred for the 30 generations. After 30 generations of selection, there was a threefold difference between the high and low lines in ambulation, and there was no overlap in distribution of ambulation scores between the high and low lines. The defecation scores of the low lines were seven times higher than those of the high lines. Based on the response to selection, the heritability was .26 for ambulation and .11 for defecation, and the genetic correlation for ambulation and defecation scores was .86.

An F3 of one of these high and low lines was bred to map QTLs for the strain differences (Flint et al., 1995). The most active and least active mice were screened for 84 DNAmarkers for which there were two alleles, and these 84 markers were spread across the 20 chromosomes of mice; this is known as a genome-wide scan. These were used to determine the chromosomal region (QTLs) associated with the openfield activity. Significant QTLs were found on Chromosomes 1, 4, 12, 15, 17, and 18. These six QTLs accounted for 26% of the phenotypic variance. There are several issues in evaluating such genome-wide scans to localize QTLs associated with a behavior. These are

• Because the association of the behavior with many DNA markers is tested, there is a risk of false positives (for mice, see Belknap et al., 1997; for humans, see Plomin & Crabbe, 2000). For this reason, QTLs should be confirmed by additional studies.
• The QTL is often a large region on a chromosome consisting of millions of base pairs of DNA and hundreds of genes. In order to positionally clone or identify positional candidates for the gene or genes underlying the QTL, the map distance between the QTL to the markers needs to be greatly reduced. The following illustrates one approach to resolving these issues (Talbot et al., 1999). For open-field behavior, the most active 20% and least active 20% of 751 mice of an heterogenous stock derived from an eight-way cross were screened for SSLPs closely spaced together in a 20 cM (centiMorgan) region of Chromosome 1. This not only confirmed that there is a QTLon Chromosome 1 with an effect on ambulation, but it also mapped the QTL more precisely.

Candidate Genes: Association of D4 Receptor Gene and Novelty Seeking in Humans

Novelty seeking is one of four personality traits in Cloninger’s theory of temperament development (Cloninger, Svrakic, & Przybeck, 1993). His theory predicts that genes acting on dopaminergic transmission would affect individual differences in novelty seeking. Associations between allelic variants of the dopamine D4 receptor (DRD4) were assessed in unselected samples (Benjamin et al., 1996; Ebstein et al., 1995). The seven alleles for the DRD4 receptor gene vary in the number of a 48-bp (base pair) repeat. In both studies, individuals with the longer alleles (6–8 repeats) had higher novelty seeking scores than did those with the shorter alleles (2–5 repeats). It may be that those with the longer alleles are dopamine deficient and seek novelty in order to increase dopamine release. Regardless, these variations in the D4DR gene account for about 4% of the phenotypic variation in novelty seeking.

The aforementioned study is a candidate gene rather than whole genome scan approach for finding genes with effects on individual differences in behavior; this type of approach presents the following methodological concerns:

• For association studies using population samples of humans, there may be artificial correlations between a candidate gene and a trait that are due to ethnic differences in frequency of the alleles of a candidate gene. This does not occur with within-family designs for association between a candidate gene and a trait. The association between length of repeat of the D4DR receptor and novelty seeking was also obtained in a within-family design.
• It is possible that an association may be a false positive. Hence, there should be independent confirmations of each reported association between a genetic variant and individual differences in behavior. There have been many replications of the association between length of repeat of the D4DR receptor and novelty (Plomin & Caspi, 1998). But there also have been a few failures to replicate this finding. This is likely to occur when the genetic variant has small effects on individual differences in the trait, as is the case for the effect of D4DR variants on novelty seeking.

Differences in Expression of Genes in the Brain

The technique of using DNAmicrochips is illustrated here for genes expressed in brain areas of two strains of mice. The inbred strains of mice are C57BL6 and 129SVEv, and the brain regions are midbrain, cerebellum, hippocampus, amygdala, entorhinal cortex, and cerebral cortex (Sandberg et al., 2000). Expression of 7,000 mouse genes were detected with the DNA microchip. Twenty-four of these were differentially expressed in all brain regions of the two strains, and 73 were differentially expressed in at least one of the brain regions of the two strains. These genes may be candidates for known behavioral differences between these strains. There are two methodological issues with regard to this approach.

• The microchip DNA arrays will not detect genes with low levels of mRNA. It is currently limited to detecting genes expressed at a relative abundance of 1/100000 mRNAs.
• There may be false positives with this technique. For this reason, findings on gene expression should be conformed with other techniques for detecting mRNAs such as Northern blots, RT-PCR (reverse transcriptase polymerase chain reaction), or instu

This technique can also be used to study gene expression from human postmortem tissue. For example, it has been used to suggest that there may be differences between individuals with schizophrenia and normally functioning individuals in brain expression of two genes involved in synaptic function (Mirnics, Middleton, Lewis, & Levitt, 2001). But because premortem and postmortem factors can influence the findings with these tissues, these and similar results should be interpreted with caution, and genes implicated with this approach in phenotypic differences should be confirmed with other approaches. For example, one of the genes detected by microchip hybridization to be differentially expressed in the brains of normal and schizophrenic individuals are in a region on Chromosome 1 associated with the risk for schizophrenia.

Summary

Eventually, most (if not all) the genes that do or can cause variation in behaviors of nematodes, flies, and mice will be known, as will those that cause variation in human behavior. Because every gene does not affect every behavior in a species, behaviors can be grouped by the genes that cause them to vary. This would be the basis for a behavioral taxonomy based on gene effects as proposed by Ginsburg in 1958. Also, as genes with behavioral effects are identified in one species, they may be useful in two ways in the search for genes with behavioral effects in other species. First, DNA hybridization techniques can be used to search for homologous genes—ones similar in base pair sequence—in two species. For example, the period gene, which affects circadian rhythms, was first identified and sequenced in flies. The DNA sequence of the period gene of flies was then used to see whether there were homologous genes in mice. This approach identified three period genes in mice. Second, the sequences of genes on chromosome segments are conserved in mammals. These homology maps can be used to suggest the chromosomal location in humans of a gene or QTL mapped in mice. For example, the QTL on Chromosome 1 with effects on emotionality of mice would be located in a specific region of human Chromosome 1.

Developmental Genetics of Behavior

Background

Protein coding genes of eukaryotes essentially have two parts, the structural regions and the regulatory regions. The sequence of base pairs in the structural region codes for the sequence of amino acids in its protein. It serves as a template for synthesis of RNA; this is transcription. This RNA is first processed and the resulting mRNA is translated into the sequences of amino acids in the protein. The amino acid sequence of the protein is a determinant of its function. The base pair sequences in the regulatory region of a gene bind proteins known as transcription factors. Together, these determine when and where in the individual a gene is transcribed or expressed. In mice and humans, about 2% of the nuclear genome codes for proteins, and in mice about half of these genes are transcribed in adult brain. Further information on molecular genetics can be found in Lewin (1997) and on developmental genetics in Gilbert (2000).

In this context, three issues in developmental genetics of behavior are considered. These issues concern identifying (a) the pathway from DNA to a behavior, (b) the interaction of genes or epistasis and behavior, and (c) the effects of the interaction between genes and environments on behavior.

Pathways From the Gene to the Behavior

The initial step is to know the DNA sequences of a gene, thereby identifying the amino acid sequence of its protein. These sequences aid in identifying the cellular function of the gene. After this is known, the question becomes how varying that protein has behavioral effects.

For example, the gene for the enzyme nitric oxide synthase-1 (NOS-1) was knocked out. The homozygotes missing the enzyme in neurons are more aggressive than are those having the enzyme; these mice lack the gaseous neurotransmitter nitric oxide (NO; Nelson et al., 1995). The next steps would be to determine how the lack of NO increases aggression and how the presence of NO decreases aggression. It has recently been shown that the knockout mice have reduced serotonin (5-HT) turnover and are deficient in 5HT_1A and 5HT_1B receptor function (Chiavegatto et al., 2001). It remains to be determined just how the presence and absence of NO affects the serotonin system and just how it in turn effects aggression. However, there is much evidence that the serotonin system is a major player in regulating aggressive behavior.

Gene Interactions and Behavior

Although it is possible to trace the effect of some individual genes from DNA to behavior, it is becoming increasingly clear that the effects of alleles of one gene on behavior depend on alleles of other genes. The interaction of the alleles of different genes is known as epistasis. The following is an example: The Y chromosomes of DBA1 and C57BL10 inbred mouse strains mice can differ in their effects on mouse aggression (Maxson, Trattner, & Ginsburg, 1979). This difference only occurs, however, if all or half the autosomal genes are from DBA1; it does not occur if all the autosomal genes are from C57BL10. Also, these types of interactions are often detected when knockout mutants in mice are transferred from one genetic background to another. For example, the knockout for the NOS-1 gene that increases attacks in one genetic background is lost after generations of backcrossing to C57BL6 mice (LeRoy et al., 2000).

In these examples with mice, the interacting genes are not known; however, interactions of pairs of genes are now being investigated in fruit flies. Epistatic interactions have been shown for recessive mutants with effects on olfaction (Fedorowicz,Fry,Arholt,&McKay,1998).Othersarelooking at how allelic substitutions in a gene affect patterns of expression of many genes as a way to identify systems of interacting genes (Wahlsten, 1999). DNA microchips are used in this research. Greenspan (2001) has suggested that these systems of interactions may be very complex—networks of different genes can have the same behavioral effects and networks of the same genes can have different behavioral effects.

Genes, Environment, and Behavior

Effects of genes on behavior are dependent on the environment, just as effects of the environment on behavior are dependent on genes. For example, experience has effects on gene expression. Some of these are due to effects of experiences on levels of steroid hormones and thereby on gene expression in neurons. Others are due to effects of experience on synaptic transmission, which thereby affect gene expression in neurons. In this section, the effects of experience on gene expression are described for the circadian clock in mice; maternal care and pup development in rats; learning of birdsong; and long-term memory in mollusks, in fruit flies, and in mice.

If mice are kept in constant darkness, their circadian rhythm for wheel-running activity shows a cycle of about 24 hours, but the onset of activity drifts with each day in total darkness. The active period is known as subjective day, and the inactive period is known as subjective night. When mice and other rodents are exposed to light during the subjective night, the onset time for activity is shifted and the transcription of c-fos and several other immediate early genes in the suprachiasmatic nucleus (SCN) is induced (Kronhauser, Mayo, & Takahashi, 1996). Immediate early genes such as c-fos code for transcription factors that regulate the expression of other genes. It is believed that the expression of these genes resets the circadian clock in the SCN.

When crouching over their pups in arched-back nursing (ABN), mother rats lick and groom (LG) their pups (Meaney, 2001). This licking subsequently affects gene expression in the brains and the behavioral responses to stress of adult rats. There is variation in how much a mother rat licks and grooms her pups. As adults, pups with more LG have higher levels of corticotropin-releasing factor (CRF) mRNA in the hypothalamus and of glucocorticoid receptor mRNA in the hippocampus than do those with less LG. As adults, those with high-LG mothers are less fearful than are those with low-LG mothers. Also, females that had high-LG mothers lick their own pups more than do those that have low-LG mothers. This may represent one type of mechanism for nongenomic transfer of behavior across generations.

In some birds, there is a sensitive period for song learning. If at that time canaries or zebra finches are exposed to the song oftapedtutorsoftheirownspecies,theimmediateearlygenes, Zenk and c-fos are induced in the forebrain structures NCM (caudal part of the neostriatum) and cHV (hyperstriatum ventral; Mello, Vicario, & Clayton, 1992). These regions appear to be involved in song learning and not in song production. This increase does not occur when each species is exposed to the song of another species or to simple bursts of sound.

Long-term memory for sensitization and classical conditioning in the mollusk Aplysia, classical conditioning in fruit flies, and spatial learning in mice involve changes in gene expression in the brain (Squire & Kandel, 2000). In each of these, the neural events of learning activate the transcription factor CREB (cyclic adenosine monophosphate response element binding protein); this then turns on other genes that cause changes in structure and function of nerve cells involved in long-term memory. For example, fruit flies are trained to associate an odor with an electric shock and another odor with the absence of an electric shock (Dubnau & Tully, 1998). The memory for these associations is tested by allowing them to choose between chambers with the two different odors. Flies have short-term memory lasting less than an hour and long-term memory lasting 24 hours or more. Flies were engineered to make CREB in response to heat shock. In non-heat-shocked flies, a single odor and electric shock pairing produces short-term but not long-term memory, whereas in heat-shocked flies, a single odor and electric shock pairing also produces long-term memory. Conversely, flies were engineered to make a protein that blocks transcription by CREB. Induction of this protein by heat shock blocked long-term but not short-term memory.

Summary

Genes are being used as Ginsburg (1958) proposed as tools to study the brain mechanisms of behavior and the effects of the environment on the development of behavior. It is of interest not only that the protein products of genes are involved in both, but also that the expression of genes are involved in both; this implies that although we can experimentally separate and manipulate genes and environment, they are in fact two sides of the same coin. For this reason, we should look beyond the old and tired nature-nurture controversy.

The Genetics of Behavioral Evolution

The evolution of behavioral adaptations is due to effects of natural selection on gene frequency. Details on population genetics may be found in Falconer and McKay (1997) and on molecular evolution in Page and Holmes (1998). In behavior genetics, the goal is to identify the genes and the changes in them that are the basis of the evolution of species differences in behavioral adaptations. Three examples are discussed in this section: the evolution of courtship song in fruit flies, the evolution of mating systems in voles, and the evolution of color vision in mammals. Respectively, these examples illustrate behavioral evolution due to change in the structural part of a gene, change in the regulatory part of the gene, and gene duplication with changes in both parts of the gene.

Evolution of the Courtship Song in Fruit Flies

Courtship by male fruit flies consists of several motor patterns. In one of these patterns, the male vibrates a wing; the sound produced is known as the courtship song. In D. melanogaster, variants of the period (per) gene affect this song. One of these abolishes the song; it is known as per⁰¹. Others change the timing of the wing vibrations and thereby the quality of the song. The song also varies among fruit fly species; furthermore, females of a species may prefer to mate with males singing their species song. This species variation is in part due to variation in the per gene (Colot, Hall, & Robash, 1998). When a cloned copy of the D. similans per gene was inserted into the genome of D. melanaogaster per⁰¹, the male courtship song is restored but it resembles that of D. similansratherthanthatofD.melanogaster(Wheeleret al., 1991). Further analysis suggested that four or fewer amino acid substitution in a nonconserved region of the per protein may control the species difference in courtship song.

Evolution of Mating Systems in Voles

Prairie voles are monogamous and montane voles are polygamous.These two species also differ in partner preference, parenting, social contact, and aggression (DeVries, Taymans, & Carter, 1997). These species differences are due—at least in part—to species variants of an arginine vasopressin receptor (V1aR) gene (Young, Winslow, Nilsen, & Insel, 1997). The structural part of the gene is the same in the two species, but the regulatory regions differ between the two species. As a consequence, there is species variation in the expression of mRNA in regions of the forebrain and therefore in the forebrain distribution of the V1aR. Several lines of evidence suggest that this between-species neural variation may be involved in the between-species behavioral variation:

• Injectionofargininevasopressin(AVP)intothebrainfacilitates aggression in prairie voles and not in montane voles, whereas it facilitates auto-grooming in montane but not in prairie voles. Also, partner preference in prairie voles is blocked by brain injection of anAVPantagonist and facilitated by brain injection ofAVP(Winslow et al., 1993).
• There are mice with a transgene for the prairie vole V1aR gene(Young,Nilsen,Waymire,MacGregor,&Insel,1999). These mice have the same brain distribution of this receptor as prairie voles, and these mice exhibit increased affiliative behavior after brain injection of arginine vasopressin.
• Aviral vector was used to insert another copy of the V1aR into neurons of a region of the basal ganglia in prairie voles (Pitkow et al., 2001). This increased V1aR binding in this region of the basal ganglia, and these males exhibited increased anxiety, affiliative behavior, and partner preference.

Evolution of Color Vision in Mammals

Most old world monkeys, great apes, and humans can discriminate between red and green, whereas all other mammals cannot (Bowmaker, 1991); this is—at least in part—due to the former’s having three types of cones in the retina and the latter’s having two types of cones. Each cone has its own opsin. Opsins are the proteins that confer spectral sensitivity on the visual receptors. Old world monkeys, great apes, and humans—but not other mammals—have cones with an opsin that is sensitive to the red end of the spectrum (Nathans, Thomas, & Hogness, 1986; Nathans, Merbs, Sung, Weitz, & Wang, 1992). The gene for the opsin sensitive to the red end of the spectrum is on the X chromosome of old world monkeys, great apes, and humans, and the gene for the opsin sensitive to the green region of the spectrum is on the X chromosome of all mammals. About 30 million years ago, the gene for the red opsin arose by gene duplication from the gene for the green opsin.At first, there were two green opsin genes. Then mutations accumulated in the structural region of one of these genes, changing its spectral sensitivity from green to red. There are seven amino acid differences between the red and the green opsins that appear to be critical for the difference in their spectral sensitivity. Changes in the DNA sequence of regulatory elements are likely; for example, the red opsin is synthesized in one cone and the green in another. Thus, after gene duplication, there are accumulated mutations in both the structural and regulatory regions of the second copy of the gene such that it came to code for a new protein with new expression patterns. This evolution by gene duplication has occurred many other times for other genes. For example, all the ion channel genes arose by gene duplications in common with the evolution of most of the serotonin receptor genes (Smith, 1991). Such genes are in families with similar DNA sequences. It is very likely that there were also behavioral consequences of these gene duplications.

Summary

It has been suggested that many genes are usually involved in the evolution of adaptations and speciation. But sometimes, as described previously, a single gene contributes most to the evolution of a trait. Also, it is possible for these genes to identify the DNAvariant with effect on the behavior that has been subject to natural selection. In time, the molecular basis of many species differences in behavior and their evolution will be understood. For example, we will know just what are the sequence and functions of the 2% of DNA that differs between chimpanzees and humans and how this difference contributes to the species differences in behaviors.

Future Directions

The completion of the respective genome projects in nematodes, flies, mice, and humans will make it possible to identify all of the protein coding genes of these species as well as whereandwhenthegenesaretranscribed,andthenewprotein initiative will eventually identify the structural conformation as well as metabolic or cellular function of each protein. This will greatly ease the task of identifying all the genes that can and do cause a behavior to vary in these four species, as well as that of tracing the pathways from gene to behavior. The great challenge will then be to understand how genes interact with each other, how they interact with the environment in the developmentandexpressionofbehaviors,andhowtheyrelate to behavioral evolution.

The study of the genetics of behaviors in animals can and should be for more than just the development of models relevant to human behaviors. The genetics of animal behaviors should also be researched in order to discover general principles relating genes to behavior across animal species and to have a comparative genetics of adaptive behaviors within related species. For this, there will need to be genome projects in other taxonomic groups; such work is already taking place on bees and other insects, many farm animals, domestic dogs, domesticcats,otherrodents,andmanyprimates;Ibelievethat this process represents the future of behavior genetics.

Bibliography:

1. Archer, J. (1973). Tests of emotionality on rats and mice: A review. Animal Behavior, 21, 205–235.
2. Allada, R., White, N. E., So, W. V., Hall, J. C., & Rosbash, M. (1998). A mutant Drosophila homolog of mammalian Clock disrupts circadian rhythms and transcription of period and timeless. Cell, 93, 781–804.
3. Antoch, M. P., Song, E. J., Chang,A. M., Vitaterna, M. H., Zhao, Y., Wilsbacher, L. D., Sangoram, A. M., King, D. P., Pinto, L. H., & Takahashi,J.S.(1997).Functionalidentificationofthemousecircadian Clock gene by transgenic BAC rescue. Cell, 89, 655–667.
4. Bakker, Th. C. M. (1994). Evolution of aggressive behavior in the threespine stickelback. In M. A. Bell & S. A. Foster (Eds.), The evolutionary biology of the threespine stickleback (pp. 345–380). Oxford, UK: Oxford University Press.
5. Belknap, J. K., Dubay, C., Crabbe, J. C., & Buck, K. J. (1997). Mapping quantitative trait loci for behavioral traits in the mouse. In K. Blum & E. P. Npbel (Eds.), Handbook of psychiatric genetics (pp. 435–453). Boca Raton, FL: CRC Press.
6. Benjamin, J., Li, L., Patterson, C., Greenburg, B. D., Murphy, D. L., &Hamer,D.H.(1996).Populationandfamilialassociation between the D4 dopamine receptor gene and measures of novelty seeking. Nature Genetics, 12, 81–84.
7. Benzer, S. (1971). From genes to behavior. Journal of the American Medical Association, 218, 1015–1022.
8. Bowmaker, J. K. (1991).The evolution of vertebrate visual pigments and photoreceptors. In J. R. Cronly-Dilllon & R. L. Gregory (Eds.), Vision and visual dysfunction: Vol. 2. Evolution of the eye and visual systems (pp. 63–81). Boca Raton, FL: CRC Press.
9. Broadhurst,P.L.(1960).Experimentsinpsychogenetics:Application of biometrical genetics to the inheritance of behavior. In H. J. Eysenck(Ed.),Experimentsinpersonality:Vol.1.Psychogenetics and psychopharmacology (pp. 1–102). London: Routledge.
10. Chiavegatto, S., Dawson, V. L., Mamounas, L. A., Koliatsos, V. E., Dawson, T. M., & Nelson, R. J. (2001). Brain serotonin dysfunction accounts for aggression in male mice lacking neuronal nitric oxide synthase. Proceedings of the National Academy of Sciences, USA, 98, 1277–1281.
11. Cloninger, C. R., Svrakic, D. M., & Przybeck, T. R. (1993). A psychobiological model of temperament and character. Archives of General Psychiatry, 50, 975–990.
12. Colot, H. V., Hall, J. C., & Robash, M. (1988). Interspecific comparison of the period gene of Drosophila reveals large blocks of non-conserved coding DNA. European Molecular Biology Organization Journal, 7, 3929–3937.
13. Crawley, J. N. (2000). What’s wrong with my mouse? New York: Wiley-Liss.
14. DeFries, J. C., Gervais, M. C., & Thomas, E. A. (1978). Response to 30 generations of selection for open-field activity in laboratory mice. Behavior Genetics, 8, 3–13.
15. DeVries, A. C., Taymans, S. E., & Carter, C. S. (1997). Social modulation of corticosteroid responses in male prairie voles. Annals of the New York Academy of Sciences, USA, 807, 494–497.
16. Dubnau, J., & Tully, T. (1998). Gene discovery in Drosophila: New insights for learning and memory. Annual Review of Neuroscience, 21, 407–444.
17. Ehrman, L., & Parsons, P. A. (1981). Behavior genetics and evolution. New York: McGraw-Hill.
18. Ebstien, R. B., Novick, O., Umansky, R., Priel, B., Osher, Y., Blaine, D., Benett, E. R., Nemanov, R., Katz, M., & Belmaker, R. H. (1995). Dopamine D4 receptor (D4DR) exon III repeat polymorphism associated with the human personality trait of novelty seeking. Nature Genetics, 12, 78–80.
19. Falconer, D. S., & McKay, T. F. C. (1997). Introduction to quantitative genetics (4th ed.). Harlow, UK: Longman.
20. Fedorowicz, G. M., Fry, J. D., Arholt, R., & McKay, T. F. (1998). Epistatic interaction between smell-impaired loci in Drosophila melanogaster. Genetics, 148, 1885–1891.
21. Flint, J., Corely, R., DeFries, J. C., Fulker, D. W., Gray, J. A., Miller, S., & Collins, A. C. (1995). A simple genetic basis for a complex psychological trait in laboratory mice. Science, 269, 1432–1435.
22. Gerlai, R. (1996). Gene-targeting studies of mammalian behavior: Is it the mutation or the background genotype? Trends in Neuroscience, 19, 177–181.
23. Gilbert, S. F. (2000). Developmental biology (6th ed.). Sunderland, USA: Sinauer.
24. Ginsburg, B. E. (1958). Genetics as a tool in the study of behavior. Perspectives in Biology and Medicine, 1, 397–424.
25. Greenspan, R. J. (2001). The flexible genome. Nature Reviews Genetics, 2, 383–387.
26. Grilo, C. M., & Pogue-Geille, M. F. (1991). The nature of environmental influences on weight and obesity: A behavior genetic analysis. Psychological Bulletin, 10, 520–537.
27. Jallon, J.-M. (1999). Genetic analysis of Drosophila sexual behavior. In B. C. Jones & P. Mormede (Eds.), Neurobehavioral genetics: Methods and applications (pp. 365–372). Boca Raton, FL: CRC Press.
28. King, D. P., & Takahashi, J. S. (2000). Molecular genetics of circadian rhythms in mammals. Annual Review of Neuroscience, 23, 713–742.
29. King, D. P., Zhao, Y., Sangoram, A. M., Wilsbacher, L. D., Tanaka, ,Antoch,M.P.,Steeves,T.D.L.,Vitaterna,M.H.,Kornhauser, J. M., Lowery, P. L., Turek, F. W., & Takahashi, J. S. (1997). Positional cloning of the mouse circadian Clock gene. Cell, 89, 641–643.
30. Kronhauser, J. M., Mayo, K. E., & Takahashi, J. S. (1996). Light, immediate early-genes, and circadian rhythms. Behavior Genetics, 26, 221–240.
31. LeRoy, I., Pothion, S., Mortaud, S., Chabert, C., Nicolas, L., Cherfouh, A., & Roubertoux, P. L. (2000). Loss of aggression after transfer onto a C57BL/6J background in mice carrying a targeted disruption of the neuronal nitric oxide synthase gene. Behavior Genetics, 30, 367–373.
32. Lewin, B. (1997). Genes (Vol. 6). Oxford, UK: Oxford University Press.
33. Lynch, C. B. (1994). Evolutionary inferences from genetic analyses of cold adaptation in laboratory and wild populations of the house mouse. In C. R. B. Boake (Ed.), Quantitative genetic studies of behavioral evolution (pp. 278–301). Chicago: University of Chicago Press.
34. Manning, A. (1961). The effect of artificial selection for mating speed in Drosophila melanogaster. Animal Behavior, 9, 82–92.
35. Maxson, S. C. (1992). Methodological issues in genetic analyses of an agonistic behavior (offense) in mice. In D. Goldowitz, D. Wahlsten, & R. Wimer (Eds.), Techniques for the analysis of brain and behavior: Focus on the mouse (pp. 349–373). New York: Elsevier.
36. Maxson, S. C. (1999). Aggression: Concepts and methods relevant to genetic analyses in mice and humans. In B. C. Jones & P. Mormede (Eds.), Neurobehavioral genetics: Methods and applications (pp. 293–300). Boca Raton, FL: CRC Press.
37. Maxson, S. C., Roubetoux, P. L., Guillot, P.-V., & Goldman, D. (2001). The genetics of aggression: From mice to humans. In M. Martinez (Ed.), Prevention and control of aggression and the impact on its victims (pp. 71–81). New York: Kluwer.
38. Maxson, S. C., Trattner, A., & Ginsburg (1979). Interaction of Y-chromosomal and autosomal genes in the development of aggression in mice. Behavior Genetics, 9, 219–226.
39. Meaney, M. J. (2001). Maternal care, gene expression and the transmission of individual differences in stress reactivity across generations. Annual Review of Neuroscience, 24, 1161–1192.
40. Mello, C., Vicario, D. S., & Clayton, D. F. (1992). Song presentation induces gene expression in the songbird forebrain. Proceedings of the National Academy of Sciences, USA, 89, 6819–6822.
41. Miczek, K. A., Maxson, S. C., Fish, E. W., & Faccidomo, S. (2001). Aggressive behavioral phenotypes in mice. Behavioral Brain Research, 125, 167–181.
42. Mirnics, K., Middleton, F. A., Lewis, D. A., & Levitt, P. (2001). Analysis of complex brain disorders with gene expression microarrays: Schizophrenia as a disease of the synapse. Trends in Neuroscience, 24, 479–486.
43. Nathans, J., Merbs, S. L., Sung, C.-H., Weitz, C. J., & Wang, Y. (1992). Molecular genetics of human visual pigments. Annual Review of Genetics, 26, 403–424.
44. Nathans, J., Thomas, D., & Hogness, D. S. (1986). Molecular genetics of human color vision: The genes encoding blue, green, and red pigments. Science, 232, 193–202.
45. Nelson, R. J. (1997). The use of genetic “knock-out” mice in behavioral endocrinology research. Hormones and Behavior, 31, 188–196.
46. Nelson, R. J., & Chiavegatto, S. (2001). Molecular basis of aggression. Trends in Neuroscience, 24, 713–719.
47. Nelson, R. J., Demas, G. E., Huang, P., Fishman, M. C., Dawson, V., Dawson, T. M., & Snyder, S. H. (1995). Behavioural abnormalities in male mice lacking neuronal nitric oxide synthase. Nature, 378, 383–386.
48. Nelson, R. J., & Young, K. A. (1998). Behavior in mice with targeted disruption of single genes. Neuroscience and Biobehavioral Reviews, 22, 453–462.
49. Nisenbaum, L. K. (2002). The ultimate chip shot: Can microarray technology deliver for neuroscience? Genes, Brain, and Behavior, 1, 27–34.
50. Page, R. D. M., & Holmes, E. C. (1998). Molecular evolution: A phylogenetic approach. Malden: Blackwell.
51. Pitkow, L. J., Sharer, C. A., Ren, X., Insel, T. R., Terwilliger, E. F., & Young, L. J. (2001). Facilitation of affiliation and pair-bond formationbyvasopressinreceptorgenetransferintotheventralforebrain of a monogamous vole. Journal of Neuroscience, 21, 7392–7396.
52. Plomin, R., & Caspi, A. (1998). DNA and personality. European Journal of Personality, 12, 387–407.
53. Plomin, R., & Crabbe, J. (2000). DNA. Psychological Bulletin, 126, 806–828.
54. Plomin, R., DeFries, J. C., McClearn, G. E., & McGuffin, P. (2000). Behavioral genetics (4th ed.). New York: Worth.
55. Sandberg, R., Yasuda, R., Pankratz, D. G., Carter, T. A., Del Rio, J. A., Wodicka, L., Mayford, M., Lockhart, D., & Barlow, C. (2000). Regional and strain-specific gene expression mapping in the adult mouse brain. Proceedings of the National Academy of Sciences, USA, 97, 11039–11043.
56. Segalat, L. (1999). Genetic analysis of behavior in the nematode, Caenorhabditis elegans. In B. C. Jones & P. Mormede (Eds.), Neurobehavioral genetics: Methods and applications (pp. 373– 381). Boca Raton, FL: CRC Press.
57. Smith, C. U. M. (1991). Elements of molecular neurobiology (2nd ed.). New York: Wiley.
58. Sokolowski, M. B. (1999). Genetic analysis of food search behavior in the fruit fly, Drosphilia melanogaster. In B. C. Jones & P. Mormede (Eds.), Neurobehavioral genetics: Methods and applications (pp. 357–364). Boca Raton, FL: CRC Press.
59. Sokolowski, M. B., & Riedl, C. A. L. (1999). Behavior-genetic and molecularanalysisofnaturallyoccurringvariationinDrospophila larval foraging behavior. In W. E. Crusio & R. T. Gerlai (Eds.), Handbook of molecular-genetic techniques for brain and behavior research (pp. 496–511).Amsterdam: Elsevier.
60. Squire, L. R., & Kandel, E. R. (2000). Memory: From mind to molecules. New York: Scientific American Library.
61. Takahashi, J. S., Pinto, L. H., & Vitaterna, M. H. (1994). Forward and reverse genetic approaches to behavior in the mouse. Science, 264, 1724–1732.
62. Talbot,C.J.,Nicod,A.,Cherny,S.S.,Fulker,D.W.,Collins,A.C.,& Flint, J. (1999). High resolution mapping of quantitative trait loci in outbred mice. Nature Genetics, 21, 305–308.
63. Tsien, J. Z. (1999). Brain region-specific and temporally restricted gene knockout using the Cre recombinase system. In W. E. Crusio & R. T. Gerlai (Eds.), Handbook of molecular-genetic techniques for brain and behavior research (pp. 282–290). Amsterdam: Elsevier.
64. Vitaterna, M. H., King, D. P., Chang, A. M., Kornhauser, J. M., Lowrey, P. L., McDonald, J. D., Dove, W. F., Turek, F. W., & Takahashi, J. S. (1994). Mutagenesis and mapping of a mouse gene, Clock, essential for circadian behavior. Science, 264, 719–725.
65. Wahlsten, D. (1999). Single-gene influences on brain and behavior. Annual Review of Psychology, 50, 599–624.
66. Wallace, D. C. (1999). Mitochondrial diseases of man and mouse. Science, 238, 1482–1488.
67. Watson, J. D., & Crick, F. H. C. (1953a). Molecular structure of nucleic acids: A structure for deoxyribose nucleic acid. Nature, 171, 737–738.
68. Watson, J. D., & Crick, F. H. C. (1953b). Genetical implications of the structure of deoxyribosenucleic acid. Nature, 171, 964–967.
69. Winslow, J., Hastings, N., Carter, C. S., Harbaugh, C., & Insel, T. R. (1993). A role of central vasopressin in pair bonding in monogamous prairie voles. Nature, 365, 545–548.
70. Wheeler, D. A., Kyriacou, C. P., Greenacre, M. L., Yu, Q., Rutila, J. E., Rosbash, M., & Hall, J. C. (1991). Molecular transfer of a species-specific behavior from Drospohila simulans to Drosophila melanogaster. Science, 251, 1082–1085.
71. Young, L. J., Nilsen, R., Waymire, K. G., MacGregor, G. R., & Insel, T. R. (1999). Increased affiliative response in vassopressin mice expressing the V₁a receptor from a monogamous vole. Nature, 400, 766–768.
72. Young, L. J., Winslow, J. T., Nilsen, R., & Insel, T. R. (1997). Species differences in V₁a receptor gene expression in monogamous and nonmanogamous voles: Behavioral consequences. Behavioral Neuroscience, 111, 599–605.