P. L. Broadhurst

Studies in genotype-environment interaction. I. Methodology and preliminary multivariate analysis of a diallel cross of eight strains of rat

Previous studies concerned with genotype-environment interaction in behavior are discussed from methodological and gene-action points of view. A number of methodological advances have been made recently. In particular, designs introducing environmental variables factorially into standard genetic designs are extremely powerful. It is evident that all the methods appropriate to the analysis of factorial designs may be employed to investigate the interaction between such macroenvironmental variables and genotype. The main advantage of adopting such designs in psychogenetic studies is that much more complex behaviors may be investigated. Studying the gene-action controlling behavior as a means of increasing our understanding of its adaptive and evolutionary significance has been the aim of a number of recent psychogenetic studies. This approach applied to studies of genotype-environment interaction makes investigating the adaptive significance of these complex behaviors possible. A number of previous interaction studies are discussed from this viewpoint. The general finding is that gene action ensures that the hybrid genotypes, typical of natural populations, are buffered against environmental treatments aimed at producing deleterious and disruptive effects. However, an interesting study by Henderson, where the treatment was aimed at enrichment of the environment, revealed no such buffering effect. Instead, an increase in the range of inherited individual differences was found, differences which remained hidden in the control group. The main purpose of this report is to present findings of a replicated diallel cross of eight strains of rats employing an environmental treatment designed to investigate genotype-environment interaction. One group was subjected to infantile stimulation by early handling, and a control group was not handled. The principal measures taken in adulthood were open-field ambulation and defecation, two escape-avoidance behaviors, and body-build index. A preliminary analysis, employing a multivariate extension of Haymans analysis of variance, is carried out describing the main findings. Each of the five measures in the analysis is shown to be under substantial independent genetic control, although a number of interesting interrelationships also emerge. Many interactions with stimulation are indicated for all measures. It is concluded that each measure should be subjected to a univariate analysis.

Genetics of escape-avoidance conditioning in laboratory and wild populations of rats: A biometrical approach

The interest of biometrical geneticists in the genetic architecture of behavior is explained with reference to the additive, dominance, and epistatic components of variation and their relation to evolutionary pressures. For one phenotype, escape-avoidance conditioning inRattus norvegicus, a fairly complete description of its genetic architecture has been gradually built and the major conclusions from four studies of this phenotype are reported: a selection study initially demonstrated the presence of large amounts of additive genetic variation and produced phenotypically extreme lines needed for later work; a diallel cross provided the opportunity for detailed examination of the dominance effects; a triple test cross permitted a similar examination of epistatic effects; and finally, another triple test cross using wild rats provided a confirmatory first attempt to test the assumption that a wild populations genetic architecture did not differ markedly from that found in laboratory populations. In relating the genetic findings to the evolutionary significance of behaviors in the escape-avoidance paradigm, it is argued that interspecific comparisons might play a major role.

A simplified triple-test cross analysis of alcohol preference in the rat

Reanalyses of first-degree biometrical genetic data from previous studies of alcohol preference in the mouse revealed little consistency beyond a basic additive genetic component. A simplified triple-test cross in the rat investigated the genetic architecture of alcohol preference for a 10% (w/v) alcohol solution or water. An initial survey of eight selected and inbred strains identified high- and low-scoring strains, the MNR and the ACI, respectively, which were crossed as tester lines to six strains (the RHA, RLA, TMB, TMD, MNR, and ACI) to produce the required set of largely F1 families. The additive-dominance model proved adequate for males, and directional dominance for low alcohol preference was found on all three measures: alcohol intake, alcohol preference ratio, and alcohol calorie contribution ratio. For females the model was adequate only for alcohol preference ratio, which showed ambidirectional dominance. The relevance of such genetic architecture to an animal model of alcoholism and to the evolution of alcohol drinking in the rat is discussed.

Analysis of two-way escape-avoidance conditioning measures from a diallel cross of eight strains of rats

In a study designed to investigate genotype-environment interaction, eight strains of laboratory rats were crossbred in a replicated diallel cross employing infantile stimulation and its absence as environment treatments. This paper reports on measures of the acquisition of two-way escape-avoidance conditioning, comprising number of avoidances, avoidance and escape latencies, and intertrial and presessional crossings, which were subjected to biometrical genetical analysis, all but the last successfully. Additive variation was prominent throughout and some measures showed directional dominance. Effects of stimulation were seen in avoidance number and crossings. The analysis of avoidances by successive blocks of trials using covariance:variance graphs revealed differences in the way the strains varied with respect to the changing relationships of proportions of dominant and recessive alleles governing this behavior. The results are discussed in the light of previous data and of their evolutionary implications.

The genetic architecture of hyponeophagia and the action of diazepam in rats

Rats of 18 genotypes derived from the Roman selected strains were tested for inhibition of feeding due to novelty (hyponeophagia) in the absence or presence of 1 mg/kg diazepam. The resulting data from three behavioral indices were subjected to the Hayman [(1954). Biometrics10:235–244], variance/covariance, Mendelian cross, and single-test cross analyses. Additive genetic variation, directional dominance for high neophobia, and some nonallelic interaction were detected. The genetic architecture of the separate behavioral indices, and its modification in the drugged subjects, was discussed in relation to evolutionary adaptation and the anxiolytic and appetite-enhancing actions of the drug.

The genetics of alcohol preference in the female rat confirmed by a full triple-test cross

Because our previous (1981) simplified triple-test cross analysis of alcohol preference for a 10% (w/v) alcohol solution or water showed failure of the simple additive-dominance genetic model among females, additional breeding was undertaken to produce the full triple-test cross. An L3 tester strain, and F1 cross between the MNR and the ACI strains, previously used as L1 and L2 testers, respectively, was bred and crossed to six strains: RHA, RLA, TMB, TMD, MNR, and ACI. The previous failure of the model was found to be due to inadequate testers sharing common loci, rather than to epistasis. Unbiased estimates of genetic architecture in the females were similar to those already found for males, revealing directional dominance for low alcohol preference.

The Maudsley reactive and nonreactive strains of rats: A clarification

The origin of the Maudsley reactive and nonreactive strains of rats and the construction of the survey of researches using them are briefly reviewed in the light of Archers criticisms. The reasons for reasserting that they represent a valid dichotomy in emotionality are summarized.