Personality traits in animals are often measured using standardised behavioural tests for activity, boldness/shyness, sociability, aggression and exploration. These tests are quick and convenient, as well as easy to repeat. As the interest in studying the impact of animal personality on ecological and evolutionary consequences has been growing rapidly, there is increasing focus on cross‐validating measurements taken during these tests with behaviours shown under natural situations. In our experiment, we aimed to study the relationship between standardised measurements for activity, exploration and anxiety‐like behaviour measured in Open Field, Novel Object and Elevated Plus Maze tests with exploration and colonisation in semi‐natural conditions. We carried out a semi‐natural enclosure experiment in parallel with standardised behavioural tests, creating a scenario similar to an invasion or dispersal event. We compared behaviours in standardised tests and in enclosures for animals of two populations of wild house mice (Mus musculus domesticus). Several behavioural variables taken during the standardised tests, such as distance moved and time spent with novel object, were negatively correlated with space‐use in the enclosure while being highly positively correlated among each other. Based on their relationship with space use, we refer to behavioural measurements from standardised tests as activity/exploration. The time spent near the walls in an open field, probably reflecting anxiety, was not correlated to any other variable or the behaviour in the enclosure. In addition, we found differences in activity/exploration behaviour between the two populations in the standardised tests, but not during the colonisation of the novel environment. These results emphasise that researchers have to be careful when trying to extrapolate behaviour shown in standardised laboratory test setups to more natural, ecologically relevant situations. This has to be taken into account in distantly related species but even when studying the wild relative of laboratory rodents, for which these standardised tests have originally been developed.

The Prader-Willi Syndrome (PWS) gene region is an imprinted gene complex involved in behavioral, metabolic and osteogenic functions. We have analyzed here the variation of two families of regulatory small nucleolar RNAs (SNORD115 and SNORD116) that are coded within the PWS and are 30 expressed from the paternal chromosome. They are organized in two tandemly repeated clusters which are naturally copy number variable between individuals. We find that the copy numbers at these loci correlate with repeatable individual test scores for anxiety that are considered to constitute a component of the "personality" of individuals. We show this for different populations and species of mice, cavies and for the anxiety component of personality tests in humans. This is also the case for an 35 inbred mouse strain (C57Bl6) implying that copy number variation creates phenotypic variability even in an isogenic background. In transcriptome data from brain samples of this strain we find SNORD copy-number correlated regulation of target genes that are known to be involved in 2 influencing behavior. SNORD115 has previously been suggested to regulate splicing of the serotonin receptor Htr2c and we confirm this in our data. For SNORD116 we provide evidence that it regulates 40 the expression level of the chromatin regulator Ankrd11, which itself regulates GABA receptors, metabolic pathways, cell differentiation and osteogenesis. Intriguingly, we find that craniofacial shapes in mice correlate also with SNORD116 copy numbers. New copy number variants are generated at very high rates in mice, possibly at every generation, explaining why conventional genetic mapping could not detect this association. Our results suggest that the variable dosage of two 45 regulatory RNAs are major determinants of individual behavioral differences and correlated traits in mammals.