Life-history trade-offs: are they linked to personality in a precocial mammal (Cavia aperea)?

Abstract

Life-history trade-offs are predicted to contribute to the maintenance of personality variation. Individuals with ‘fast’ lifestyles should develop faster, reproduce earlier and exhibit more risky behaviours. Evidence for such predicted links, however, remains equivocal. Here, I test how growth rate, timing of maturation, litter size and maternal effort correlate with exploration, boldness, fearlessness, docility and escape latency. I found several links that were predicted by recent theory while others were against theoretical predictions, e.g. fast-growing individuals were more fearful. Thus, while I found personality to be integrated with life history, I cannot fully support recent hypotheses aiming to explain such behaviour–life-history associations.

1. Introduction

Within populations, individuals typically vary consistently in behaviour, often referred to as animal personality. Life-history trade-offs may explain variation in behaviour when alternative life-history strategies have similar fitness outcomes [1–3]. Several major trade-offs such as growth versus mortality, reproduction versus self-maintenance or current versus future reproduction have received considerable attention in explaining between-individual variation in behaviour [2–4]. Individuals with a fast lifestyle were predicted to grow fast, reproduce early, exhibit risk-prone behaviours and die younger than individuals with a slow lifestyle. Behaviours facilitating resource acquisition but potentially reducing survival, such as exploration, boldness or aggressiveness, were predicted to correlate with life-history traits at both the genetic and phenotypic level [4,5].

Theoretical predictions and empirical evidence from human populations in addition predict links between risk-taking behaviours and parental investment into offspring quantity versus quality [6], suggesting two distinct reproductive strategies with associated behavioural tendencies: fast-living, proactive and risk-taking individuals should favour offspring quantity over quality while reactive, risk-averse individuals should emphasize offspring quality [6]. Although the importance of parental allocation in linking between-individual differences in behaviour with fitness has recently been emphasized [7], studies in animals have not yet tested links between risk-taking behaviours and allocation into offspring number versus size.

Empirical studies found contradictory results regarding predicted links between behaviour and life history [8–10]. Reasons might be that behaviour–life-history relationships are only apparent at specific levels of variation, i.e. at the genetic or the phenotypic level [10] or are population or environment-specific [4,11]. In addition, studies often focused on one or few behavioural and life-history traits and thus might have missed the relevant trait combinations in the respective species [3]. To test behaviour–life-history relationships, we should therefore investigate several behavioural and life-history traits simultaneously in the same individuals.

Here, I test the predicted association between four life-history traits (growth, maturation timing and after onset of reproduction: litter size and maternal effort) and five behavioural traits (boldness, exploration, fearlessness, escape latency and docility) in female wild cavies (Cavia aperea). First, I estimate the repeatability of life-history and behavioural traits to estimate their consistency, then, I estimate among-individual correlations between all traits.

2. Material and methods

(a). Animals

Females were bred under standard laboratory housing conditions (enclosure size 0.8 m²) at 20°C. Hay, guinea pig pellets (Höveler, Germany) and water were available ad libitum and supplemented with fresh green every other day throughout life.

I investigated only females, because determination of maturation in males requires regular blood sampling, potentially causing stress and affecting their behaviour. At weaning, females were separated from their mother and siblings and kept in pairs with other young females or were introduced to sexually mature males. Individuals in female–female pairs were checked daily for signs of vaginal opening, indicating oestrus. In females that were kept together with males, timing of maturation was determined by subtracting the length of pregnancy (60 days) from the date of first parturition. To obtain estimates for growth rate, females were weighed weekly (six times) until an age of 45 days.

Reproduction occurred under conditions of ad libitum food and water either in the laboratory or in outside enclosures with a hut providing shelter. Maternal mass was determined when the male was introduced for breeding. Maternal mass divided by litter mass at birth was interpreted as maternal effort.

Escape latency was measured twice during early development (day 3–4 and 8–9). Likewise, docility was measured at an age of 6–7 days and 12–14 days. Boldness, exploration and fearlessness were measured for the first time at weaning before females were separated from their mothers. Thereafter, behavioural tests were repeated at different ages to a maximum of 1 year when females were not reproductively active. Not all females participated repeatedly in every behavioural test (table 1). The order in which animals were tested for boldness, exploration and fearlessness was randomized.

Table 1.

Life-history and behavioural traits. Traits are measured in 100 female C. aperea. Data for escape latency and docility were obtained from n = 54 animals while all 100 animals were tested at least once for all other variables. N_total gives the total amount of data for the respective trait, N_{per animal} shows the mean number of measurements per individual, mean and s.d. refer to the population mean and standard deviation. R is the trait repeatability (credible interval) and the associated p-value. Maturation was measured as the first opening of the vaginal membrane and hence no repeatability could be calculated.

trait	Ntotal	Nper animal	mean	s.d.	R	p-MCMC
litter size (no.)	233	2.4	2.85	1.18	0.10 (0–0.26)	0.6
maternal effort (%)	231	2.3	34.0	11	0.17 (0.01–0.32)	0.02
growth rate (g/d)	288	2.8	5.18	1.57	0.31 (0.14–0.38)	0.01
maturation (d)	100	1	51.5	21.88	—	—
boldness (s)	203	2.2	1370.7	1472	0.44 (0.27–0.58)	<0.001
exploration (no.)	125	2.4	13.85	26	0.26 (0.03–0.47)	0.003
fearlessness (cm)	145	2.1	3283.7	4066.8	0.42 (0.20–0.59)	<0.001
escape latency (s)	105	2.0	18.43	17.98	0.59 (0.39–0.75)	<0.001
docility (s)	102	2.0	8.78	7.65	0.53 (0.30–0.70)	<0.001

Boldness was measured as the latency to touch a novel object that was introduced into the home enclosure after the partner animal had been gently removed. The animal's interactions with the novel object were video-recorded for 1 h.

To measure exploration, the individual was introduced to a standard housing enclosure at 17.00 and left there overnight to habituate. At 9.00 the following day, a trap-door was opened, allowing the individual to explore an unknown area for 3 h. The number of trips into this unknown area was recorded by light sensors.

Fearlessness was measured as the distance moved (cm) when individuals were introduced into an open field for 20 min. The first 10 min, a shelter was present in the arena under which animals could hide. For the second 10 min, this shelter was lifted out of the arena.

To measure escape latency, the individual was put on the observer's hand approximately 5 cm above the floor of the home enclosure. The latency to leave the hand was scored within 60 s; animals that did not escape scored ‘60’.

To record docility, an individual was turned on its back and held in the hand of an observer. For 30 s, the time an individual actively struggled to escape that situation was recorded. Fearlessness, escape latency and docility measure behavioural expressions of stress-coping. For a detailed description of behavioural tests, see [12,13].

(b). Statistical methods

All analyses were performed using R v. 3.1.2 (The R Foundation for Statistical Computing, Vienna, Austria, http://www.r-project.org) with the package MCMCglmm for Bayesian mixed models [14].

Repeatability was calculated as variance(ID)/total variance (ID + residual) using univariate Bayesian mixed models assuming a Gaussian distribution fitting non-informative inverse-Wishart priors [14]. Each model contained animal ID as random effect. The models for litter size and maternal effort additionally contained maternal age and litter number. All models were run with 525 000 iterations, a burn-in of 25 000, and a thinning interval of 500. This yielded Monte Carlo Markov chains with a sample size of 1000 and low autocorrelation.

To test for correlations among traits, one Bayesian multi-response mixed model containing all life-history traits as well as the behavioural traits as response variables was fitted using the same model parameters described above. A correlation was accepted as significant if the credible interval did not overlap zero. Model convergence was confirmed by applying Gelman and Rubin's test [15] using three separate chains with overdispersed starting values.

3. Results

All traits showed considerable among-individual variation with litter size being the only non-repeatable trait (table 1).

Early developmental traits (growth, maturation) correlate but were not associated with reproduction. Developmental traits were correlated with behaviour (figure 1; electronic supplementary material, table A1) while reproductive traits were not correlated. Behaviours form two behavioural syndromes, a proactive–reactive syndrome (docility, escape latency, fearlessness) and a separate slow–fast (exploration, boldness) syndrome (figure 1).

Figure 1.

Pearson correlations between life-history and personality traits. Large, dark circles represent strong correlations. Red circles are negative, blue circles positive correlations. Correlation coefficients are given below the diagonal, * indicates CI's non-overlapping zero. (Online version in colour.)

4. Discussion

Developmental life-history traits are positively correlated but do not trade-off with reproduction, at least under artificial housing conditions. Fast-growing individuals matured early and expressed more risk-prone behaviour (boldness and exploration), consistent with theoretical expectations [4,5]. However, fearlessness and escape latency showed a negative relationship with growth indicating that fast-growing individuals are passive stress-copers (i.e. showing freezing behaviour in the open field, the escape test and the struggle test). This can be explained by the behaviours' relationship to metabolism as they correlate negatively with resting metabolic rate (RMR) [12,13]. RMR is expected to be higher in fast-growing individuals because a high RMR can support the energy necessary for fast growth [16]. According to [16], behaviours requiring energy without increasing food intake, such as fearlessness and low escape latency, should be coupled to low RMR and lead to slow growth.

The repeatability of maternal effort indicates that females follow stable strategies either investing more heavily into reproduction or self-maintenance, while no trade-off between litter size and maternal effort was observed. Females favouring self-maintenance might benefit from increased survival [17], which was however not measured in this captive population. Litter size in these opportunistic breeders is positively correlated with female mass but negatively with offspring birth mass. Litter size and the ratio of offspring mass to litter size is adjusted to the prevailing environment, e.g. to season [18] or temperature [19], rather than showing consistent individual-specific variation. None of the behaviours correlated with litter size or maternal effort, questioning the broad importance of life-history traits related to offspring quality versus quantity. Although measured in a captive population, individuals still experience environmental fluctuations (e.g. season, temperature) which are thought to maintain variation in the quality–quantity continuum. Potentially, traits that are more directly related to reproduction such as parenting behaviour or affiliative behaviours might correlate with these life-history traits as found for other rodent species [20].

Although I did not measure survival and therefore cannot draw conclusions about the growth-mortality trade-off, my results support the life-history productivity hypothesis [3] stating that only behaviours increasing productivity should be associated with the growth-mortality and/or the reproduction versus self-maintenance trade-off. On the other hand, the results do only partly support the pace-of-life syndrome (POLS) hypothesis [4] assuming fast-living animals to be bold and explorative and at the same time cope with stress actively. None of the behaviours correlated with all pace-of-life indicating life-history traits (i.e. growth, maturation) although these were highly correlated with each other. My results suggest that a general association between life-history and personality traits as predicted by the POLS is too simplistic and ignores biological and ecological complexity such as environmental influences, sex differences, neuroendocrine processes and their interaction with the traits involved in POLS [21].

Ethics

All experimental procedures were in accordance with German animal protection laws and approved by the local government authority responsible for health, veterinary and food monitoring (21.07.2014).

Data accessibility

The data are available from the Dryad Digital Repository at http://dx.doi.org/10.5061/dryad.tn476hs [22].

Competing interests

I declare I have no competing interests.

Funding

A.G. was supported by a Leopoldina Postdoc Scholarship (LPDS 2015-04).