Animals and husbandry

The guinea pigs used in this experiment were bred at the Department of Behavioural Biology at the University of Münster. The breeding program was established in 1975 with 40 founder multicolored shorthair guinea pigs from a professional breeder, and individuals from other breeders are routinely added to the breeding stock to prevent inbreeding. The breeding program consists of multiple breeding harems, whereby one male and two to three females are housed together in an enclosure with an area of either 1m2 or 1.5m2. The offspring of these individuals remain in the harem groups throughout weaning and are removed at 21 ± 1 days of age.

The twelve animals used in this study were transferred upon weaning to two groups of six females. Since the females were born at slightly different times, the groups were established when the oldest two females were at least 20 days old, and the younger females were added as they reached 21 days of age. The age difference between the oldest and youngest female of each group was 59 and 57 days. This age variation is random due to the continuous breeding program. No full siblings were used for this experiment; any half-siblings shared a father and were subsequently housed in separate groups.

The area of each enclosure was 1.5m2; the walls had a height of 0.5 m and were made of wood with a red plastic segment at the bottom, and the floor was covered with wood shavings (Tierwohl Super, J. Rettenmaier & Söhne GmbH + Co KG, Rosenberg, Germany). Each enclosure contained two large shelters made of red plastic, hay that was refreshed daily, and commercial guinea pig food (Höveler Meerschweinchenfutter 10,700, Höveler Spezialfutterwerke GmbH & Co. KG) and water were supplied ad libitum; water was supplemented weekly with vitamin C powder. The two groups were housed in separate rooms under controlled conditions, with a light dark cycle of 12:12 (lights on 7:00–19:00), average temperature of 22 °C, and average humidity of 42%.

Experimental procedure

Testing occurred in two four-week testing phases (Fig. 6). In the first week of each testing phase, videos were recorded in the home enclosures to be later analyzed to determine dominance rank indices of individuals. In the following three weeks, a cortisol response test was carried out for each female. Two females from each group were tested each week and only one female from each group was tested each day. The second testing phase began after a two week break so that there were six weeks between each measurement, and the sequence of individuals tested in weeks 8 through 10 was identical to the order of individuals tested in weeks 2 through 4. Mean age at the first cortisol response test was 172 days (range 151–197 days); mean age at the second cortisol response test was 213 days (range 193–238 days). Females had lived in their social groups for at least four months prior to the first observation.

Fig. 6

Experimental procedure of rank index determination and cortisol response tests. Videos analyzed for rank index were recorded in weeks 1 and 7. Weeks 2 through 4 and 8 through 10 consisted of cortisol response tests; four females were tested per week. The same testing order was maintained for the second measurement; for example, the same four females that underwent the cortisol response test in week 2 had their second cortisol response test in week 8

Rank index determination

Rank index was determined from videos recording the social behavior of the groups over the course of a week. Videos were recorded for multiple hours in the afternoon every other day (Monday, Wednesday, and Friday), and this was repeated after a six week break to assess the stability of rank index. The behavioral coding software Interact (Interact, Lab Suite Version 2017, Mangold International GmbH) was used for video analysis. Individuals were observed using focal animal sampling, and each individual was observed until it was involved in 10 interactions resulting in a retreat for each of the three days. Therefore, each female had a total of 30 interactions in which she retreated from another female or another female retreated from the focal female. A retreat was defined as the following: A female moves away from another female so that she maintains a distance of more than one body length; this behavior is shown either after an interaction of the females or after an approach of one of the females involved. Female guinea pigs do not often interact agonistically and can be quite subtle in their dominance interactions [12, 39]. Therefore, this retreat definition allowed for responses to interactions as well as responses to approaches, as long as the females had been within one body length of one another. A retreat was counted even if other females were nearby as long as the two interacting females were within one body length of one another. A rank order index was calculated for each female as the ratio of the number of times the focal female was retreated from divided by the total number of retreats (30). Therefore, the rank order index was on a scale from 0 to 1, with 0 being completely subdominant and 1 being completely dominant. This rank order index was calculated separately for the first and second measurement for each individual.

Cortisol response test

The cortisol response test is used to measure the endocrine stress response to a challenge. Cortisol is the predominant glucocorticoid in guinea pigs [52], and guinea pigs show an increase in plasma cortisol when exposed to a novel environment void of shelter [53]. It is not known how long cortisol concentrations increase in response to a challenging situation before peaking and decreasing in female guinea pigs. For this experiment, individual females were placed in a novel enclosure for 2 h, and blood samples were taken directly before and after 1 and 2 h to capture the baseline and response values of plasma cortisol.

The cortisol response test began at 13:00 ± 15 min, as plasma cortisol concentrations fluctuate throughout the day and a peak is observed at 13:00 [40]. By starting the cortisol response test at the daily cortisol peak, any increase in cortisol concentration measured would be due to the challenge and not circadian fluctuations. The dimensions of the test enclosure were 1 m × 1 m, with walls that were 50 cm high constructed of wood with a red plastic section at the bottom. Similar to housing conditions, the floor of the enclosure was covered with wood shavings and water and guinea pig food was provided. The test enclosure was in a different guinea pig housing room from where the focal individual was housed.

The housing room of the focal individual was locked 1 h prior to the beginning of the test to prevent any influence of human activity on baseline cortisol levels. At the beginning of the cortisol response test, a stopwatch was started directly when the experimenter knocked on the door to enter the housing room of the focal animal. The focal animal was collected, brought to a separate room, and placed on the lap of an assistant who then applied a small amount of muscle salve (Finalgon® salve, Boehringer Ingelheim International GmbH, Ingelheim on the Rhine, Germany) to the ears of the focal animal; excess salve was removed directly afterward. The assistant then held the ear of the animal taut so that the experimenter could illuminate the blood vessels via a cold-point lamp held beneath the ear and prick a visible blood vessel with a sterile blood lancet (Solofix® Blutlanzetten, B. Braun Melsungen AG, Melsungen, Germany). The assistant massaged the ear to stimulate blood flow while the experimenter collected approximately 150 μl of blood with two capillary tubes (Capillary tubes for microhaematocrits, 100 μl, Paul Marienfeld GmbH & Co KG, Lauda-Königshofen, Germany) within three minutes of entering the housing room of the focal animal. A swab was applied to the ear to stop the blood flow, and the focal animal was weighed (to the decigram) and placed into the test enclosure with the rump against the center of the closest wall. The room was then locked to prevent disturbance during the test, and the blood sample and weighing procedure was repeated 60 min and 120 min after the initial entering of the focal animal housing room.

The blood plasma was isolated directly after the blood sampling procedure. The capillary tubes were sealed at one end with hematocrit sealing compound (Brand GmbH & Co. KG, Wertheim, Germany). The sealed capillary tubes were then centrifuged at 13,000 rpm (16,060 g) for five minutes, after which the plasma was separated from the rest of the blood. The capillary tube was broken at this separation point with an electronic file, and the plasma was pipetted into an Eppendorf tube and centrifuged at 13,000 rpm (13,800 g) for three minutes. The plasma was pipetted into a new Eppendorf tube and centrifuged under the same conditions until no visible pellet remained. The plasma samples were then frozen at − 20 °C.

The concentration of cortisol in the blood plasma was determined using an enzyme-linked immunosorbent assay (Cortisol ELISA, RE52061, IBL International GmbH, Hamburg, Germany). The samples were analyzed in two batches consisting of six individuals each. The principle of the analysis is based on the following description (IBL International GmbH 2014):

A certain amount of enzyme-labelled antigen and the antigen in the sample compete for the binding sites of the antibody-coated wells. After a certain incubation time, the enzyme-labelled antigens that had not bound were removed by washing. The substances prednisolone (30%), 11-desoxy-cortisol (7%), corticosterone (1.4%), cortisone (4.2%), prednisone (2.5%), 17α-oh-progesterone (0.4%), desoxy-corticosterone (0.9%) and 6α-methyl-17α-oh-progesterone cross-reacted with the antibody. The intra-assay variances were on average CV = 2.98% and the inter-assay variances were on average CV = 3.51%.

Statistical analysis

Statistical analysis was carried out with R version 4.0.3 [54]. The package rptR (version 0.9.22) [55] was used to estimate adjusted repeatability for baseline cortisol, cortisol responsiveness, rank index, and body weight lost during the cortisol response test. In addition, the packages lme4 (version 1.1.25) [56] and lmerTest (version 3.1.3) [57] were used to assess the influence of the fixed effects on rank index and body weight lost during the cortisol response test. Performance (version 0.7.3) [58] was used to verify that the models fulfilled assumptions. When using rptR, permutation was set to 500 and bootstrapping was set to 1000. Two-tailed tests were used and the significance threshold was set at 0.05.

Linear mixed-effect models were fitted to estimate adjusted repeatability of baseline cortisol (log transformed), cortisol responsiveness after 1 h (log transformed), cortisol responsiveness after 2 h (log transformed), rank index, and body weight lost during the cortisol response test (absolute values log transformed). Continuous fixed effects were mean-centered and individual identity was fitted as a random effect in all models. To control for any influence of time or habituation to the testing regime or any influence of the housing group, measurement (first or second) and housing group were included as fixed effects in all models. Baseline cortisol and cortisol responsiveness after 1 and 2 h included rank index and body weight as additional fixed effects. Age and body weight were included as fixed effects for rank index. For body weight lost during the cortisol response test, cortisol responsiveness after 2 h and age were additionally included as fixed effects.

After running the repeatability analysis for the three cortisol sampling times (baseline, responsiveness after 1 h, responsiveness after 2 h), further analyses were carried out to gain a better understanding of why cortisol responsiveness after 1 and 2 h were repeatable but baseline cortisol concentration was not. To do this, the variance was partitioned into within- and between-individual components for each cortisol sampling time. These variance components were then pairwise compared between the three cortisol sampling times. An asymptotic two-tailed P value was calculated as twice the proportion of samples in which the difference (within-individual variance component from baseline cortisol minus the within-individual variance component from responsiveness after 1 h, etc.) was smaller/greater than zero.

To assess whether there was an association between cortisol concentrations and rank index, a linear mixed-effect model was fit with cortisol concentration (untransformed) as the response variable and sampling time (baseline, responsiveness 1 h, responsiveness 2 h), rank index, and the interaction between sampling time and rank index as fixed effects. Sampling time was additionally included as a random slope, and individual identity was included as a random intercept. Covariation of slopes and intercepts was not constrained. Additional fixed effects included body weight (mean centered), measurement (first or second), and housing group. To determine whether the cortisol concentrations measured at the three sampling times (baseline, responsiveness after 1 h, responsiveness after 2 h) significantly differed, the package emmeans (version 1.6.2-1) [59] was used.

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit The Creative Commons Public Domain Dedication waiver ( applies to the data made available in this article, unless otherwise stated in a credit line to the data.


This article is autogenerated using RSS feeds and has not been created or edited by OA JF.

Click here for Source link (