Ethics statement

The project (evaluating small mammal abundances and handling with Tengmalm’s owl chicks) was approved by the Ministry of the Environment of the Czech Republic (permit No. 35016/02-OOP/8751/02, 530/758 R/08-Abt/UL, 01220/LP/2008, and 48429/ENV/14-2831/630/14). Owl chicks were handled and ringed under the Ringing Centre of the National Museum in Prague permit No. 329 and 942. Maximum effort was made to reduce handling time.

Study area

The study was conducted in the northern part of the Czech Republic (50°N, 13°E), on the Ore Mountain plateau (elevation: 730–960 m a. s. l.; area: ca 100 km²), from 2006 to 2014, except 2007 (blood of nestlings not available) and 2013 (no nestlings hatched). The study site is characterized by a mosaic of small patches of mature Norway Spruce Picea abies forests, fragments of young secondary forests, and open areas with no trees; such a pattern resulted from significant impacts of air pollution towards the end of the 20th century (for details see [40]). In this study site, Tengmalm’s owls breed mainly in nest boxes (> 90% of breeding pairs), as naturally occurring cavities for nesting are scarce [40]. The nest boxes were evenly distributed within the study area, and their numbers varied between 116 and 212 (mean ± SE; 159.3 ± 11.5) in different years.

Field procedures

Each year, all nest boxes were inspected from the onset of the breeding period (late March) to the end of the breeding period (July) in intervals of one to three weeks to detect new breeders. Nest boxes found to be occupied were subsequently revisited at a sufficient frequency to assess the number of eggs, hatchlings and hatching date (± 1 day). During the study period of 2006–2014 (except 2007 and 2013) we found a total of 116 nesting attempts from which 301 nestlings hatched. We weighed, measured and identified the sex of a total of 137 nestlings from 43 nests (Table 1). Owl nestlings were measured at age from 1 to 42 days after hatching (n = 418 measurements). Each nestling was measured and weighted on average 3.1-times (SE = 0.1, n = 137 nestlings) during its stay on the nest. The body mass and wing length of the nestlings were measured usually at intervals between one to two weeks. The nestlings were mostly identified according to their rings. When they were too small to be ringed, they were identified according to colour marks under wings or on legs. Number of measurements was during early phases of ontogeny lower compared to late phases (mean chick age during measurement was 21.4 days, SE = 0.4). This was a consequence of tactful research approach during early phase of nestling ontogeny when female often exhibit disturbed behaviour, especially during cold days. Blood samples for molecular sexing (see below) were taken from each nestling by brachial vein puncture under the wing, around 14–21 days after hatching.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Prey abundance index (expressed by number of trapped individuals per 100 trap nights) of separate prey groups, number of nests and number of nestlings measured in each study year.

Means (per one hectare trapping area or nest) ± SE are shown.

https://doi.org/10.1371/journal.pone.0138177.t001

Prey abundance

The abundance of small mammals was assessed using snap-traps (baited with flour roasted on bacon). Trapping was carried out at the beginning of June each year by setting up snap-traps in three one-hectare areas (11 × 11 trap grid; span of 10 m). The traps were left out for three days and checked every morning. We calculated the abundance index of small mammals as the number of captured individuals per 100 trap nights in each trapping area. All captured mammals (n = 186 individuals, 7 years) were identified to species level and grouped into three prey categories according to the Tengmalm’s owl diet: Microtus voles (field vole M. agrestis, common vole M. arvalis and European pine vole M. subterraneus), Apodemus mice (yellow-necked mouse A. flavicollis and wood mouse A. sylvaticus), and Sorex shrews (common shrew S. araneus and pygmy shrew S. minutus) (for details see [34,41]).

Laboratory work

Sex identification from blood samples was carried out using molecular biology techniques involving PCR with primers related to the CHD gene [42]. Genomic DNA from blood samples was extracted by alkaline lysis following a neutralization step [43]. A small amount (3–5 μl) of blood suspension in 96% ethanol was briefly spun down, alcohol was decanted and samples were dried out at room temperature. Into each sample, 50 μl of alkaline lysis buffer (25 mM NaOH, 0.2 mM disodium EDTA, pH 12) was added and heated to 96°C for 10 min. The samples were then cooled down on ice and 50 μl of neutralization buffer (40 mM Tris-HCl, pH 5) was added. As a template for PCR, 1–2 μl of mixture was used. Alternatively, the Chelex100 Resin (BioRad) DNA extraction method can also be used [44]. The 150 μl of 5% Chelex100 solution in sterile water was mixed in a microcentrifuge tube with the blood sample, briefly vortexed and incubated at 56°C for 30 min. After another vortexing followed an 8 min incubation step at 96°C, and samples were then vortexed again. The final step was centrifugation at 10 000 g for 10 min. The supernatant was directly used as a template for PCR. Each PCR reaction was performed in 20 μl. We used the primers 2550F(5'-GTTACTGATTCGTCTACGAGA-3') and 2718R (5'-ATTGAAATGATCCAGTGCTTG-3'), as described by Fridolfsson and Ellegren [42]. The PCR mastermix was prepared according to the Taq DNAPolymerase (NewEngland Biolabs) manufacturer’s guidelines. The PCR conditions were optimized to initial denaturation for 5 min at 95°C, followed by 29 cycles of denaturation at 95°C for 30 s, annealing at 60°C for 40 s, an extension at 72°C for 1 min 10 s, and a final extension at 72°C for 5 min. At the end of the PCR program, the amplified samples were loaded into wells on agarose gel (1% w/v) stained with ethidium bromide. Bands along the DNA molecular marker were detected and photographs taken under UV light. In Tengmalm’s owl, this pair of primers produces a single Z-band (700 bp) in males, and Z- (700 bp) and W- (1200 bp) bands in females [45].

Statistical analyses

We ran GLMM Gaussian models with an identity link function for testing the effect of (i) year, (ii) brood size, (iii) hatching order and (iv) sex on body mass and wing length (dependent variables) data using R software, version 3.02 [46]. The distribution of dependent variables did not differ from the Gaussian distribution (Kolmogorov-Smirnov tests, P at least 0.4). A data unit was represented by each measurement; we used nestling individual (n = 137 individuals; categorical data) and nest (n = 43; categorical data) as random factors, and age of nestling (1–42 days; continuous data) as a covariate. Hatching order (1–2, 3–4, 5–8; categorical data), sex (male/female; categorical data), brood size (1–8; continuous data) and year (n = 7; categorical data) were used as independent (tested) variables. We used factor year in GLMM analysis instead of direct prey abundances to eliminate pseudoreplications and more detailed analyses on our first prediction are described below. We used the following model formula: lmer (dependent variable ~ age of nestling + tested factors + (1| nestling individual) + (1|nest box)). Statistical significance was obtained by comparing each model with previous model using the anova command, starting with comparison of first model with factor with null model without factors. We used forward selection, factors were added to the model based on Akaike’s information criterion (AIC) using Akaike’s weights. Percentages of variability explained by the tested factors were computed as the ratio between decrease of deviance of each model and previous model. We also show the values of the Chi-squared statistic. Because body mass was not meaningful to show standing alone, we present results of above mentioned analysis as body mass increase per day (g/day). Post-hoc comparisons of changes in daily increase of body mass among years and hatching order categories were calculated in R software using glht function.

To support above mentioned models, we also calculated growth and showed parameters for each year, hatching order group and sex using logistic growth curves fitted by non-linear regression based on the formula y = A / (1 + e ^ (−K × (x − T_i)) (Starck and Ricklefs [47], where y is nestling body mass, x is nestling age, A is an asymptotic—maximal—value, K is growth rate, and T_i is the inflexion point). Iterations were carried out using non-linear estimation in Statistica v. 12. 9 [48]. Using this software, we also calculated linear regressions for particular relationships between the above-mentioned parameters of the growth curve and prey abundance (Microtus voles, Apodemus mice, and pooled prey abundance index, which included Microtus voles, Apodemus mice and Sorex shrews). We present all relevant data used in the analyses in (S1 Dataset).

Body mass

In accordance with our first hypothesis, we found that the body mass of nestlings controlled for age varied among years. Simultaneously, the yearly growth rate and inflection point increased (i.e., nestlings grew faster and for a longer period) with increasing prey abundance index. This effect was most pronounced during the peak years of Apodemus mice and Microtus voles. No relationship was found between the asymptote of nestling body growth (i.e. maximal nestling mass) and prey abundance index. Eldegard & Sonerud [10] documented that supplemental feeding led to parental allocation of food for self-maintenance rather than to higher body mass of Tengmalm’s owl nestlings. On the other hand, Hipkiss et al. [3] showed that the asymptote of Tengmalm’s owl nestling body mass increased in experimentally fed broods, but no relationship was found between the rate of mass gain and food supplementation. In this light, Kouba et al. [49] demonstrated that Tengmalm’s owl parents produced nestlings with different body mass during a two-year period with extremely different prey abundance (Apodemus mouse peak compared to a poor rodent year), but Valkama et al. [16] found that the body condition (expressed by residuals of the regression between body mass and wing length) of fledged Tengmalm’s owlets was not associated with the phase of the vole cycle. Similarly inconsistent results were documented in other owl species, e.g. supplementary food resources have been found to affect the growth rate of little owl Athene noctua, burrowing owl Athene cunicularia or Ural owl Strix uralensis nestlings [4,12,19], however, the body condition of Montagu’s harrier Circus pygargus during fledging (expressed by the difference between measured body mass and the asymptotic body mass) did not vary according to vole cycles [30]. We suggest that the body mass of the Tengmalm’s owl nestlings in a temperate areas varies under natural prey abundance, but other factors, such as clutch size adjustment by females [50,51], the feeding effort of males [52,53], or variable environmental conditions (e.g. the night length which can limit foraging time of owls, as well as the level of interspecific competition with other vole-eating predators [50,54]), may play important roles.

The body mass of nestlings can be affected by brood size. For example, Eurasian kestrel Falco tinnunculus nestlings showed increased mortality and reduced growth rate when broods were experimentally enlarged, even when parents increased their hunting effort [22]. Similarly, experimental brood enlargement in barn owl Tyto alba led to increased nestling mortality and decreased body mass of the surviving male and female nestlings [24]. Also, in a northern population of Tengmalm’s owl, nestlings in larger broods suffered from higher mortality and fledglings from larger broods tended to reach a lower mass than those in reduced or control broods [16,55]. We found no effect of brood size on nestling body mass in the natural environments of this temperate Tengmalm’s owl population, thus rejecting our second hypothesis. Tengmalm’s owl females in temperate areas adjusted their clutch size better than females in northern areas, where prey availability is less predictable, which can lead to a clutch size larger than optimal and increased nestling mortality [50]. In light of the mentioned effect of latitude, we suggest that no effect of brood size on nestling body mass was found in our study population because females adjusted their clutch and brood size depending on prey availability.

In agreement with our third hypothesis, we found that early-hatched nestlings reached a larger body mass, and showed higher asymptote and inflection point, compared to late-hatched nestlings. Similarly, in Montagu’s harrier, Arroyo et al. [30] found the poorest fledgling condition in the youngest nestlings. Also, in other owl species (e.g., little owl, burrowing owl, Ural owl), hatching order has been found to affect growth rate [4,12,19]. In our study, late-hatched nestlings suffered from poor body condition, supporting that hatching asynchrony in Tengmalm’s owl is an adaptive way to optimize brood size under varying environmental conditions (see also [16]).

In keeping with our fourth hypothesis, we found that female nestlings reached a higher body mass than male nestlings during their time in the nest, and thus a higher asymptote and inflexion point was found in females compared to males. This finding corresponds with the results of Hipkiss et al. [3], who suggested that the female’s larger size gives them a competitive advantage over their male siblings during fights over food items [56].

Wing length

Despite adult female Tengmalm’s owls having slightly longer wings than males, no male-female difference in wing length was found in nestlings. Similarly, in marsh harrier Circus aeruginosus, Krijgsveld et al. [28] found no difference between the sexes in wing length, although female fledglings were heavier than males. Our results also agree with the finding of Hipkiss et al. [3] that juvenile Tengmalm’s owl females require more time to fully develop their wings than males. Furthermore, we found no effect of year, brood size, hatching order or prey abundance index on the wing length of nestlings. These findings are in contrast to the two-year study performed in the same study area by Kouba et al. [49], in which owl parents produced nestlings with longer wings during an Apodemus peak year (2010), compared to a poor rodent year (2011). In our seven-year dataset (2006–2014, excluding 2007 and 2013), we recorded no occurrence of extremely high prey abundance, except in one year (2010). The disunity of these results indicates that wing length can differ under extremely different food conditions, but the differences are not pronounced when food availability is relatively stable. This finding is quite unexpected, because the individual duration of nestling periods is related to wing length rather than body mass, sex, prey abundance index or weather conditions [5].