Body size and first-winter survival of Red-crowned Crane Grus japonensis chicks

ABSTRACT

An analysis of the biometrics of 187 individually marked Red-crowned Crane Grus japonensis chicks from Japan revealed that larger chicks were more likely to seen alive in their first-winter. Male chicks were larger than females but sex did not influence apparent survival, nor did brood size determine body size.

Although many ringing studies of cranes (family Gruidae) have clarified a basic pattern of life history in North America and Europe (Nesbitt et al. 2002, Alonso et al. 2008, Mewes et al. 2010, King et al. 2013), such research has rarely been conducted in East Asia. According to the current International Union for Conservation of Nature Red List of Threatened Species, 11 out of 15 Gruidae species are currently at risk of global extinction (BirdLife International 2016) and 5 of these 11 species inhabit East Asia; which highlights the importance of the regional gaps in knowledge in relation to monitoring populations and identifying extinction threats in East Asia. It is particularly important to investigate determinants of the survival of chicks by using long-term data, as mortality of chicks is higher than at any other life stage in large birds (Sutherland et al. 2004).

The Red-crowned Crane Grus japonensis is an endangered species in East Asia (BirdLife International 2016) and ecological research is necessary to formulate appropriate conservation and management for a sustainable population (Masatomi & Masatomi 2018). Previous studies reported that survival rate was lowest at the chick stage compared to other stages of their life (Masatomi et al. 2007). Here, we aimed to use data collected over 10 years to investigate how body size of Red-crowned Crane chicks might influence apparent survival later in life. First, we examined whether the 10 biometric variables measured in this study were correlated with each other. Second, we asked whether male chicks had larger body sizes than female chicks, as is seen in the adults (Inoue et al. 2013). In addition, we investigated whether the number of chicks in a family negatively affected body size as a result of limited resources such as parental provisioning effort. Finally, we asked whether chick body size was a useful predictor of apparent survival over the first winter, as determined by ring re-sightings.

This study was conducted on a population of resident Red-crowned Cranes in eastern Hokkaido, Japan (Tokachi area 42°25′ to 43°18′N 143°19′ to 143°41′E; Kushiro area 42°56′ to 43°24′N 143°58′ to 145°12′E; Nemuro area 43°00′ to 43°′38N, 144°20′ to 145°37′E). In this population, breeding pairs occupied territories and incubated up to two eggs in early spring (peak period from March to April; Masatomi & Kitagawa 1974). Chicks walk and leave their nest immediately after hatching and then stay with their parents until late winter (usually March).

The Red-crowned Crane Conservancy (RCC) individually marked crane chicks with numbered metal rings and colour rings. Approximately 25 chicks were ringed by the RCC every year between late June and late July, just before the chicks began to fly (about 100 days after hatching). Before ringing, the RCC surveyed the distribution and reproductive performances of each pair of cranes and checked the growth stage of the chicks at candidate sites. After capture, a range of biometrics (see below) and the body mass of each chick were measured. A blood sample was also collected from a blood vessel in the wing or leg so that the length of the CHD genes on the Z and W chromosomes could be analysed to determine the sex of chicks since this cannot be done based on appearance (Miura et al. 2013). The RCC attempted to minimize the length of time that chicks were kept captive in order to avoid putting them under undue stress. When the chicks were released, the RCC confirmed that they re-joined their parents.

The following 10 morphometric variables were recorded: head size (culmen to feathering, gape, nostril, total head length), leg size (tarsus length, ankle to tip of claw), wing length (wing chord), total length (tip of bill to tip of tail feathers), length from bill to the tip of the claw and the body mass. They also recorded the number of chicks (1 or 2) per family.

Whether the ringed individuals were observed in their first winter (i.e. re-sighting rate) was recorded. Every year, the RCC conducted a census of the whole area for 10 days with the help of more than 50 volunteers and follow-up checks at feeding sites from the end of January to early February. In order to prevent missing observations of survived individuals, we also checked for the presence of ringed individuals in their second winter when the ringed individuals were not observed in the first winter. Therefore, while ring re-sighting data can never exactly measure true survival rates, they can provide estimates of apparent survival of the ringed individuals and in this study we are confident that we recorded the majority of birds that were alive.

In total, we used data on 187 individuals in 168 families collected between 2005 and 2014. We analysed these data using linear mixed models (LMMs) or generalized linear mixed models (GLMMs). To avoid pseudo-replication, the year that ringing was conducted was set as a random variable in the GLMMs. We selected the model with significant (P > 0.05, two-tailed) terms. In all models, we used the Julian date when ringing was conducted as an additional dependent variable to control for the effect of the growth of the chicks. We checked for possible collinearity, which can lead to unreliable estimates in GLMMs (Freckleton 2011), by calculating variance inflation factors (VIFs; O’Brien 2007). The highest VIF was 1.32, indicating that there were no strong correlations among independent terms in our models.

First, we summarized all 10 biometric measurements using principal component analysis (PCA; Jolliffe 2011) in order to investigate whether the 10 variables were correlated with each other and to extract principal components for the body size of chicks.

Second, we analysed factors affecting the body size of chicks. Since all body size variables were highly correlated (see the result of PCA), we used the first principal component (PC1) value as an index of the body size of chicks. We ran LMMs in which body size (PC1 value) was set as a response variable. We included the following categorical variables as explanatory variables: chick sex (male or female) and the number of chicks in a family (1 or 2).

Finally, we conducted GLMMs featuring binomial error structure, in which the ringed chicks’ re-sighting during winter was set as a response variable. We analysed body size (continuous, PC1 value), sex (male or female) and number of chicks (1 or 2) as explanatory variables.

Regarding the correlation of physical measurements, PCA generated a first principle component (PC1 value) with an eigenvalue (index of the data’s covariance) higher than 1.0 (Table 1). This indicated that PC1 explained 89.6% of the variation in the original variables with the same positive loading from these variables (Table 1, eigenvector: mean ± se = 0.316 ± 0.002). High PC1 values indicated large body sizes. Eigenvalues of other principal components were less than 1.0 (PC2 = 0.366).

Table 1. Principal component analysis and mean dimensions for the biometrics of Red-crowned Crane chicks.

	PC1	Mean size ± sd
Eigenvalue	8.967
Proportion of variance	0.896
Rotation
Culmen to feathering (mm)	0.323	95.2 ± 10.9
Gape (mm)	0.322	105.4 ± 11.4
Nostril (mm)	0.302	67.7 ± 8.6
Total head length	0.317	164.5 ± 15.7
Tarsus length (mm)	0.317	274.2 ± 32.5
Ankle to tip of claw (mm)	0.309	393.8 ± 38.6
Total length (mm)	0.319	852.0 ± 121.0
Natural wing length (mm)	0.312	267.0 ± 102.0
Body mass (g)	0.320	3923.3 ± 916.8
Bill to tip of claw (mm)	0.318	1285.0 ± 140.0

Male chicks had larger body sizes (PC1 value) than female chicks after controlling for the effects of date (Table 2(a)). The number of chicks in a family did not affect body size (PC1; Table 2(a)). The body size of the chicks (PC1) was positively related with the winter re-sighting rate (Table 2(b)). No other variables affected the re-sighting rate (Table 2(b)).

Table 2. Summary of mixed effects models of (a) determinants of body size (linear mixed effects model, with year as a random effect) and (b) determinants of ring re-sighting probability (binomial generalized linear mixed model, with year as a random effect) for Red-crowned Cranes ringed as chicks.

(a) Body size (PC1)
Independent terms	b ± se	t	P
Intercept	−2.895 ± 0.444	−6.516	< 0.001
Julian date	0.113 ± 0.014	7.671	< 0.001
Sex (male > female)	0.927 ± 0.372	2.493	0.013
Number of chicks	0.358 ± 0.416	0.861	0.390
(b) Re-sighted in first winter
Independent terms	b ± se	z	P
Intercept	1.804 ± 0.244	7.376	< 0.001
Body size (PC1)	0.163 ± 0.071	2.278	0.022
Julian date	−0.011 ± 0.019	−0.574	0.565
Sex (male = female)^a	0.048 ± 0.431	0.112	0.910
Brood size (1 or 2)	−0.261 ± 0.460	−0.568	0.570

^aEstimate of b ± se when the effect of the right side was set to 0.

To our knowledge, this study is the first to report variables affecting the size of Red-crowned Crane chicks in the wild. Our analyses show that chick body size positively affected the probability of re-sighting over the first winter, after controlling for the effect of confounding factors such as the Julian date of ringing (Table 2(b)). This result suggested that chick body size was positively related to the survival rate since the re-sighting rate was likely to be representative of the true survival rate.

The number of chicks in a family (1 or 2) affected neither size nor re-sighting rate (Table 2). This might be because the Red-crowned Crane is a precocial species whose young are mobile immediately after hatching and can find food by themselves at a relatively early stage. Alternatively, it might be that the total amount of parental effort is greater in families with two chicks than in families with one chick. As a result, parental investment may have been similar for each chick irrespective of the number of chicks in the family. To test this idea, we need observational data on parental care.

We also found that males had larger body sizes (PC1 values) than females, even at the chick stage (Table 2(b)). This sexual difference is also seen in adults (Inoue et al. 2013).

A previous study reported that mortality rate is highest in Red-crowned Cranes during the chick stage, before the first winter (Masatomi et al. 2007). Therefore, the size of crane chicks is an important factor to consider when evaluating future population dynamics and when planning conservation actions. Future studies are needed to clarify environmental factors explaining chick body size. The habitat of the crane is highly variable, from natural marshes to agricultural areas such as dairy farms. Such variation might influence the kinds of foods available, which may in turn influence chick growth. It is essential to collect these basic data to plan and conduct the conservation of these rare birds.

Ackowledgments

We thank members of the Red-crowned Crane Conservancy, especially, K. Momose for collecting ringing data and giving advice, F. Mastumoto and M. Inoue for managing the observation data of ringed individuals, H. Teraoka for conducting sex determination and volunteers for helping the field research. We also thank the ministry of the environment of Japan and Yamashina institute for Ornithology for permitting the research.

ORCID

Kohei F. Takeda http://orcid.org/0000-0002-6550-6681

Nobuyuki Kutsukake http://orcid.org/0000-0001-8637-6233