- Top of page
- Material and Methods
- Data Accessibility
- Conflict of Interest
- Supporting Information
Seasonal conditions are known to influence postnatal mortality in mammals (Descamps et al. 2008) through direct effects on infant survival (Mumby et al. 2013) or indirectly through its effects on maternal body condition and nutrition quality and availability (Pettorelli et al. 2006). For this reason, in seasonal environments mothers may match their timing of conception and birth to the months or seasons with lower offspring mortality (Stearns 1992). Benefits of seasonal breeding have been clearly shown in many short- and midlived species, which concentrate all their reproductive effort into a short period of time characterized by high resource availability (Ogutu et al. 2010; Schaper et al. 2012). For example, in red deer (Cervus elaphus) all females give birth within a 2-month time period from mid-May to mid-July (Clutton-Brock et al. 1987), with temperature variation of just a few degrees during gestation having long-term effects on offspring mortality and reproduction (Albon et al. 1983; Kruuk et al. 1999).
In long-lived species, which produce few offspring with long periods of parental dependence and whose reproductive life spans cover several years or decades, the effects of gestational and early-life environment may be manifested both in the first few months after birth when mortality is highest, and in the long term due to the extended duration of offspring dependency (Caughley 1966). However, few studies have investigated the effects of seasonal variation in early-life environment on either birth timing or offspring mortality in long-lived mammals.
Even long-lived aseasonal breeders such as wild African elephants (Loxodonta africana) (Laws et al. 1975; Dublin 1983; Moss 1988), killer whales (Orcinus orca) (Robeck et al. 1993), and humans (Moore et al. 1999) can concentrate births into certain periods of the year. However, this pattern is not universal: the chimpanzee (Pan troglodytes) (Wallis 2002), for example, breeds throughout the year and exhibit no seasonality of births. Despite the ability of humans to reproduce any time of the year, in the Northern hemisphere more human births occur in the spring months (Lummaa et al. 1998). Season of birth in humans is also linked with both short- and long-term effects on mortality (Moore et al. 1998). However, the seasonal pattern of births in human populations does not always maximize offspring survival, possibly because it is influenced both by biological and cultural factors (Lummaa et al. 1998). These two components could also be relevant in natural animal populations interacting with humans, such as working elephants whose seasonal workloads and rest schedules are set by humans but nonetheless reproduce and feed with little human intervention. Elephants have 16-week reproductive cycles throughout the year, each with 2–3 receptive days (Fowler and Mikota 2006). There is variation in cycle duration (range 12–19 weeks), and more variation between than within individuals (Glaeser et al. 2012).
In elephant species, there is some evidence that wild African elephants from Amboseli (Kenya) experiencing poorer early-life environments showed higher mortality before age 2 years (Moss et al. 2011), and that individuals experiencing two drought years during the first years of life had reduced total life expectancy (Moss et al. 2011). Male mortality was more responsive to poor environmental conditions, due to higher energetic demands of male calves (Lee and Moss 1986). In Kenya, 81% of 1030 recorded births occurred between November and May, indicating clear seasonal variation in births (Moss 2001), but the study did not investigate whether seasonal variation in mortality was associated with seasonal changes in birth rates. Female African elephants that reproduce before the start of the rainy season give birth at the time of maximum resource availability (Wittemyer et al. 2007a,b), but there is conflicting evidence as to whether their offspring are more likely to survive their first year (Dublin 1983). Females reproducing before the rains were also more likely to be dominant, and therefore it is not clear whether differential offspring mortality is attributable to environmental conditions, maternal dominance, or their interaction. In summary, previous studies of elephants have not yet fully clarified the relationship between birth timing, offspring mortality and seasonal variation in environmental conditions, and in addition they have primarily focused on wild African elephants. African and Asian elephants are known to differ in reproductive endocrinology with a higher rate of noncycling females (Freeman et al. 2004) and synchronicity of cycling (Weissenboeck et al. 2009) in the former species.
Asian elephants (Elephas maximus; Fig. 1) are characterized by long gestations of around 22 months (Lueders et al. 2012) and an extended infant dependency period of up to 6 years (Fowler and Mikota 2006). They inhabit regions characterized by a seasonal climate, which has been shown to influence mortality rates (Mumby et al. 2013). In addition, this species includes large populations of working elephants that also face the possible effects of seasonal work schedules. They are for this reason a relevant system for studies of effects of seasonal variation on birth timing and short- and long-term mortality of calves, with results that could be applied to management practices. Due to paucity of longitudinal data on Asian elephants, it is not clear whether there is a seasonal pattern of births in wild or captive Asian elephants, or whether females time conceptions or births to the most favorable periods of the year in terms of offspring survival. Two studies of elephants in India found signs of birth seasonality and a peak of births in January: logging elephants in the south (Sukumar et al. 1997), and elephants in a wildlife sanctuary in northwest India (Baskaran et al. 2009). However, the first study miscalculated conception dates (by subtracting 26 instead of 22 months from birth dates), whereas the second includes only 58 individuals. Therefore, further investigation into the links between seasonal conditions, birth seasonality, and calf survival is required.
Figure 1. Asian elephant (Elephas maximus) female and her calf. Pan Cho Lwin, born on 3 April 1977, is shown with her fourth calf; two older female offspring are still alive, while a male died at age 8 months. Both mother and offspring are being longitudinally weighed and measured as part of the Myanmar Timber Elephant Project. Photograph by Virpi Lummaa.
Download figure to PowerPoint
Here, we investigate seasonal timing of births, stillbirths, and mortality and their relation to birth month and maternal work schedule in a population of semicaptive Asian elephants. We use a unique demographic data set of 2350 working timber elephants from Myanmar born between 1948 and 2000. We investigate (1) the distribution of births through the year, and the effects of climate seasonality (temperature and rainfall) on birth timing. We also analyze (2) the short-term effect of birth month on mortality to age 1 month, during which 10% of calf deaths occur (Mar et al. 2012), (3) the effect of birth month on mortality to age 1 years, a high mortality phase during which calves are completely dependent upon their mothers (Mar et al. 2012), (4) the effect of birth month on mortality to age 5 years, the age at which elephants are weaned and begin the training, and (5) the long-term effect of birth month on mortality to age 17 years, the age at which elephants become full-time working elephants. In all analyses, we test whether effects of birth month on mortality are modulated by sex and birth order. Finally (5) we discuss how such patterns coincide with the known seasonal workload of the mothers.
- Top of page
- Material and Methods
- Data Accessibility
- Conflict of Interest
- Supporting Information
Both the presence and the pattern of birth seasonality vary widely across long-lived species. We investigated whether Asian elephants exhibit seasonal variation in births and whether birth seasonality is associated with variation in infant and subadult mortality rates in a population of working elephants. We found strong evidence for seasonal variation in births, with birth number peaking from December to March and showing their lowest levels from May to August. These 4-month periods, respectively, included 41% and 24.8% of the 2350 birth events recorded. Reproducing during this peak birth period led to better calf survival between ages 1 and 5 years. These findings help to design improved management strategies of the working elephant population, and have general relevance for conservation of Asian elephants given that because of its low birth rate and high calf mortality risk, the semicaptive working population is supplemented by capturing endangered wild elephants.
We observed a higher birth seasonality among firstborn individuals. We included birth order in the analyses because mortality of firstborn calves was shown to be higher than of later-born counterparts in Myanmar elephants (Mar et al. 2012) as well as other species such as mountain lions (Puma concolor) (Jansen and Jenks 2012), and also because it is a proxy for maternal dominance: later offspring are born to older mothers, and both age and parity are associated with female dominance. Seasonal factors affecting conceptions seem to have a more pronounced effect in nulliparous females, or alternatively, firstborn individuals may be more vulnerable to seasonal factors. It is also possible that those mothers with only a single offspring are of lower “quality” in general, and thus there may be more such mothers among the primiparae; in this data set, 365 females of a total of 1103 included mothers contributed only a single firstborn offspring into the data set. In contrast, sex has no effect on the association between month and probability of birth. This differs from previous reports showing that African elephant males show higher infant mortality under extreme climatic conditions (Moss et al. 2011); however, the lack of sex effect is consistent with our own studies on survival by climatic conditions across all ages in Asian elephants (Mumby et al. 2013).
Given such clear seasonal variation in birth rate, we investigated whether the seasonal timing of birth could be maximizing calf survival prospects in the short or long term, and if mothers are timing their births to the most favorable periods for calf mortality risk. Our results showed that birth month is associated with seasonal variation in mortality of calves aged between 1 and 5 years, with odds ratios for survival being 44% higher for calves born between December and March compared to those born between May and August. The lack of association between birth month and mortality in individuals in other age categories could mean that mortality at those ages is determined by other factors including developmental factors for younger calves (Levitis 2011) or current seasonal conditions for postindependence juveniles (Mumby et al. 2013). Also, before the period between 1 and 5 years, conditions faced by calves may be buffered by maternal care, whereas between the ages of 1 and 5 years the occurrence of weaning exposes offspring to the effects of unfavorable conditions. At the age of 5 years, an increase in mortality was observed in this population (Mar et al. 2012) most likely due to taming. The lack of association in the age 5- to 17-year category could represent selective disappearance of weaker individuals before the age of 5 years (Nussey et al. 2011). Our analyses do not preclude delayed fitness effects of birth seasonality manifested through survival and/or reproduction beyond the age of 17 years (Lee et al. 2013).
We suggest that the factors determining the seasonal pattern of conceptions and number of pregnancies carried to term are either the schedules of work and rest periods determined by humans or seasonal climatic variation, through their effects on stress levels, maternal condition, reproductive cycling, access to mates, adaptation to maximize offspring survival, or a combination of those processes. It is difficult to tease apart these effects in the population as all elephants have the same working pattern and experience very similar climatic variation; in fact, the work undertaken by elephants is determined by the specific climatic conditions at that time of year (Zaw 1997). Asian elephants are the only endangered mammal used as a draught animal. At the beginning of the rainy season (June–August), elephants are primarily used for “aunging” (log pushing). In the late part of the rainy season (September–October), logs often form jams in rivers, these are cleared by elephants through “yelaiking,” an operation that may cause injuries including leg fractures and joint dislocation and is arguably the most stressful work-related activity for timber elephants (Mar 2007). From November to January, elephants drag logs left behind along the river banks. All elephants finish their work season by mid-February, and are moved to rest camps for the hottest and driest part of the year until work resumes around mid-June. Taking a pregnancy duration of 22 months, the resting period corresponds exactly to the 4 months with the highest number of conceptions we identified (December–March), whereas the lowest number of conceptions (May–August) correspond to the period of heavy workload. Elephants conceived in the rest season also experience higher survival, potentially reflecting that the elephants have benefited from the resting period, or that the rest period defined by humans and based on the environment matches the optimum conception or delivery time for elephants. The former explanation appears to be more likely based on our finding that early mortality is not seasonal and previous results indicating that the hot, dry months that characterize the rest period are not ideal conditions for elephants; highest elephant survival is found instead at intermediate temperatures of around 26°C and high rainfall (Mumby et al. 2013).
There are several possible reasons behind higher conception rates during the rest period. In principle, seasonal variation in conception rate can be passive (determined by variable access to mates) or active (with changes in likelihood of females coming into estrus). On one hand, females may have more opportunities to interact with potential reproductive mates (which in the case of Myanmar elephants include both wild and captive males) due to increased time spent in the forest. On the other hand, workload may actively affect reproductive cycling of females, which was shown to be disrupted under poor or stressful conditions (Schneider 2004). Studies of the reproductive cycle of African (Wittemyer et al. 2007a,b) and Asian elephants (Thitaram et al. 2008) indicate that part of the typical 16-week duration of the estrus cycle is extended in poor climatic conditions. In Asian elephants living in a comparable climate, such extension is primarily driven by a prolonging of the luteinizing stage of the estrus cycle, leading to delay of ovulation, by around 10 days on average in the wet season (Thitaram et al. 2008). Postponing conception may avoid the high energetic cost of reproduction until conditions improve (Canale et al. 2012). In this population of timber elephants, we propose that this could imply a delay of conceptions from work season into the rest season, although a 10-day delay of ovulation alone is not sufficient to do so.
Another explanation of the observed birth pattern is that it reflects seasonal variation in the number or proportion of pregnancies carried to full term. Even if conception rate is constant through the year, the same maternal and environmental factors described above may be causing an increase in rates of abortion during the work season. However, early-term abortions (known to frequently occur in Asian elephants in zoos; Lueders et al. 2010) are not recorded in the logbooks and therefore this possibility could not be directly tested. An indirect test for differences in the probability of pregnancies being carried to full term would involve an analysis of seasonal variation in number of stillbirths. A total of 101 stillbirths were included in this analysis and appear to show some variation across months (Table S4), but unfortunately the sample is too small for robust investigation.
Our findings have relevance for conservation of Asian elephants because the working population of elephants in Myanmar represents a large population of an endangered large mammal that, despite having considerably better survival rates than Asian elephants in Western zoos (Clubb et al. 2008), is currently not viable because of consistently too low birth rates and high calf mortality rates. On a wider scale, captive and wild populations of elephants are declining worldwide (Sukumar 2006) and zoo populations may not be sustainable (Faust et al. 2006; Clubb et al. 2008). The semicaptive working population is boosted by capturing wild elephants, a strategy that is costly to the timber industry because they are more difficult to train (Lair 1997) and contributes to the decline in the wild elephant population. Supplementing captive populations with the capture of wild elephants has been legislated against in many range countries including Myanmar, which limits capture of wild elephants to work in the timber industry to a maximum of 50 per year (Mar 2007). However, elephants involved in human–wildlife conflict can also be taken into captivity and even the capture of elephants at a low level could lead to extinction of the wild population in as little as 100 years (Leimgruber et al. 2008).
Therefore, breeding of captive elephants has become increasingly important (Kurt et al. 2008), but current programs are challenged by high mortality rates of calves, particularly in Western zoos (Clubb et al. 2008) and by low birth rates among the logging elephant population (Mar 2007). If the high conception rates reached in the resting season were reached throughout the year, based on birth numbers averages from the 1990s (the most recent decade in the data set), the annual calving rate could rise from 6.9% to 8.8%, meaning increase in the number of calves born from 70 to 88 per year, an increase of 25.7%. The strong implication for management of the semicaptive Asian elephant population is that placing some cycling females on prolonged rest from work may increase fertility rates and bolster the population of Myanmar logging elephants, at the same time protecting the wild population. The cost of having these females on prolonged rest would be countered by the increase in fertility of those females and number of working elephants produced by them.