This study illustrated body temperature (Tb) fluctuation during reproduction and discussed the effectiveness of measuring Tb for predicting the reproductive profile of female Apodemus speciosus both under laboratory and field conditions. Tb fluctuation was monitored for four breeding events in the laboratory and for three in the field. Individual variation in Tb fluctuation during reproduction was larger in the field than in the laboratory, while its temporal pattern was clearer in the field than in the laboratory. Tb fluctuating patterns according to the progress of the reproductive stage were similar between the laboratory and the field. Daily mean Tb increased after the start of pregnancy, decreased during late pregnancy, rapidly increased after parturition, and remained higher through the lactation period. In particular, the following three characteristics should be apparent indices of parturition: increase of daily mean Tb and daily minimum Tb in early pregnancy, decrease of daily max Tb and daily mean Tb in late pregnancy, and increase of daily max Tb at the day of parturition. These results indicate that implanting small-sized loggers enables us to obtain a representative pattern of Tb fluctuation and to predict the reproductive profile of female A. speciosus, both under laboratory and field conditions.
Reproductive events of small terrestrial mammals are not well-known, because most of them are nocturnal and live in nest burrows. Thus, their reproductive events also happen in the dark. Actually, it is not easy to address reproductive events of such small mammals even in captive conditions. For example, direct observation or palpation of mother and/or pups of the captive large Japanese wood mouse (Apodemus speciosus) during reproduction often results in infanticide (killing of a live pup) or abandonment of pups (personal observations by K. Tsuchiya and S. Sakamoto). Therefore, an indirect monitoring method is desirable.
Various and drastic changes occur in the body of female mammals through the maternal adaptation from gestation to lactation. Liver glycogen increases at the latter gestation period and decreases rapidly after parturition (Shelley 1961). The concentration of prolactin fluctuates, increasing between days 1 and 3 of the gestation period and decreasing between days 8 and 12 in the golden hamster (Mesocricetus auratus; Bast and Greenwald 1974). It is widely known that progesterone secretion is elevated during pregnancy, and it causes body temperature (Tb) elevation (e.g., Buxton and Atkinson 1948; McCormick and Greenwald 1974; Nakayama et al. 1975). Then, Tb characteristically changes according to the progress of reproduction. Measurement of Tb is also less invasive and permits continuous or repeated measurements from individual animals, thus its preferable over the former techniques. Therefore, Tb is widely used to illustrate female reproductive profiles in captive mammals, livestock such as beef cattle (Miwa et al. 2019), replacement gilts (Johnson and Shade 2017), and ewes (de Freitas et al. 2018), as well as in laboratory animals, such as dwarf hamsters (Phodopus roborovskii; Scribner and Wynne-Edwards 1994), Mongolian gerbils (Meriones unguiculatus; Weinandy and Gattermann 1995), and laboratory mice (Gamo et al. 2013).
Monitoring female reproductive profiles by Tb measurement could also be adapted to diverse wild mammals. However, such studies are still not common, especially in small mammals. This is because to measure core body temperature (Tcb) some devices need to be attached inside the animal's body. This is necessary even when predicting Tb from surface body temperature (Tsb) by infrared thermography, to ensure a strong correlation between Tcb and Tsb in target species. Therefore, costs and weights of devices, ecological features and body size of animal models, and study situations limit effective Tb measurement.
Measurement of the Tb of small wild mammals was used exclusively in the study of torpor. Recently, procedures for miniaturizing temperature data loggers (Thermochron iButtons) were developed to measure Tcb from smaller mammals, both in the laboratory and the field (Lovegrove 2009; Boratyński et al. 2018; Virens and Cree 2018). Tb of small female mammals during reproduction was reported in “torpor mammals”, such as yellow-footed antechinus (Antechinus flavipes) in an outdoor aviary and indoors (body mass, 22.8 g; Stawski and Rojas 2016), mulgaras [Dasycercus blythi, 57.4 ± 3.9 (mean ± SD) g; Körtner et al. 2008], edible dormice (Glis glis, 78–238 g; Bieber et al. 2014), greater hedgehog tenrecs (Setifer setosus, > 150 g; Levesque et al. 2014), and arctic ground squirrels (Urocitellus parryii, 600–1000 g depending on season; Williams et al. 2011). These inexpensive techniques enable us to explore the reproductive events of smaller terrestrial mammals.
We focused on wild mice whose body size and reproductive characteristics, e.g., duration of pregnancy and nursing, are equivalent to those of laboratory mice, because data for laboratory mice could become the best reference. Apodemus speciosus is nocturnal and endemic to Japan (Nakata et al. 2009). Its body mass (ranging approximately from 25 to 55 g for adults) and gestation period (19 to 21 days, Oh and Mōri 1998) are similar to laboratory mice. In addition, genus Apodemus comprise a dominant group of murid rodents in the Palaearctic region (Corbet 1978) and is the Old-World equivalent of the genus Peromyscus in the North American continent. Widely distributed species of both groups are used as environmental indicators to monitor the effects of environmental radiation from nuclear power plants (Okano et al. 2016; Kawagoshi et al. 2017) and dioxins (Ishiniwa et al. 2010) on wildlife. In such cases, reproductive parameters are monitored as good indices. In addition, these species demonstrate geographical variation in breeding season (Murakami 1974; Bronson 1985). Among these, A. speciosus living in Miyazaki prefecture breeds from autumn to spring (Sakai et al. 2013), providing unique opportunities for studying reproductive plasticity in a wild rodent.
Studies suggest that Tb during torpor was modified by ambient temperature (Ta) (e.g., Ruf et al. 1993; Ortmann et al. 1997; Boratyński et al. 2018). In contrast, it is an unclarified question whether Tb of winter-breeding A. speciosus females demonstrates a typical pattern during reproduction or not because it can be largely modified by cold Ta. A previous study reported that late pregnancy Tb in rats decreased irrespective of Ta (Eliason and Fewell 1997), suggesting that Tb during reproduction should be strongly controlled by reproductive hormone even under the harsh winter condition. Therefore, we hypothesized that Tb of winter-breeding females would demonstrate a typical pattern during reproduction. Tb measurement in female A. speciosus was already established through studies on daily torpor (Eto et al. 2014, 2015, 2018) by implantation of smaller-sized Tb measuring devices in their abdominal cavity. In addition, we achieved successful mating and parenting of these wild mice in laboratory conditions (Sakai et al. 2013; Sakamoto et al. 2015). Therefore, to test the hypothesis, first we experimentally verified that females with implanted devices were able to get pregnant through natural mating and give birth under laboratory condition. We then illustrated the characteristics of Tb fluctuation during reproduction under constant Ta condition in the laboratory, as well as in free-living females under fluctuating Ta conditions in the field. Finally, we compared Tb fluctuation during reproduction between the laboratory and the field.
Materials and methods
The Animal Experimentation Committee at the University of Miyazaki (Permission Nos. 2012-002-5 to 7 and 2017-026) reviewed and approved all experimental procedures.
Trapping, animals, and housing condition
To monitor Tb fluctuation in female A. speciosus during reproduction, a trapping study was conducted from October 2015, at the start of the breeding season for the population, to April 2016, at Sumiyoshi Livestock Science Station, University of Miyazaki (39°59′N, 131°28′E). Live traps (HB Sherman Traps Inc., FL, USA), which were baited with oatmeal and pieces of raw sweet potato, and wrapped in polythene to stabilize internal Ta and exclude rain, were set in the late afternoon and checked at the night or early in the next morning according to weather conditions.
In the laboratory experiment, five pregnant females were introduced to the laboratory [22.9°C ± 1.2°C (mean ± SD), and a natural photoperiod] and housed individually in opaque plastic cages (345 × 403 × 177 mm). Their pups were housed individually in transparent cages (225 × 338 × 140 mm) after weaning. Among these pups, 6–12-month aged females [n = 4, 32.0 ± 2.3 g (mean ± SD)] and males (n = 8, 43.2 ± 4.6 g) were used for paring. Females were housed individually in transparent cages until they recovered from surgery for data logger implantation (see the latter section), then they were transferred to opaque plastic cages with a male and pairing commenced. In the field experiment, eight non-breeding females were captured on October 2015 and kept for surgery and recovery for two weeks in transparent cages at the same laboratory. They were released in the trapped points after recovery. Wood chips for bedding were provided in all cages and commercial rodent diets (Labo MR Stock, Nosan Corporation, Kanagawa, Japan) and water were provided to the individuals ad libitum throughout the study.
Measurement of body temperature of females
Tcb of the females was measured as a Tb index using Thermochron data loggers (iButtons, DS1922L, Maxim Integrated, CA, USA) set to record the nearest 0.1°C every 60 minutes. Each data logger was made smaller according to a part of the procedures described by Lovegrove (2009) and coated with a thin layer of a paraffin-Evaflex (EV220, Du Pont-Mitsui Polychemical Co., Ltd., Tokyo, Japan) mixture. The weights of the coated data loggers were 2.0 ± 0.1 g. The maximum value of the loggers represented 6.6% of the body weight of a female in non-pregnancy, which was below 10% of the body weight, the maximum recommended by Rojas et al. (2010). Each data logger was calibrated in a water bath of known temperature before being implanted into the abdominal cavity of experimental animals. Implantation surgery took place in animals anesthetized with simultaneous administration of triple combination anesthetic 0.1 ml/10 g body weight; medetomidine (10 mg/10 ml, Domitor, Meiji Seika Pharma Co, Ltd, Tokyo, Japan), midazolam (10 mg/2 ml, Dormicum, Astellas Pharma Inc, Tokyo, Japan), and butorphanol (5 mg/1 ml, Vetorphale, Meiji Seika Pharma Co, Ltd, Tokyo, Japan). At the end of the experiments, all implanted animals were anesthetized again to remove the data loggers. After that, all animals were kept through life. Data loggers were recalibrated after the experiments.
The laboratory experiments
Four females (FA, FB, FC, and FD) were implanted from October 2016 to April 2017 (term 1) and three females (FA, FB, and FC) from May to September 2017 (term 2). After the 2-week recovery time from surgery, a female and a male were paired and Tb measurement was conducted. Female reproductive status was diagnosed from observations of the vagina (open or closed), nipples, abdomen, and changes in body weight once a week. A parturition day was detected by pup emergence. Duration of pregnancy was estimated backward from parturition day (day 0). Based on the pre-observation, breeding stages were determined as follows, pre-pregnancy (days –30 to –21), early pregnancy (days –20 to –11), late pregnancy (days –10 to –1), early lactation (days 1–10), late lactation (days 11–20), declining phase (days 21–30), and post-reproduction (days 31–40). If female pregnancy was confirmed, the paired male was removed from the cage three to six days before parturition. If female pregnancy was not confirmed through gain in body mass for three to four weeks, the male was replaced. A female whose pregnancy was not confirmed during the study period was treated as a control.
The field experiments
Eight non-breeding adult females were implanted with a data logger and individually marked with a passive integrated transponder tag (PIT; Trovan ID-100, 2.12 mm × 11.5 mm, 0.1 g). After the 2-week recovery time, implanted females were released at their captured points. A mark and recapture study was conducted from October 2015 to April 2016 to monitor breeding condition with Tb logging in the field. A trapping session comprised two or three consecutive nights. Sessions were conducted every two weeks. At each capture, individual identity, trap location, sex, body weight, and reproductive status were recorded. Female reproductive status was diagnosed from observations of the vagina (open or closed), nipples, abdomen, and changes in body weight. Apodemus speciosus were then released at the capture location. All implanted females were collected and introduced to the laboratory three to four months later. Females whose breeding were verified during December 2015 were treated as winter-breeding females. We could not obtain Tb data for non-breeding females during December 2015; hence a female (F92) whose pregnancy was not confirmed from January to April 2016 was used as a control in the field.
Data analysis
This study aimed to illustrate the characteristics of Tb fluctuation during reproduction in the laboratory, as well as in the field. However, we could not know which Tb parameters would be largely modified by fluctuating Ta in the field. Therefore, at the first step, daily Tb parameters, daily mean Tb, daily maximum Tb (daily max Tb), and daily minimum Tb (daily min Tb), were calculated on a daily basis for all individuals and compared these Tb parameters during reproduction between the laboratory and the field.
Daily Tb rhythm and daily activity rhythm are synchronous (Shimatani et al. 2021); hence we should take intraday fluctuation in Tb into consideration when analyzing Tb changes according to breeding progress. Mean hourly Tb of each stage was calculated from the middle seven in every ten-day intervals, to eliminate the values of marginal days between two consecutive breeding status, according to a previous study (Gamo et al. 2013). Gamo et al. (2013) directly monitored daily Tb and daily activity patterns during reproduction of laboratory mice in the constant laboratory condition, at the same time. Therefore, they directly analyzed Tb in active and inactive time. In contrast, we could not directly obtain the activity data in this study. Apodemus speciosus demonstrated the activity patterns of nocturnal animals. Therefore, we determined that A. speciosus would be inactive at daytime [6:00 to 18:00 JST (Japan Standard Time: UTC + 9)], with the rest being active at nighttime (18:00 to 6:00). This daily rhythm may change according to seasonal changes in day length, whereas our data unfortunately did not have adequate variation to analyze these seasonal effects on Tb. Therefore, it should be better to exclude Tb patterns around sunset and sunrise times when analyzing daytime and nighttime Tb patterns, separately. Hence, Tb changes corresponding to the breeding stages were compared for the middle 8 h in inactive daytime (8:00 to 16:00) and those in active nighttime (20:00 to 4:00), respectively.
All statistical analyses were performed using R version 3.6.1 (R Core Team 2019). To analyze whether Tb changes were explained by the breeding stage, Tb for the middle 8 h was treated as the response variable and the breeding stage, a set of consecutive three breeding stages, as the explanatory variable. The distribution of Tb for the middle 8 h showed Gaussian distribution; hence linear mixed effects models (LMMs) were used with ‘lmer’ function in the statistical package ‘lme4’ (Bates et al. 2011). Models for a set of consecutive three breeding stages were made at daytime and at nighttime, respectively.
Results
In the laboratory experiments, FA, FB, and FC, gave birth twice (FA-1st and FA-2nd), once (FB-1st), and once (FC-1st), respectively in term 1, while FA, FB, and FC, gave birth once (FA-3rd), once (FB-2nd), and once (FC-2nd), respectively in term 2, whereas FD did not; hence, in total, three in four mated females gave birth seven times. Tb data were not obtained from FA-1st, FA-2nd, and FC-1st because the battery of the implanted logger came out of position when coating of the loggers was removed. As a result, Tb data was obtained for four in seven breeding events, which resulted in three (FB-1st) or four pups (FA-3rd, FB-2nd, and FC-2nd). Tb data from a non-breeding individual (FD) was used as a control data in the laboratory. In the field experiments, Tb data during reproduction were obtained from three females, F87, F66, and F54, whose reproduction could be addressed by the mark and recapture study. F87, F66, and F54 breed in winter on December 16th, 17th, and 21th, 2015, respectively. F92 did not demonstrate breeding from January to April 2016; hence Tb data of F92 for this period was used as a control data in the field. Tb fluctuating patterns in breeding females differed from the usual patterns in non-breeding females, both in the laboratory and field (Fig. 1), while it was not different between a female with three pups (FB-1st) and females with four pups (FA-3rd, FB-2nd, and FC-2nd).
During early pregnancy (from days –20 to –11), daily max Tb did not significantly change, while daily mean Tb and daily min Tb increased (Fig. 2A and B). During late pregnancy (from days –10 to –1), daily max Tb decreased quickly, daily mean Tb decreased gradually, and daily min Tb remained high (Fig. 2A and B). On the day of parturition (day 0), daily max Tb increased by 1.0°C ± 0.6°C from day –1, with daily mean Tb and daily min Tb also increasing (Fig. 2A and B). During early (from days 1 to 10) and late (from days 11–20) lactation periods, daily max Tb, daily mean Tb, and daily min Tb remained high (Fig. 2A and B). In contrast, daily mean amplitude of Tb was slightly unclear, especially in the field (Fig. 2C and D). It decreased through pregnancy (from days –20 to –1) and reached a minimum value on day –1 in the laboratory, whereas this happened on day –7 in the field (Fig. 2C and D). It significantly increased on parturition day (2.6°C ± 0.9°C on day 0), and was unchanged through early and late lactation (from days 1 to 20) in the laboratory, whereas it gradually increased in the field (Fig. 2C and D). It increased again in the declining phase (from days 21 to 30), when it returned to values similar to the period before pregnancy (Fig. 2C and D).
Fig. 1.
Patterns of body temperature (Tb) fluctuation for breeding females in the laboratory (A; FA-3rd reproduction) and in the field (B; F87) were clearly different from the usual pattern of non-breeding females in the laboratory (C; FD) and in the field (D; F92), respectively. The individual ID was presented in the upper right of each panel.
Tb fluctuations changed according to the progress of the reproductive stage, both in the laboratory and the field (Fig. 3). Intra-day Tb fluctuations were smaller as breeding progressed and larger as breeding regressed. From pre-pregnancy to late pregnancy (Fig. 3A and B), mean Tb was higher in early and late pregnancy than in the pre-pregnancy period at daytime, inactive time [in the laboratory: pre-pregnancy vs. early pregnancy = 0.490 ± 0.043 (estimate ± SE), t = 11.471, P < 0.001, and pre-pregnancy vs. late pregnancy = 0.375 ± 0.043, t = 8.775, P < 0.001; in the field: pre-pregnancy vs. early pregnancy = 0.228 ± 0.043, t = 5.312, P < 0.001, and pre-pregnancy vs. late pregnancy = 0.136 ± 0.043, t = 3.176, P < 0.005]. In contrast, at active nighttime, mean Tb was higher in early pregnancy than in the pre-pregnancy period (in the laboratory: early pregnancy vs. pre-pregnancy = 0.233 ± 0.072, t = 3.247, P < 0.005; in the field: early pregnancy vs. pre-pregnancy = 0.199 ± 0.063, t = 3.169, P < 0.005), and mean Tb was lower in late pregnancy than in the pre-pregnancy periods (in the laboratory: late pregnancy vs. pre-pregnancy = –0.238 ± 0.072, t = –3.312, P < 0.001; in the field: late pregnancy vs. pre-pregnancy = –0.223 ± 0.063, t = –3.558, P < 0.001). From late pregnancy to late lactation (Fig. 3C and D), mean Tb was higher in the early and late lactation periods than in late pregnancy at day and night (in the laboratory: late pregnancy vs. early lactation at daytime = 0.181 ± 0.034, t = 5.327, P < 0.001, late pregnancy vs. late lactation at daytime = 0.323 ± 0.034, t = 9.507, P < 0.001, late pregnancy vs early lactation at nighttime = 0.210 ± 0.058, t = 3.636, P < 0.001, and late pregnancy vs. late lactation at nighttime = 0.336 ± 0.058, t = 5.814, P < 0.001; in the field: late pregnancy vs. early lactation at daytime = 0.356 ± 0.035, t = 10.08, P < 0.001, late pregnancy vs. late lactation at daytime = 0.428 ± 0.035, t = 12.14, P < 0.001, late pregnancy vs. early lactation at nighttime = 0.570 ± 0.055, t = 10.37, P < 0.001, and late pregnancy vs. late lactation at nighttime = 0.802 ± 0.055, t = 14.61, P < 0.001). The highest was mean Tb in late lactation, followed by early lactation and late pregnancy (Fig. 3C and D). From late lactation to post reproduction (Fig. 3E and F), mean Tb was highest in late lactation, followed by declining phase and post reproduction at daytime (in the laboratory: late lactation vs. declining phase at daytime = –0.399 ± 0.039, t = –10.220, P < 0.001, and late lactation vs. post reproduction at daytime = –0.824 ± 0.043, t = –18.980, P < 0.001; in the field, late lactation vs. declining phase = –0.304 ± 0.050, t = –6.042, P < 0.001, late lactation vs. post reproduction = –0.582 ± 0.050, t = –11.564, P < 0.001). In contrast, Tb differences between periods were not observed in the laboratory but happened in the field at nighttime (in the laboratory: late lactation vs. declining phase = 0.006 ± 0.069, t = 0.086, P = 0.932, and late lactation vs. post reproduction = –0.098 ± 0.077, t = –1.269, P = 0.205; in the field, late lactation vs. declining phase = –0.230 ± 0.082, t = –2.806, P < 0.01, and late lactation vs. post reproduction = –0.764 ± 0.082, t = –9.328, P < 0.001). Hourly Tb of post-reproduction was closely matched by that of pre-pregnancy both in the laboratory (G) and the field (H), but especially in the field.
Fig. 2.
Daily body temperature (Tb) fluctuations during reproduction in the laboratory (A) and in the field (B) and mean amplitude of Tb in the laboratory (C) and in the field (D). Four and three breeding events were analyzed for the laboratory and for the field, respectively. Square: daily max Tb, circle: daily mean Tb, and triangle: daily min Tb in A and B. Error bars indicate standard deviation (SD). Breeding stages were shown as follows, a: pre-pregnancy (days –30 to –21), b: early pregnancy (days –20 to –11), c: late pregnancy (days –10 to –1), d: early lactation (days 1–10), e: late lactation (days 11–20), f: declining phase (days 21–30) and g: post-reproduction (days 31–40).
Fig. 3.
Comparisons of fluctuations in hourly body temperature (Tb) (mean ± SE) among consecutive reproductive status: from pre-pregnancy to late pregnancy in the laboratory (A) and in the field (B), from late pregnancy to late lactation in the laboratory (C) and in the field (D), from late lactation to post-reproduction in the laboratory (E) and in the field (F), and from pre-pregnancy to post-reproduction in the laboratory (G) and in the field (H). Four and three breeding events were analyzed for the laboratory and for the field, respectively. Gray bars represent inactive daytime [6:00 to 18:00 JST (Japan Standard Time: UTC + 9)], with the rest being active nighttime (18:00 to 6:00), respectively. Allows indicate the middle 8 h in inactive daytime (8:00 to 16:00) and those in active nighttime (20:00 to 4:00).
Discussion
This study verified experimentally that female A. speciosus with implanted loggers were able to get pregnant through natural mating and give birth both in the laboratory and the field. The mean litter size in the laboratory was 3.80 ± 0.40 (mean ± SD). This was equivalent to those of previous studies: three to four (Nakata et al. 2009) or 4.29 ± 0.82 from laboratory mating (Oh and Mōri 1998). Furthermore, all weaned pups matured as usual. Therefore, we could monitor Tb fluctuations for a typical situation of female reproduction in the laboratory. Tb fluctuations in breeding females were quite different from those of non-breeding females, both in the laboratory and the field (Fig. 1). Individual variation in Tb fluctuations during reproduction was larger in females in the field than in the laboratory (Fig. 2), while its temporal pattern was clearer in the field than in the laboratory (Fig. 3). Hourly Tb of post-reproduction was very similar to that of pre-pregnancy both in the laboratory (Fig. 3G) and the field (Fig. 3H). Thus, implanting small-sized data loggers enabled us to get a representative pattern of Tb fluctuation during reproduction of female A. speciosus, similarly to laboratory mice (Gamo et al. 2013) and other larger wild rodents (Williams et al. 2011), both under constant Ta conditions in the laboratory and variable Ta conditions in the field.
Daily max Tb, daily mean Tb, and daily min Tb were useful to address the reproductive status of females (Fig. 2). Daily min Tb and daily mean Tb increased from days –21 and –20 in the laboratory and in the field, respectively. Oh and Mōri (1998) mated female and male A. speciosus only at one night in the laboratory, and determined that their gestation period was 19 to 21 days (Oh and Mōri 1998). Female laboratory mice showed a peak of activity on day –20, which was the first day of mating, because the male continuously harassed the female and the female spent a lot of time running away from the male (Gamo et al. 2013). Subsequently, female Tb increased during the mating period from day –20 compared to the baseline period, though activity eventually decreased to baseline level (Gamo et al. 2013). In the present study, Tb fluctuation during breeding of females was quite similar to that reported in Gamo et al. (2013) and mean copulation date was suggested to be around day –21.
From day –11, daily max Tb and daily mean Tb decreased rapidly and gradually, respectively (Fig. 2A and B). Decrease of Tb in late pregnancy has been observed in several rodents, such as rats (Fewell 1995), Mongolian jerbils (Weinandy and Gattermann 1995), and arctic ground squirrels (Williams et al. 2011). At parturition day, mother Tb (especially max Tb) abruptly increased in A. speciosus. Subsequently, Tb indices were maintained high throughout the lactation period. This pattern is widely observed in mammals, such as mice (Gamo et al. 2013), rats (Fewell 1995), arctic ground squirrels (Williams et al. 2011), dwarf hamsters (Scribner and Wynne-Edwards 1994), dairy cattle (Wrenn et al. 1958), and beef cattle (Miwa et al. 2019). Therefore, this pattern of Tb fluctuation around parturition should be strong evidence indicating parturition. It may be the general pattern among mammals which ovulate spontaneously, although in the dwarf hamster Tb does not decrease but only slightly increases during gestation (Scribner and Wynne-Edwards 1994). In this study, it was difficult to identify the end of lactation from daily Tb parameters, as intra-day variation in mean hourly Tb became smaller as breeding progressed, especially from pre-pregnancy to late pregnancy periods (Fig. 3A–D), but larger as breeding regressed, from late lactation to post reproduction periods (Fig. 3E and F). This was because the inactive Tb at daytime was higher, but the active Tb at nighttime was lower from late pregnancy to late lactation periods than from late lactation to post-reproduction (pre-pregnancy) period. Hence, daily amplitude was lower during breeding than during non-breeding (Fig. 2C and D).
Decline in Tb is partially explained by the lower levels of physical activity at this time in mice (Gamo et al. 2013). Eliason and Fewell (1997) hypothesized that Tb reduction in late pregnancy might be an energy conservation strategy. However, reduction in the Tb of mice in late pregnancy only reversed the Tb increase in early pregnancy, hence inactive Tb was not different between late pregnancy and baseline (Gamo et al. 2013). Thus, lowered Tb (reflecting lowered resting metabolism and/or reduced energy intake) does not contribute to energy savings in pregnancy, whereas reduction in physical activity, particularly from day –11 onwards, may be an important part of balancing the energy budget for pregnant females that need to allocate a lot of energy for fetal growth (Gamo et al. 2013). In the present study, daily max Tb and daily mean Tb, which are under the effect of Tb in active time, decreased in late pregnancy, particularly from day –11 (Fig. 2A and B). Therefore, reduction in Tb subsequent to reduction in physical activity might also occur in A. speciosus.
Through pregnancy periods, very high inactive Tb was recorded. This is consistent with the result in mice (Gamo et al. 2013). In mammals, progesterone secretion is elevated during pregnancy and it causes Tb elevation (e.g., Buxton and Atkinson 1948; Wrenn et al. 1958; Nakayama et al. 1975). Among rodents, a marked increase from mid-pregnancy until just before parturition was reported in mice (McCormick and Greenwald 1974; Murr et al. 1974; Virgo and Bellward 1974) and rats (Bridges 1984). In addition, in the IV CS strain of mice, a sharp rise in the progesterone level was observed from day 0 to day 2 of pregnancy, followed by a plateau, and then a marked increase from day 13 until the day before parturition (Kosaka et al. 1988). On the other hand, the peak of progesterone level in pregnant Mongolian gerbils emerged in early pregnancy from day –14 to day –8 (Kai et al. 1997). This is consistent with the observed elevation in inactive Tb during pregnancy in A. speciosus (Fig. 3A and B) and in laboratory mice (Gamo et al. 2013). Decrease in Tb during late pregnancy is widely observed in mammals other than rodents. In dairy cattle, this Tb decrease is caused by a reduction in the progesterone level (Wrenn et al. 1958). Therefore, high inactive Tb through pregnancy periods may be general among mammals. However, it is unclear whether the timing, duration, and increases in Tb elevation directly link to the level and timing of progesterone secretion, which would be diverse among species. Comparative approaches controlling progesterone levels will help to reveal such interspecific variations in Tb elevation and subsequent reduction during pregnancy.
Most huddling endotherms maintain a higher and more constant body temperature than their isolated counterparts (Gilbert et al. 2010). Hence, huddling with a male was also a possible explanation for the high inactive Tb through pregnancy periods. However, a male was present in a cage from pre-pregnancy (< day –30) to late pregnancy (day –3 to day –6) only in the laboratory. Thus, huddling with male cannot explain the high inactive Tb starting at early pregnancy at least in the laboratory. On the other hand, declining of the Tb of breeding females in late pregnancy started before male removal. Therefore, presence or absence of a male would not be the cause for the difference in Tb patterns between breeding and non-breeding females in the laboratory. Apodemus speciosus exhibits a promiscuous mating system, and females are territorial, while males occupy larger ranges overlapping those of females and other males (Oka 1992). In previous study, released males rarely entered the burrow entrances of pregnant females (Sakamoto et al. 2015). Therefore, huddling with male may not explain the high inactive Tb starting at early pregnancy in the field, too. In contrast, the presence of neonatal pups was a possible explanation for the high inactive Tb through lactation periods. We did not wean pups before day 40 in this study; hence declining of the min Tb of breeding females in late lactation started before removal of pups. Therefore, the decrease in body temperature might be caused by the decrease in lactating or huddling frequency. To reveal this, an experiment to control lactating or huddling frequency is needed in a future study.
This Tb monitoring method works well both in the laboratory and in the field. In future, we will detect the end of lactation from Tb by experimentally manipulating pup presence, and monitor reproductive schedules of individual females in the field. This may reveal the understudied reproductive ecology of nocturnal small mammals, which is probably diverse and “quietly” occurring in the dark. This approach will also be useful to study reproduction of endangered and invasive species.
Acknowledgments:
We appreciate the support from staff at the Sumiyoshi Livestock Science Station, University of Miyazaki. This work was supported by JSPS KAKENHI Grant Numbers JP 24657018 and JP 17K07537. We wish to thank the anonymous reviewers for critical comments on earlier version of this paper.
