Introduction

Matings in Japanese macaques, Macaca fuscata, normally occur during the autumn and winter (September–February) and, following a gestation period of about 173 days, births normally occur during the spring and summer (March–August; Kawai et al. 1967, p. 40; Nozaki et al. 1992, p. 301). Birth seasonality obviously is a direct consequence of mating seasonality (cf. Takahashi 2002, p. 145). However, because field data concerning birth seasonality are much more abundant than those concerning mating seasonality, the present study of variation in reproductive seasonality in M. fuscata focuses on birth seasonality. Available evidence, including old and new information, concerning the nature and causes of birth-season variation in this species is suggestive but not conclusive.

Birth seasonality in in-situ troops in Japan

The seminal article on geographic variation of birth seasonality in M. fuscata was published by Kawai et al. (1967, p. 38). These authors studied birth-season dates in 25 troops of this species (Fig. 1; Table 1; Appendix 1); nineteen of these (including two with incomplete data) were natural troops observed in native habitats, five were natural troops that had been transported to new localities, and one was an artificial group composed of individuals captured on Yakushima and transported to Honshu. Concentrating attention on in-situ natural troops with complete data, Kawai et al. (1967, p. 44) noted a general tendency for birth seasons at northern latitudes in Japan to occur earlier than those at southern latitudes (cf. Lancaster and Lee 1965, p. 498); as exceptions, birth seasons on Shodoshima and at Funakoshiyama appeared to be earlier than expected, relative to the latitude of these localities (Kawai et al. 1967, p. 43; cf. Fig. 2; Appendix 1).

Fig. 1.
figure 1

Distribution in Japan of Macaca fuscata troops for which birth seasonality data are available. In-situ troops (see Appendix 1). Latitude 31–37°N (localities studied by Kawai et al. 1967, pp. 38–39): 2 Toimisaki; 3 Kojima; 4 Takasakiyama; 5 Nametoko; 7 Kochi; 8 Taishakukyo; 9 Gagyusan; 10 Kanbanotaki; 13 Funakoshiyama; 14 Rosando; 15 Choshikei; 17 Minoo-A; 18 Arashiyama; 23 Hagachi; 25 Takagoyama-S; 26 Takagoyama-A; 29 Jigokudani. Latitude <31°N or >37°N: 1 Nina-A; 30 Kinkazan A; 31 Wakinosawa. Supplementary localities: 11 Katsuyama; 16 Kaminada. 19 Otoumi; 20 Ryozen-A; 24 Hakone. Translocated troops (see Table 1): 6 Miyajima; 12 Takeno; 21 Ohirayama; 22 Okinoshima; 27 Takaosan; 28 Hodosan

Full size image
Table 1. Birth seasonality comparisons of in-situ and translocated troops of Macaca fuscata
Full size table
Fig. 2.
figure 2

Latitudinal variation in birth seasonality in in-situ troops of M. fuscata. Seasonality in 18 troops distributed within the Toimisaki–Kinkazan latitudinal zone (31°22′–38°17′N) is compared with seasonality in troops on Yakushima (Nina-A, 30°21′N) and Shimokita Hanto (Wakinosawa, 41°08′N); for documentation, see Appendix 1. (Regression equation, y=−10.07x+493.4; r=−0.764, F=16, P<0.01)

Full size image

The data of Kawai et al. were reexamined by Van Horn (1980, p. 192) and Fedigan and Griffin (1996, p. 377). The authors of each of these studies selected data for 11 of Kawai et al.'s troops (10 troops in common to both studies) and confirmed the correlation between latitude and birth-season dates.

Cozzolino et al. (1992, p. 332) also reexamined the data of Kawai et al., including both in-situ and translocated troops in their analysis. These authors found a positive correlation between temperature at nearby weather stations and mean birth date in 25 troops [including two translocated troops not studied by Kawai et al. (1967, p. 38)]; it should be noted that local temperature generally decreases as latitude increases (cf. Hamada et al. 1996, p. 99).

All of the above analyses of geographic variation in birth seasonality in in-situ M. fuscata troops are restricted to localities north of 31°N and south of 37°N, which excludes the southernmost and northernmost extremes of the species range (ca. 30°15′–41°30′N; Fig. 1). Data are now available for troops at three localities outside of the 31–37°N latitudinal zone (Nina-A, Yakushima, 30°21′N; Kinkazan A, 38°17′N; and Wakinosawa, Shimokita Hanto [Peninsula], 41°08′N) and for five supplementary localities within the 31–37°N latitudinal zone (Appendix 1).

Of troops at the three localities outside of the 31–37°N latitudinal zone, the mean birth date of one—Kinkazan A—fits well on the birth-date/latitude regression line determined by mean birth dates of the 17 in-situ troops with complete data that were studied by Kawai et al. (1967, pp. 38–39; Fig. 2). For these 18 troops, from Toimisaki (31°22′N) in the south to Kinkazan A (38°17′N) in the north, the negative regression between latitude and mean birth date (r=−0.764, P<0.01) slightly exceeds that for Kawai et al.'s 17 troops, excluding Kinkazan A (r=-0.737, P<0.01). At 31°N on the Toimisaki–Kinkazan regression line, the predicted mean birth date is 30 June; at 38°N, the predicted mean birth date is 20 April, more than 2 months earlier than at 31°N.

Available evidence for troops at the remaining two localities outside of the 31–37°N latitudinal zone (Nina-A, Yakushima; Wakinosawa, Shimokita Hanto) indicates that birth seasonality in these troops does not conform to the Toimisaki–Kinkazan regression (Fig. 2); the regression that aggregates Nina-A and Wakinosawa with the 18 Toimisaki–Kinkazan localities is not significant (r=−0.441, P>0.05), unlike the regression that excludes Nina-A and Wakinosawa (see above). South of 31°N, in the Nina-A troop (30°21′N), the mean birth date (30 April) apparently is approximately 2 months earlier than—not, as predicted, later than—the mean birth dates (30 June and 2 July, respectively) at Toimisaki and on Kojima (ca. 31°30′N); the mean birth date is similarly early (12 May) in the Yakushima population that was translocated to Ohirayama, Honshu (Table 1). Suzuki et al. (1998, p. 318) have previously noted that the birth season on Yakushima, at the southern limit of distribution of M. fuscata, occurs during approximately the same months as the birth season on Kinkazan, which is nearly 8° north of Yakushima and only approximately 3° south of the northern limit of distribution of the species.

North of 37°N, at Wakinosawa (41°08′N), the mean birth date (13 May) is slightly later than—not, as predicted, about 35 days earlier than—the mean birth date (26 April) on Kinkazan (38°17′N; Fig. 2). The apparently anomalous timing of the birth season on Shimokita Hanto was previously suggested by Kawai et al. (1967, p. 42). Information concerning earliest and latest birth dates (mean birth dates unknown) for troops at five supplementary localities within the 31–37°N latitudinal zone (Appendix 1) suggests that birth seasonality in these troops conforms reasonably well with the Toimisaki–Kinkazan regression (cf. Fig. 2).

Birth seasonality in translocated troops

Data concerning birth seasonality are available for ten translocated troops—six translocated within Japan, three translocated from Japan to northern hemisphere localities in the United States or Italy, and one translocated from Japan to Tasmania, in the southern hemisphere (Table 1; Appendix 2). Birth seasonality in troops translocated across up to 11° of latitude (Mihara/Beaverton) to northern hemisphere localities in Japan, North America, and Europe is generally similar to that in corresponding in-situ troops in Japan, as noted by previous authors (Kawai et al. 1967, p. 47; Cozzolino et al. 1992, p. 329; Nozaki et al. 1992, p. 308; Nozaki 1993, p. j97; Fedigan and Griffin 1996, p. 374). In contrast, translocation of a troop across about 80° of latitude, from the northern hemisphere to the southern hemisphere (Kasuga/Launceston), has resulted, within 18 months, in forward displacement of the birth season by approximately 6 months.

Long-term birth-season variation in a translocated population

Birth records of a M. fuscata population translocated from Yakushima to Honshu have been carefully maintained for 45 years by the Japan Monkey Centre (Fig. 3; Appendix 2). The progenitors of this population were 81 monkeys captured at various unrecorded localities on Yakushima (ca. 30°20′N) and released at Ohirayama (35°23′N), Honshu, in March 1957 (Kawai 1960, pp. 204, 222). The population was permitted to range freely, with minimal provisioning, at Ohirayama until August 1986, when, as a result of complaints by local residents, the monkeys were recaptured, confined to a fenced enclosure at the site, and fed pellets; to control population growth, some of the female monkeys were given contraceptive pills, and the most prolific females were removed from the group (A. Katoh, Japan Monkey Centre, personal communication). Finally, in 1997, the population was moved from the enclosure at Ohirayama to an enclosure at the nearby Japan Monkey Centre.

Fig. 3.
figure 3

Interannual variation in mean birth date in an M. fuscata population translocated from Yakushima to Ohirayama, Honshu, in 1957 and to nearby Japan Monkey Centre in 1997; this population was free ranging until August 1986 and subsequently has been confined to enclosures. For documentation, see Appendix 2

Full size image

From 1957 through 1986, when the population ranged freely at Ohirayama, the 30-year mean birth date was 12 May (Fig. 3; Appendix 2), which, as previously noted (Table 1), is close to the mean birth date of an in-situ troop on Yakushima (30 April); during these 30 years, the annual mean birth date fluctuated between 19 April and 16 June, but these fluctuations did not trend either earlier or later (regression, y=0.1529x−167.6; r=0.109; F=28; P>0.05). From 1987 through 2001, when the population was confined to enclosures at Ohirayama and the Japan Monkey Centre, the 15-year mean birth date was 3 May, the annual mean birth date fluctuated between 16 April and 22 May, and the annual mean birth date tended weakly to become slightly earlier (regression, y=−1.134x+2384; r=−0.519; F=13; 0.05>P>0.01). The birth seasonality of this population apparently was not affected by its translocation to an unconfined habitat (1957–1986), but it may have been affected by conditions of the population's subsequent confinement (1987–2001).

Laboratory studies of birth seasonality

Nozaki and colleagues have conducted an important series of laboratory studies of birth seasonality in M. fuscata (Nozaki 1991, p. 104; Nozaki et al. 1992, pp. 302, 313). These authors have concluded that the annual cycle of this species "is governed by the periodicity of internal reproductive rhythms in combination with multiple annual cyclic environmental factors, including photoperiod, ambient temperature, rainfall and social influences." In captives kept indoors under conditions of constant artificial day length and temperature, normal ovulatory cyclicity and normal birth cyclicity persisted for at least 12 years. Further studies revealed that experimental manipulation of the photoperiod experienced by captive females had little or no effect on the females' natural ovulatory cycles (Nozaki et al. 1992, p. 309). Simultaneous manipulation of photoperiod and temperature yielded equivocal results; these simultaneous manipulations succeeded in reversing the natural ovulatory cycles in two of four experimental monkeys (Nozaki et al. 1992, p. 310).

Discussion

Laboratory studies indicate that M. fuscata females manifest an intrinsic reproductive rhythm that can persist independently of external photoperiod or temperature cues. This intrinsic rhythm also is evident in free-ranging troops that have been translocated across up to 11° of latitude and nonetheless have maintained their original reproductive cycles virtually unchanged (Table 1); translocation of a troop from Japan to the southern hemisphere, however, has resulted in reversal of the original northern hemisphere cycle—presumably in response to the reversed seasons.

Available information concerning in-situ M. fuscata troops that inhabit the latitudinal zone between Toimisaki (31°22′N) and Kinkazan (38°17′N) suggests that the intrinsic reproductive rhythm of these troops has become locally coordinated with some aspect of latitudinal variation and that the mean birth date in these troops tends to become earlier as the latitude of troop localities increases (Fig. 2). However, in troops that have been sampled south of Toimisaki (i.e., Nina-A, Yakushima) and north of Kinkazan (i.e., Wakinosawa, Shimokita Hanto), the mean birth date apparently does not have the same relationship to latitude as in troops within the Toimisaki–Kinkazan latitudinal zone. Following are speculative answers to two major questions that arise concerning latitudinal variation in mean birth date in M. fuscata.

  1. (1)

    In M. fuscata troops that inhabit the Toimisaki–Kinkazan latitudinal zone, why is it apparently adaptive for the mean birth date to become earlier as the latitude of troop localities increases?

    Winter, which generally arrives earlier at higher latitudes than at lower latitudes, is the poorest food season for Japanese macaques (Suzuki 1965, p. 42; Iwamoto 1982, p. 168; Nakagawa 1997, p. 275; Agetsuma and Nakagawa 1998, p. 285). If infants have not reached a minimum level of development before the onset of winter, they presumably will be unlikely to survive the poor food conditions of that season (cf. Eaton 1978, p. 37); poor food conditions affect the survival of infants directly and indirectly, via their lactating mothers. Therefore, it is adaptive for Japanese macaques living at higher latitudes—with earlier winters—to give birth earlier than those living at lower latitudes; as a result of this adaptation, when winter begins, infants born at higher latitudes will be at the same level of development as those born at lower latitudes.

  2. (2)

    Why does the above relationship between mean birth date and latitude not apply to M. fuscata troops on Yakushima (south of Toimisaki) and on Shimokita Hanto (north of Kinkazan)?

    Yakushima and Shimokita Hanto are geographically marginal to the main distribution of M. fuscata (Fig. 1), and, as a result, monkey troops living in these two areas are affected by special ecological conditions; significantly, during the last glacial maximum, Shimokita Hanto apparently was completely uninhabitable by M. fuscata (Tsukada 1982, p. 1092; Kawamoto 1997, p. 37, 1998, p. 54, 1999, p. 303; Takahara et al. 2000, p. 673). Winters on Yakushima are relatively mild (Agetsuma and Nakagawa 1998, p. 276) and therefore are less of a factor in infant survival than elsewhere in the range of M. fuscata. Shimokita Hanto, by contrast, is so far north (ca. 41°N) that extrapolation of the Toimisaki–Kinkazan regression to this locality is biologically unfeasible; the mean birth date predicted at this latitude by the Toimisaki–Kinkazan regression is 22 March, which would be well before the end of winter on Shimokita Hanto.

    The 2-month difference in mean birth dates between the Nina-A (Yakushima) troop and the Toimisaki and Kojima troops is particularly striking and puzzling, because these localities are in close geographic proximity and therefore are fairly similar in climate (Figs. 1, 2). Additional data collected from in-situ troops at strategically dispersed localities are required to investigate further the patterns and processes of birth-season variation in M. fuscata.