We monitored reproduction of 11 female Amur tigers (Panthera tigris altaica) on and near the Sikhote-Alin Biosphere Zapovednik, Russia, 1992–2000, using radiotelemetry, capture, and conventional tracking (using snow and soil substrates). Tigers gave birth in all but 3 months of the year, with a peak in late summer (x2 = 10.68, d.f. = 3, P = 0.014; n = 19 litters from 11 mothers). Minimum age of 1st reproduction for 4 tigers was 4 ± 0.4 years (mean ± 95% confidence interval). Mean interval between litters was 21.4 ± 4.4 months (n = 7 pairs of consecutive litters for 4 tigers). Mean litter size was 2.4 ± 0.6 cubs (n = 16 litters of 9 tigers) when litter size was 1st determined but, due to 41–47% cub mortality (n = 19 litters), decreased to 1.3 ± 0.5 cubs (range = 0–4, n = 19 litters) by the time cubs were 12 months old. At least 57% of cub mortality was anthropogenic. Mean age at dispersal was 18.8 ± 1.5 months (n = 5 litters). Mean reproductive rate was 1.4 cubs/year, but only 0.7 cubs/year survived up to 12 months old. We believe that recent conclusions that tiger populations can grow and recover rapidly from substantial losses may be overly optimistic.
We studied reproductive activities of tigers on and near the 390,184-ha Sikhote-Alin Biosphere Zapovednik (hereafter, Zapovednik) in Primorye Province, Russia (44°46′N, 135°48′E). Russian Zapovedniks (state reserves) are highly protected lands with minimal human disturbance; access is restricted to scientists and forest guards (Stepenitski 1996). The eastern border of the reserve is the coast of the Sea of Japan, and its central feature is the Sikhote-Alin Mountain Range, which parallels the coastline. Elevations range from 0 m to nearly 1,600 m, but most mountain peaks are below 1,200 m. Secondary Mongolian oak (Quercus mongolica) forests are common near the coast and a mixture of Korean pine (Pinus koraensis), larch (Larix komarovii), birches (Betula costata, B. lanata, and others), Amur basswood (Tilia amurensis), and Khingan fir (Abies nephrolepis) persists more inland and at higher elevations. The land surrounding the Zapovednik includes a 70,350-ha buffer zone (1-8 km wide) where human activities include fishing, hunting, tourism, and some agricultural practices such as livestock grazing and hay cutting. More descriptions of the region and environmental variables potentially influencing tiger reproduction can be found elsewhere (Knystau-tas and Flint 1987; Miquelle and Smirnov 1999; Miquelle et al. 1996, 1999a, 1999b; Newell and Wilson 1996).
We monitored reproductive activities of 11 adult female tigers from January 1992 through December 2000; 9 were radiocollared, and 2 were monitored by tracking 1 litter each of radiocollared cubs (Table 1). We captured and fitted tigers with radiocollars using methods described in Goodrich et al. (2001). Each animal was assigned a letter designating sex (F = female and M = male) and a unique number. We monitored and recorded locations of radiocollared tigers from the ground (on foot and from vehicles) and from the air in a biplane (Antonov-2) or helicopter (MI-8). Ground locations were obtained by triangulation, approaching within 100–400 m and partially circling the tiger, visual observations, and subsequently locating tracks (especially in snow, but also in various soil substrates) in an area where we detected a tiger's radiosignal. In using track data (in snow or soil), we did not attempt to identify individuals based on specific track characteristics (Das and Sanyai 1995; Karanth 1987) but compared track measurements (particularly pad width) found in the field to known measurements of captured females to verify location, presence or absence of litter, and litter size of females. General estimates of track size were used to identify adult males, females, and cubs following Smirnov and Miquelle (1999) and Matyushkin et al. (1999).
↵a Female no. 007 was never radiocollared, but her litter was monitored by tracking her 2 radiocollared offspring.
↵b Calculated from time of captures to time when cubs were at least 12 months old (after which cubs were rarely with their mother).
Reproductive parameters.—We estimated the following reproductive parameters: age at 1st reproduction, seasonality of conception and birth, length of gestation, mean interval between litters (when the 1st litter in a pair of consecutive litters survived at least 2 months), mean litter size, cub mortality, sex ratio of litters, and cub age at dispersal (when cubs left their natal home ranges). We used the following age categories: cubs (associated with and dependent on their mothers), subadults (no longer associating with their mothers but not yet having reproduced), and adults (having reproduced at least once). We estimated ages by comparing tooth wear (Ashman et al. 1983), body weight, and body and track measurements (Nikolaev and Yudin 1993) and by behavioral observation, i.e., young tigers traveling with their mothers (cubs), scent marking and other territorial behaviors (associated with adults or newly established subadults—Smith et al. 1989), or females traveling with cubs (adults).
We estimated minimum age of 1st reproduction for all females monitored from 2 years of age until their death or disappearance, assuming that successful reproduction does not occur before 2 years of age (Christie and Walter 2000; Smith 1984). We estimated birth dates of litters by 1 or a combination of the following indicators: females usually localized their movements just before and up to 2 months after the birth of a litter (Smith and McDougal 1991); extrapolation of birth dates from observations and radiolocations of female-male pairings and indications of mating behavior, e.g., roaring (Smith and McDougal 1991); by tooth replacement patterns of young (canines are replaced at approximately 12 months—Mazak 1981); and by estimating cub ages based on body size (when captured or observed) and track size (as above). We include an estimate of error for each birth date based on an assessment of accuracy (day of localization for denning may have an error of 1–7 days, depending on frequency of locations, but for estimates of tooth replacement we considered error to be ±1 month). Recognizing the potential errors in our age estimates, we looked for seasonality in birth dates at a gross scale by comparing the number of litters born in each of 4 seasons, early winter (November-January), late winter (February-April), early summer (May-July), and late summer (August-October), using a chi-square test for equal proportions (SAS Institute Inc. 1996). Because of small sample sizes, we selected these seasons post hoc to maximize the probability of detecting a difference. We estimated length of gestation for 1 female whose pairing with a male and birth date were well defined. We estimated month of conception for other females from known birth dates by assuming that gestation lasts, on average, for 103 days (Kitchener 1991; review in Nowell and Jackson 1996; Sankhala 1978).
We determined litter sizes, sex ratio, cub mortality and survival by visual observation, radiolocations of cubs themselves, or, more often, by the presence or absence of cub tracks (especially in snow) in association with a radiocollared tigress. If cubs were not sexed by direct observations, we determined their sex using track measurements at about 1 year of age, when males are 24% larger than females (Smith and McDougal 1991), tracks of male cubs (pad width >10 cm) are generally larger than tracks of adult females (pad width generally 8.5-9.5 cm), and tracks of female cubs are still smaller than adult females (pad width generally <9 cm—Smirnov et al. 1999; Yudakov and Niko-laev 1987). Because the sex of most cubs was determined using track measurements at 12 months, our estimation of litter sex ratio likely does not represent sex ratio at birth.
If cubs were not radiocollared but we found their tracks with those of their mothers at or beyond 12 months of age, we assumed that they survived and were recruited into the population; if we lost their tracks before 12 months of age, we assumed they died. If action was taken to prevent the death of cubs after the death of their mother (i.e., removal from the wild or supplemental feeding in the wild), we assumed these cubs would have died without those actions and included them in our estimates of mortality. To include all litters for calculating cub mortality rates, for those litters whose size was unconfirmed in the first 12 months, we used 2 values for litter size: we assumed litter size to be the minimum (i.e., 1) or we assumed litter size was the average for all others at 1st detection. We calculated age of dispersal for radiocollared cubs that dispersed as the age at which we 1st located each cub ≥5 km from the boundary of their mothers home range.
We estimated reproductive rate of female tigers using mean litter size/mean birth interval (J. J. Craighead et al., in litt.; Wielgus and Bunnell 1994), total lifetime productivity (total number of cubs produced in a lifetime—Sunquist 1981), and total lifetime reproduction of young surviving their 1st year. This last value is slightly different from that estimated by Smith and McDougal (1991) of “lifetime reproduction of dispersal young,” but it provides a basis for comparison. In estimating total lifetime productivity, we assumed that females are reproduc-tively active until 14 years of age (Crandall 1964; Kleiman 1974; review in Nowell and Jackson 1996): the oldest known-age tigress was in Nepal and was killed when she was at least 15.5 years old (McDougal 1991). Data are reported as mean ± 1 SD.
We estimated minimum age of 1 st reproduction as 4 ± 0.4 years (mean ±95% confidence interval) based on observations of 4 female tigers: female 1 produced her 1st litter at 3.5 years of age (±1 month); female 3 at 4.5 years of age (±1 year); female 23 had still not reproduced at 4 years of age (±0.25 months) when her collar failed; and female 27 had not yet reproduced at 4 years of age (±6 months) when she was killed by poachers. We consider our estimate a minimum age of 1st reproduction because 2 of the 4 animals had not yet reproduced at 4 years of age. Our assumption that reproduction does not occur before 2 years of age was supported by 2 additional observations: female 25 had not yet reproduced at 2 years of age when she was killed by poachers, and female 26 dispersed and we lost her signal at 21 months of age, before she reproduced.
Tigers gave birth in all but 3 months of the year (n = 19 litters from 11 mothers; Fig. 1), but births were most frequent in late summer (x2 = 10.68, d.f = 3, P = 0.014). Based on these estimates of birth dates, corresponding conceptions were most frequent from March to May (Fig. 1). Because litters were detected at different ages and by using different methods, error in estimating litter birth dates varied from 2 to 30 days (Table 2).
Times of birth and conception for Amur tiger litters on and near the Sikhote-Alin Biosphere Zapovednik in Russia, 1991–2000. Conception dates were birth dates minus estimated gestation period (103 days).
↵d An unknown number of cubs were presumed dead after their mother was poached.
The gestation period of 1 female tiger (female 15) with well-defined conception and birth dates was between 101 and 108 days. This female (female 15) also gave birth 7 months after she lost her 1-week-old litter to unknown causes. For females whose litter survived at least 2 months, the mean interval until the next litter was 21.4 ± 4.4 months (range = 13–26 months, n = 7 pairs of consecutive litters born to 4 tigers; Table 2). Females had shorter birth intervals after raising single cubs (15.5 ± 4.9 months, n = 2 intervals and 2 tigers) than after raising >1 cub (23.8 ±1.1 months, n = 5 intervals and 3 tigers; t = −3.39, d.f. = 1, P = 0.09).
On average, we 1st determined litter size when cubs were 3.9 ± 1.6 months old (n = 16); nearly 70% of the litters were >2 months old when litter size was 1st detected. Mean litter size was 2.4 ± 0.6 cubs/ litter (range = 1–4 cubs, median = 2.5 cubs, mode = 1 cub, n = 16 litters of 9 females; Table 2) when litter size was 1st determined but decreased to 1.3 ± 0.5 cubs/ litter (range = 0–4 cubs) by the time litters were 12 months old. Cub mortality was 41–47% during the period from 1st detection to 12 months of age, the range depending on our methods, i.e., whether litters of unknown size were allocated initial sizes of 1 (minimum) or 2.4 (mean; n = 19 litters and ≥41 cubs). Sex ratio of litters when cub sex was 1 st detected was female biased ( 1.45 ± 0.49 females/male cub, n = 10 litters); 3 litters of single cubs were all females.
Cubs died, or were removed from the wild and therefore considered mortalities in our analysis, from several causes (Table 2). Two cubs representing 2 litters died of natural causes at <1 month of age (1 was killed by a small predator and 1 died of unknown but nonanthropogenic causes at <1 week of age), ≥6 cubs from 6 litters (including 2 complete litters of unknown sizes) died of unknown causes, and ≥9 cubs from 4 litters (53% of total cub mortality, n ≥ 17 dead cubs) died, or would have died without our intervention, because of anthropogenic factors. The last included 1 cub that was shot by a Zapovednik forest guard in self defense, 1 litter of 4 cubs that were captured and removed from the wild after poachers killed their mother, 1 litter of ≥1 cub that almost certainly died after its mother disappeared (presumably poached), and 1 litter of 3 cubs that were kept alive by supplemental feeding from their 7th through 11th month after poachers killed their mother.
Mean age at dispersal was 18.8 ± 1.5 months (n = 5). Female 1 and female 23 did not disperse but settled in their natal home ranges. We only located female 23 with her mother (female 4) once between 15 and 22 months of age, at which time female 4 was poached. The mother of female 1 was not collared, but tracks and observations indicated that they were not together after female 1 was 17–18 months old. Of 14 litters with unmarked cubs, cubs were never detected (via tracks or visual observation) in association with their radio-collared mother after 17 months of age.
Reproductive rate was 1.4 cubs female−1 year−1, but only half of these (0.7 cubs/year) survived up to 12 months of age. Assuming a female is reproductively active from 4–15 years of age, total lifetime productivity would be 15.4 cubs and total successful productivity would be 7.7 cubs. However, our data suggest that few, if any, females achieve such high lifetime productivity. Of the 14 females monitored for this study, the status of 3 is unknown (never collared, collar failure, and dispersal), 6 died at ages estimated to be <10 years, and 4 are still alive (3 of these are also <10 years). Only 1 female survived the entire study period and beyond (January 1992-December 2001). At 10 years of age, she had produced 5 litters, with a lifetime production of at least 9 cubs but with only 4 cubs surviving to 12 months. At a reproductive rate of 0.8 cubs/year, she is well below our estimated average reproductive rate.
Although litters were born in all but 3 months of the year (November, February, and March), Amur tigers on our study site were seasonal breeders, with over 50% of all births occurring in late summer (August-October) and corresponding conception rates peaking in March-May. These results contradict the majority of literature on wild Amur tiger populations that report peak conception rates in January and February (Baikov 1925; Dunishenko and Ku-likov 1999; Kucherenko 1985), suggesting a spring (April-May) birth peak. Because these reports are based on observations made via snow tracking, the potential for bias is great because they had no information during snow-free months. Our data also contradict data from captive Amur tigers, which demonstrated a distinct spring (April-June) birth peak (Seal et al. 1987). Indeed, our data contradict all other studies on all subspecies that report breeding throughout the year but with seasonal conception peaks in November-early April and, hence, birth peaks in February-July (reviews in Mazak 1981; Nowell and Jackson 1996; Table 3). Small sample size and use of multiple births from individual tigers may have resulted in skewed estimates of birth and conception peaks in our study because multiple litters of some individual tigers tended to be produced at the same time of the year (i.e., female 1 gave birth to 2 litters in October and female 4 gave birth to 2 litters in August). Nonetheless, this data set represents one of the largest for a wild population of tigers, and the pattern coincides with birthing seasons of mountain lions (Puma concolor) in north-temperate environments (review in Logan and Swean-or 2001).
Comparison among studies on reproductive parameters of tigers.
Age at 1st reproduction (years)
Mean litter size
Interbirth interval (months)
Mean dispersal age (months)
Smith and McDougal (1991), Smith (1993)
Seal et al., (1987)
Abramov K. (1977), Abramov V. (1962)
Abramov K. (1965)
Kucherenko (1972, 1985)
Smirnov and Miquelle (1999)
Amur and Bengal
Nowell and Jackson (1996, review), Mazak (1981, review)
E. N. Matyushkin et al. (in litt.)
Dunishenko and Kulikov (1999)
Our results on other reproductive parameters of wild Amur tigers also differed from those reported by many other observers (Table 3). Mean birth interval (21.4 months) was shorter than those in all previous reports for Amur tigers (Table 3) but almost identical to the 21.6 months reported by Smith and McDougal (1991) for Bengal tigers. Mean litter size at 1st detection (2.5 cubs) was at the high end of the range reported for Amur tigers (Table 3) but less than the nearly 3 cubs/litter reported by Smith and McDougal (1991). However, our estimate of litter size at 12 months of age (1.5 cubs) was at the low end of the range reported for Amur tigers (Table 3). This discrepancy with other reports for Amur tigers is probably due to different methods used to measure reproductive parameters in the field. Radiotrack-ing allowed us to estimate birth dates, birth intervals, litter sizes, and cub mortality more accurately than traditional snow tracking methodology because radiocollared tigers could be tracked continuously for long periods of time, including snow-free periods. Our results suggest that the reproductive potential of Amur tigers is higher than was previously believed, i.e., birth intervals are shorter and mean litter size greater, but due to cub mortality, estimates of recruitment of young (survival to 12 months) turned out to be similar to those of earlier reports (Abramov K. G. 1965; Smirnov 1986; E. N. Matyushkin et al., in litt.).
Although our data may be more accurate than earlier studies, our inability to determine litter size during the 1st few months of life no doubt led to an underestimate of initial litter size and an underestimate of overall mortality rates of cubs. Even in captivity, mortality rates of nearly 40% have been reported during the first 2 months of life (Christie and Walter 2000). Human-caused mortality, including poaching of mothers with cubs, accounted for 53% of all cub mortality identified during our study and is a primary factor depressing lifetime productivity of individuals and recruitment rates for the population (Kerley et al. 2002).
On average, Amur tigers in this study dispersed 4 months earlier (19 versus 23 months) than Bengal tigers in Nepal (Table 3); although in Nepal, a cub's association with it's mother declined sharply after 18 months and most cubs were completely independent of their mothers by 21–22 months. Higher tiger densities (3/100 km2 compared with 0.3/100 km2—Smirnov and Miquelle 1999; Sunquist 1981) and the insular nature of the Nepal study area may have inhibited dispersal there.
With the exception of birthing season and age at dispersal, reproductive parameters of female Amur tigers on our study site were similar to those reported for Bengal tigers in Nepal (Smith and McDougal 1991; Table 3), where study techniques were similar to ours. This was true even though Bengal tigers occur where prey densities are an order of magnitude higher and tiger home ranges an order of magnitude smaller than those in the Russian Far East (Goodrich et al. 1999; Miquelle et al. 1999a; Smith et al. 1987; Sunquist 1981).
Sunquist (1981) estimated that lifetime production of a tigress living under optimum conditions in the wild could reach 13–18 animals and estimated that about half that number would likely reach adulthood. More recent data, both from Nepal and our work, suggest that such large reproductive outputs are seldom, if ever, met. Smith and McDougal (1991) estimated mean lifetime reproduction of female tigers in Nepal at 4.5 dispersing young, and although our data do not allow precise calculations, it is clear that lifetime reproduction is even less in Russia. Thus, recent conclusions that tiger populations can grow rapidly and recover relatively rapidly from substantial losses (Sunquist et al. 1999) may be overly optimistic.
The Siberian Tiger Project is funded by The National Geographic Society, the “Save The Tiger Fund,” National Fish and Wildlife Foundation, the National Wildlife Federation, Exxon Corporation, the Charles Engelhard Foundation, Disney Wildlife Fund, Turner Foundation, and Richard King Mellon. We thank I. Nikolaev, B. Schleyer, N. Reebin, A. Reebin, A. Kostirya, I. Seerodkin, V. Melnikov, A. Saphonov, V. Schu-kin, and E. Gishko for their assistance with data collection. Director A. A. Astafiev and Assistant Director of Science M. N. Gromyko of Sikhote-Alin State Biosphere Zapovednik provided the logistical and administrative support necessary to conduct this work, and the Russian State Committee for Environmental Protection provided permits for capture, as well as much welcome political support, for our work.
1965. The Amur tiger: a relic of the fauna of the Far East. Pp. 106–112 in Zapiski Pri-morskiy filiala Geographichiskaya Obshchestva SSSR. Dalnevostochniy Knizhniy Izdatilstva, Vladivostok, Russia (in Russian).
1989. Population characteristics of tigers on the Sikhhote-Alin eastern macroslope. Pp. 151–153 in Vsesoyuznoe soveshchanie po probleme ka-dastra i ucheta zhivotnogo mira. Tezisy dokladov, Ufa, Russia (in Russian).
. 1999. Preliminary analysis of the system of home ranges of Pantera tigris altaica in Sikhote-Alin Biosphere Reserve. Pp. 89–97 in In ternational conference on rare mammal species in Russia and adjacent territories (AriktovA. A., ed.). Russian Academy of Sciences, Moscow, Russia (in Russian, English summaries).
1987. Tigers in India: a critical review of field censuses. Pp. 118–133 in Tigers of the world: the biology, biopolitics, management and conservation of an endangered species (TilsonR. L., SealU. S., eds.). Noyes Publications, Park Ridge, New Jersey.
. 1999. Distribution and numbers of Amur tigers in the Russian Far East in the mid-1990s. Pp. 242–271 in International conference on rare mammal species in Russia and adjacent territories (AriktovA. A., ed.). Russian Academy of Sciences, Moscow, Russia (in Russian, English summaries).
. 1999. People and tigers in the Russian Far East: searching for the “coexistence recipe.” Pp. 290–293 in Riding the tiger: tiger conservation in human-dominated landscapes (SeidenstickerJ., ChristieS., JacksonP., eds.). Cambridge University Press, Cambridge, United Kingdom.
. 1999a. Hierarchical spatial analysis of Amur tiger relationships to habitat and prey. Pp. 71–99 in Riding the tiger: tiger conservation in human-dominated landscapes (SeidenstickerJ., ChristieS., JacksonP., eds.). Cambridge University Press, Cambridge, United Kingdom.
. 1999b. A habitat protection plan for the Amur tiger: developing political and ecological criteria for a viable land-use plan. Pp. 273–295 in Riding the tiger: tiger conservation in human-dominated landscapes (SeidenstickerJ., ChristieS., JacksonP., eds.). Cambridge University Press, Cambridge, United Kingdom.
1994. The tiger in Lazovskiy Zapovednik. Pp. 98–102 in Prirodookhrannye territorii i akvatorrii Dalnego Vostoka i problemy sokhraneniya biologicheskogo raznoobrazia. Materially 2 nauch-noy konferentsii, posv. 60 letiyu Ussuriyskogo za-povednika. Far Eastern Branch of the Russian Academy of Sciences, Vladivostok, Russia (in Russian).
. 1940. Distribution, biology, and harvest of the Amur tiger, Felis tig ris longipilis Fitzin., in the Sikhote-Alin Mountain country. Nauchno-me-todicha Zapiski Po Zapovednikom, Zooparkam, i zoosadam 7:251–254 (in Russian).
. 1987. Behavioral indicators and endocrine correlates of estrous and anestrous in Siberian tigers. Pp. 244–254 in Tigers of the world: the biology, biopolitics, management, and conservation of an endangered species (TilsonR. L., SealU. S., eds.). Noyes Publications, Park Ridge, New Jersey.
1986. Large carnivores and the consequences of habitat insularization: ecology and conservation of tigers in Indonesia and Bangladesh. Pp. 1–42 in Cats of the world: biology, conservation, and management (MillerS. D., EverettD. D., eds.). National Wildlife Federation, Washington, D.C.
. 1999. Population dynamics of the Amur tiger in Sikhote-Alin State Biosphere Reserve. Pp. 61–70 in Riding the tiger: tiger conservation in human-dominated landscapes (SeidenstickerJ., ChristieS., JacksonP., eds.). Cambridge University Press, Cambridge, United Kingdom.
. 1999. New data on Amur tiger reproduction. Pp. 414–426 in International conference on rare mammal species in Russia and adjacent territories (ArictovA. A., ed.). Russian Academy of Sciences, Moscow, Russia (in Russian, English summaries).
. 1987. Female land tenure system in tigers. Pp. 97–109 in Tigers of the world: the biology, biopolitics, management, and conservation of an endangered species (TilsonR. L., SealU. S., eds.). Noyes Publications, Park Ridge, New Jersey.
. 1999. Ecology, behaviour, and resilience of the tiger and its conservation needs. Pp. 5–18 in Riding the tiger: tiger conservation in human-dominated landscapes (SeidenstickerJ., ChristieS., JacksonP., eds.). Cambridge University Press, Cambridge, United Kingdom.
. 1999. Where can tigers live in the future? A framework for identifying high-priority areas for conservation of tigers in the wild. Pp. 255–272 in Riding the tiger: tiger conservation in human-dominated landscapes (SeidenstickerJ., ChristieS., JacksonP., eds.). Cambridge University Press, Cambridge, United Kingdom.