WHAT DETERMINES PHENOLOGY AND SYNCHRONY OF UNGULATE BREEDING IN SERENGETI?
SIMON A. R. MDUMA [1,3]
PETER ARCESE [2,4]
Abstract. Birth seasons of ungulates in tropical regions show a complex pattern varying from asynchronous to highly synchronous and at different times of year. We examine the factors determining the phenology and synchrony of birth seasons of 13 species of ungulates in the Serengeti ecosystem, Tanzania. We propose that phenology of births (time of year) is determined by food supply, whereas birth synchrony (degree of coordination or spread) is an antipredator adaptation that functions in two different ways. High synchrony may occur through "predator satiation" in species with precocious newborn ("followers"), whereas asynchrony may occur through "predator avoidance" in species with nonprecocial young ("hiders"). We used green biomass of grass or tree shoots and percentage crude protein as measures of food supply. Births were determined from monthly sample counts covering the period 1967-1997. The frequency distribution of births was compared to that predicted by the abundance of green biomass and percentage protein, and by an even (asynchronous) monthly distribution. Wildebeest, topi, warthog, and Grant's gazelle births differed from all predicted distributions. Another group showed birth distributions similar to that of green biomass food (buffalo, oribi) or the distribution of percentage protein (giraffe, waterbuck, kongoni, zebra). Also giraffe, waterbuck, and Thomson's gazelle showed births spread more evenly through the year. For grazing species the lag in months between birth peak and protein peak is a positive function of metabolic body size whereas the lag with biomass is a negative function of body size. We suggest that small grazers produce their young early in the wet season ahead of the high protein peak, whereas large species produce their young in phase with high biomass and after the protein peak consistent with metabolic requirements. In terms of synchrony, large species in large herds with precocial young (wildebeest, topi, buffalo) have highly synchronized birth seasons consistent with the "pre dator satiation" hypothesis. Small species living in small groups with nonprecocial young (impala, Thomson's and Grant's gazelle, waterbuck, oribi) have births less synchronized than the food supply, as predicted by the "predator avoidance" hypothesis. In general, food supply determines the phenology of the birth season. Predation appears to shape the synchrony of births through two opposite adaptations. However, no single feature predicts all species' birth distributions. A combination of the phenology of food supply plus antipredator adaptations accounts for most but not all these distributions.
Key words: birth seasons; browser; food supply; grazer; phenology; Serengeti; synchrony; tropical ungulates.
The birth seasons of ungulates in north temperate zones are thought to be adjusted so that young are produced when food is plentiful (Bunnell 1982, Linnell and Andersen 1998). Thus, moose (Alces alces) and caribou (Rangifer tarandus) have well synchronized birth seasons in early June associated with the onset of new plant growth (Bergerud 1974, Dauphine and McClure 1974, Bronson 1989, van Ballenberg and Miguelle 1993). However, breeding becomes less synchronous as one approaches the tropics (Bronson 1989). In the widespread genus Odocoileus, for example, populations at latitudes [greater than]30[degrees] have birth periods that are highly synchronized with spring green-up. In the Florida Keys, Odocoileus deer give birth in all months of the year with a strong peak in spring. In Venezuela, breeding is asynchronous (Brokx 1972).
In temperate Africa births are also associated with spring and summer (Dasmann and Mossman 1962, Fairall 1968, Spinage 1973, Skinner et al. 1973, 1974, David 1978, Sekulic 1978; Underwood 1978, Anderson 1979, Murray 1982). The pattern is less clear-cut than that in the northern hemisphere, particularly in hotter areas such as Kruger National Park (N. Owen-Smith, personal communication). In contrast, in tropical Africa both the timing and synchrony of births differs among ungulate species. However, gross habitat, climate, and other factors confound comparisons between species in different regions. Thus, births have been recorded either at different seasons or throughout the year (Spinage 1969, 1973, Bourliere et al. 1976, Sekulic 1978, Wirtz and Kaiser 1988).
As part of a long-term study of ungulate populations in the Serengeti ecosystem of Tanzania, we have accumulated information since 1967 on the timing of births of 13 species of ungulates. The Serengeti ecosystem is a tropical savanna area close to the equator (l-3[degrees] S) and has a strong seasonal rainfall pattern (Norton-Griffiths et al. 1975, Sinclair 1979a). Rainfall is the main determinant of vegetation growth and food supply for ungulates (Sinclair 1975, 1977a, McNaughton 1979, 1983).
In this paper we analyze the phenology and synchrony of birth seasons in the context of two major limiting factors, food supply and predation. Phenology and synchrony address different questions. Phenology addresses when in the year births occur, i.e., their seasonality. Synchrony addresses the degree to which births are coordinated or spread over time irrespective of the time of year. We propose first, that the time of year when births occur is related to the timing of growth (phenology) in food supply (Rutberg 1987). Since each ungulate species differs (slightly) in its preferred food, the period in the year when births occur should reflect differences in when food becomes available. Secondly, we propose that the synchrony of births is determined both by the length of time that food is available and by antipredator behavior. If food determines synchrony (as distinct from seasonality) then the distribution of births should be similar to the seasonal distribution of food.
Alternatively, predation may produce extreme synchrony in births as an antipredator adaptation, as proposed for both birds (Kruuk 1964, Findlay and Cooke 1982) and ungulates (Estes 1966, 1976, Sinclair 1977a, b, Estes and Estes 1979, Rutberg 1987, Ims 1990). In essence, if births occur synchronously then predators become satiated and survival of the remaining newborn animals increases. This "predator-satiation" hypothesis should apply only to those species where such synchronized behavior is effective. Jarman (1974) has categorized African ungulates in terms of a suite of co-evolved ecological and behavioral traits. One such category comprises species that are relatively large, eat abundant food such as grass, form large herds, and have precocial young that follow the mother from birth and therefore are visible to predators. These are "follower" species. We predict it is this group that exhibits synchronized births as an adaptation to satiate predators and constrain predation. In Serengeti, follower species include wildebeest (Connochaetes taurinus), topi (Damaliscus korrigum), buffalo (Syncerus caffer), and zebra (Equus burchelli). Both kongoni (Alcelaphus buselaphus) and warthog (Phacochoerus aethiopicus) have precocial young but live in dispersed, small groups (see Plate 1).
Another group of adaptive traits identified by Jarman (1974) is shown by small to medium sized, territorial species that eat dispersed food, live at lower density in small groups, and hide their newborn young (see Plate 2). These are "hider" species. We predict that this group should not synchronize their births. Because they are not sufficiently abundant to satiate predators, synchrony would result in greater rather than lesser predation (Ims 1990). Instead, this group should show the opposite trait of asynchronous births as a "predator avoidance" behavior. In Serengeti, this group includes impala (Aepyceros melampus). the two gazelles (Gazella granti, G. thomsoni), waterbuck (Kobus defassa), oribi (Ourebia ourebi), and the very small dikdik (Rynchotragus (Mado qua) kirki). Giraffe (Giraffa camelopardalis) although large and visible, live at low density in small groups and leave their young in the care of a guardian while feeding. Giraffe do not fit easily into either category but are more similar to the lat ter group in their suite of adaptations. In general, local abundance distinguishes the two groups, with "followers" tending to be locally abundant whereas "hiders" tend to be locally sparse. Thomson's gazelle are anomalous, at least in the migratory Serengeti population, by being "hiders" that are locally abundant, whereas kongoni and warthog are "followers" that are locally sparse.
We test these predictions using data collected over 31 yr (1967-1997) on births for the thirteen ungulate species: unpublished data are presented for wildebeest, zebra, topi, kongoni, impala, waterbuck, warthog, and giraffe while other data are available for African buffalo (Sinclair 1977a), Grant's gazelle (Walther 1972), Thomson's gazelle (Bradley 1977), oribi (Jongejan et al. 1991; P. Arcese, personal observation), and dikdik (Hendrichs and Hendrichs 1971). Rainfall data, collected by members of the Serengeti Research Institute and its successor, the Serengeti Wildlife Research Centre, cover 34 yr (1961-1994).
For most species samples of animals were counted from a vehicle and scored according to age and sex. Size, coat color, presence of umbilical remnant, and absence of horns identified newborn animals, the combination depending on species. The sex of the newborn could not be identified. Adult females in most species were identified by size, color, absence of horns, or where present, their shape and size, and from the genital organs. In giraffe bony lumps on the forehead differ between the sexes. In zebra and warthog the sexes could not be distinguished reliably so that females and males were scored together as adults.
We collected data using vehicle transects that ran through four main areas of the Serengeti ecosystem: (i) the north (north of the Grumeti River in the northern extension), (ii) center and east (from the Research Institute northeastwards to the Grumeti River), (iii) west (from the Research Institute west along the corridor), and (iv) the Serengeti plains. All animals within a distance where age and sex could be distinguished with binoculars (about 200 m) were counted except in the case of wildebeest and zebra, where large numbers required us to count samples at systematic intervals. The timing of births in buffalo, Grant's gazelle, and Thomson's gazelle was obtained from published data on fetuses and pregnancy rates (Walther 1972, Bradley 1977, Sinclair 1977a) and those of oribi and dikdik from observations of juveniles (Jongejan et al. 1991, Hendrichs and Hendrichs 1971; P. Arcese, personal observation).
Rainfall was measured monthly from storage gauges placed in various parts of the ecosystem. Scientists, park wardens, and park rangers read the gauges. The data were stored at the Research Institute. Sometimes a gauge was not read for several months so those records were for the accumulated amount. The data set was purged of all records where the accumulated amount was for [greater than]2 mo. For two-monthly records the data were divided evenly.
For each species the age-sex records were summed by month and the mean monthly newborn/adult female ratio ([R.sub.mth]) calculated. For warthog and zebra the equivalent newborn/adult ratio was calculated. The ratios were summed over the year to give the annual newborn/adult female ratio ([R.sub.y] = [sigma][R.sub.mth]). The percentage of annual births occurring in each month (p[NB.sub.mth]) was therefore
p[NB.sub.mth] = 100[R.sub.mth]/[R.sub.y]. (1)
This procedure overcame bias from unequal samples of animals counted each month.
Rainfall gauges were grouped according to the different areas of the ecosystem. For each area the mean monthly rainfall was calculated, and from this we obtained the proportion of annual rainfall occurring in each month.
Grass biomass in the wet season was calculated from the relationship given in Sinclair (1975, 1977a) and Braun (1973) of accumulated biomass ([B.sub.wet]) as a function of accumulated rainfall ([R.sub.a]) since the start of the wet season. The relationship differed slightly between the short grass plains ([B.sub.wetp]) and the long grass plains and woodlands ([B.sub.wetw]):
[B.wetp] = 0.0496[[R.sup.1.7888].sub.a] (coefficient of variation [CV] = 0.89) (2)
[B.sub.wetw] = 3.501[[R.sup.1.1022].sub.a] (CV = 0.94). (3)
The percentage crude protein in the dry grass biomass was calculated from the relationship given in Sinclair (1977a) of percentage crude protein ([P.sub.grass]) and accumulated biomass (B). This did not differ between plains and woodlands so that
p[P.sub.grass] = 146[B.sup.-0.4269] (CV = 0.78). (4)
At the end of the wet season (June or July depending on the area), the grass persists at high biomass and remains green if rainfall continues. However, the grass senesces if no rain falls. Therefore, green biomass in July was estimated by taking the June biomass and discounting by the degree of senescence. The senescence discount was taken as the ratio of July to June rainfall (for eastern woodlands and plains May and June replace June and July):
[B.sub.july] = [B.sub.june]([R.sub.july]/[R.sub.june]). (5)
By the end of July the grass is dry and senescence is complete. Dry season rain then results in new growth from the roots, and hence a new relationship of biomass to rainfall. For the dry season months (August to start of wet season depending on area) biomass was calculated from monthly growth ([B.sub.drym]) and monthly rainfall ([R.sub.drym]) given in Sinclair (1979b) and Mduma et al. (1998):
[B.sub.drym] = 1.25 [R.sub.drym]. (6)
For herbaceous dicots and small woody shrubs we have used the same phenology as that for monocots. However, for trees a different phenology occurs. Pellew (1980) measured monthly production, biomass and protein of available shoots and leaves in Acacia trees fed upon by giraffe. We use his values for available biomass and calculate from his data the relationship between protein (p[P.sub.tree]) and biomass (B) as
p[P.sub.tree] = 95.689[B.sup.-0.2787]. (7)
To detect the seasonality of births we used the nonparametric two sample Kolmogorov-Smirnov test. With this we examined differences between observed birth distributions and (i) expected even distributions through the year ([K-S.sub.even]) and (ii) expected distributions determined by biomass ([K-S.sub.biomass]) and percentage protein ([K-S.sub.protein]). The observed numbers of births or fetuses were used for the tests.
Synchrony of births
Because births were recorded by month our measure of synchrony was the degree to which births were distributed over the 12 months. This measure was the Index of Evenness J', which for 12 months is given by
J' = ([sigma][p.sub.i] log [p.sub.i]/log 12) (8)
where [p.sub.i] is the proportion of births of species i occurring in any month. Evenness (J') ranges from zero for maximum synchrony to unity for no synchrony (Krebs 1989:366).
Mean annual rainfall in Serengeti varies from 1050 mm in the northwest woodlands to 550 mm on the southeastern plains. Rainfall is highly seasonal with a wet period from about November to June and a dry period from July to October. We illustrate this in Fig. 1 for the northcenter of the system (solid bars). On the plains (Fig. 1, open bars) the wet season starts later (December) and ends earlier (May) but otherwise the seasonality is similar throughout the region.
With the start of the rains grass and herbs begin growth and green biomass accumulates until May or June. This is illustrated for the woodlands (Fig. 2A) from data in Sinclair (1977a) (Eq. 3), and differs only in detail with growth on the plains (Eq. 2). With the drop in rainfall in July, green biomass declines. Dry season storms in the north maintain some green biomass (Eq. 6) but the plains receive too little rain for any growth at this time.
The first growth of grass is high in protein content. As green biomass increases, however, percentage protein declines as illustrated in Fig. 2B (Eq. 4 from data in Sinclair 1977a). Both protein and green biomass were calculated separately for the north, east, west, and plains regions of the ecosystem, based on their respective rainfall regimes.
Available biomass of twig shoots and green leaves of Acacia trees measured by Pellew (1980) is shown in Fig. 2a. Biomass increases in the rains and decreases in the dry season. Tree green biomass is about one-tenth that of grasses and is more evenly distributed over the year. Also, these trees differ from grasses in that they anticipate the rains by one or two months with new growth starting in September. Thus crude protein is highest in this month, declining gradually over the rest of the year (Fig. 2B).
Distribution of births
Table 1 summarizes the abundance and distribution of species from data in Campbell and Borner (1995). Table 2 gives the sample sizes of newborn and adult females (or total adults for zebra, warthog) for the eight species counted on road transects over the years 1977-1995. Table 3 gives the monthly proportion of births for all thirteen species. The Kolmogorov-Smirnov values for differences in frequency distributions between observed births and those predicted by even monthly distributions, green biomass, and percentage crude protein for the same sample size are given in Table 4.
For the six grazing species with precocial young (wildebeest, topi, kongoni, warthog, buffalo, zebra), Fig 3. shows the frequency distribution of births. Four of these species show a marked peak of births near the beginning of the wet season. However, buffalo births coincide with the peak rainfall in April, and the kongoni birth peak occurs in the middle of the dry season.
Fig. 4 shows the frequency distribution of births for the less precocial "hider" species, and giraffe. Of these the two gazelles produce their young in the early rains, while oribi, of similar size to the gazelles, have a birth peak that coincides with late wet season. Dikdik, the smallest of these antelopes, produces two young per year. Based on the small sample given in Hendrichs and Hendrichs (1971) there are two peaks, one in early wet season the other at the end of the rains. Waterbuck distribute their births over the year with a small peak in late wet season. Impala have fewer births in the main wet months and have peaks in dry season and early wet season. Giraffe, the only species feeding on tree shoots, have their peak of births in the dry season al though births occur frequently all through the year. Flowering, fruiting, and new shoots of Acacia trees (a major food of giraffe) occurs in mid dry season when crude protein levels reach 21.0 [+ or -] 1.7% of dry mass [mean [+ or -] 1 SE) compared to 15. 3 [+ or -] 0.9% for older shoots later in the year (Pellew 1980).
We divided species into grazers, mixed feeders, and browsers. The diet of each species was obtained either from our own observations or from the literature: for wildebeest (Watson 1969, Kreulen 1975), topi (Duncan 1975, Murray 1993, Murray and Brown 1993), buffalo (Sinclair 1977a), waterbuck (Field 1972, Spinage 1982), zebra (Klingel 1969), Thomson's gazelle (Gwynne and Bell 1968), kongoni (Stanley-Price 1978, Murray 1993), dikdik (Hendrichs and Hendrichs 1971), giraffe (Pellew 1980), and oribi (Mduma 1991, Mduma and Sinclair 1994).
The Kolmogorov-Smirnov values provide a measure of how different observed birth distributions were from those predicted with the same sample size (number of births corrected for monthly sample size) by food biomass, crude protein content of food, and an even distribution through the year (Table 4). Sample size for dikdik was too small to provide meaningful values and we do not consider this further. All but two species showed birth distributions significantly different from those predicted by food biomass. The two exceptions are buffalo and oribi, whose distributions are very similar to biomass (Figs. 3 and 4). Eight species showed different birth distributions from those predicted by the protein content distribution. In contrast, giraffe, waterbuck, kongoni, and zebra showed birth distributions very similar to that for percentage protein. Nine species showed birth distributions that differed from an even distribution through the year. However, births of giraffe, waterbuck, and Thomson's gazelle did not diff er from an even monthly distribution (Fig. 4). Thus, at one extreme there is a group of species that had high Kolmogorov-Smirnov values and that differed from all predicted distributions by having short, synchronized seasons (wildebeest, topi, warthog, and Grant's gazelle). At the other extreme there is a group with birth distributions similar to that of the food supply (buffalo, oribi, kongoni, and zebra), or births distributed evenly through the year (dikdik, Thomson's gazelle, giraffe, and waterbuck).
There is some indication of a pattern in the timing of the peak of births. The number of months that the birth peak occurred before or after the food-protein peak or green-biomass peak was found by cross-correlating the birth distribution with that predicted by the food protein or biomass distributions. The birth distribution was shifted one month at a time, and the maximum correlation coefficient was used as the indicator of the lag (Table 5). For grazing species the lag between birth peak and protein peak is a positive function of metabolic body size ([W.sup.0.75], in kilograms) (Spearman rank correlation [r.sub.s] = 0.73, n = 8 grazing species, P [less than] 0.05) whereas the lag with biomass is a negative function of body size ([r.sub.s] = -0.76, n = 8, P [less than] 0.05) (Fig. 5). Tentatively, small grazers produce their young early in the wet season in anticipation of the high protein peak. These species are out of phase with high green biomass. In contrast, large species tend to produce their young m ore in phase with high biomass and after the protein peak. This relationship is not evident with the five browsers and mixed feeders (Grant's gazelle, impala, oribi, dikdik, and giraffe). Giraffe births are directly in phase with their protein peak (Pearson correlation r = 0.57, P [less than] 0.05) and out of phase with biomass (r = -0.566, P [less than] 0.05).
Synchrony of births
We measured the synchrony of births using the index of evenness (J') and compared that with the evenness of rainfall, green biomass, and percentage food protein through the year (Table 6). Despite the evident seasonality of rainfall and food supply, both rain and protein are fairly evenly distributed over the months with all areas having J' [greater than] 0.9. Green biomass is slightly less evenly distributed.
In contrast, the evenness of births varies widely with species. The most uneven is wildebeest with [J'.sub.births] [approximate] 0.35 and this species is highly synchronized relative to rainfall and food supply. There is a group of four species (topi, warthog, buffalo, and dikdik) with [J'.sub.births] [sim]0.5-0.7 and these are also more synchronized than predicted by rainfall and food supply. At the other end of the scale, kongoni has the least synchronized birth season and births are more evenly distributed than rainfall and food biomass. This is also the case for Thomson's gazelle and zebra. The remaining five species show a more even distribution of births than food biomass and only slightly more synchrony than percentage protein and rainfall.
Phenology of births
We compare our results with our initial prediction that births should be related to food supply in a way consistent with their known food requirements. As a general pattern birth peaks occur sometime in the wet season when there is more green food. However, even this broad pattern is not universal, for giraffe, impala, and kongoni have their peak of births at the driest time of year. Giraffe births coincide with the flush of tree shoots and their high protein content.
Among grazers, the pattern suggested by Fig. 5 is consistent with known metabolic constraints. Births of small species appear to coincide with the protein peak and those of large species coincide with the green biomass peak. This is consistent with the idea that small ungulates have high specific metabolic rates and require high protein and energy but less bulk relative to large species (Kleiber 1947, Bell 1970, Jarman and Sinclair 1979, Clutton-Brock and Harvey 1983). Thus, grass-feeding ungulates may be adapting their birth seasons to their own specific food peaks, and these differ from species to species. This pattern of births and protein, therefore, is complex. It also does not account for all species, particularly the browsers or mixed feeders such as dikdik and oribi. Dikdik have two births per year 6 months apart. If they are to produce both births in the wet season they are constrained in timing. The first must occur in November--January if the second is to occur in the late wet season May to July.
It is not clear why the oribi, a small species, produces its young at the peak biomass, unless other factors override nutritional intake. Oribi live in habitats of very tall grass in which they hide. They are also subject to high levels of predation (from telemetry studies; P. Arcese, personal observation). They may be using the period of maximum grass height in late wet season to hide newborn young. Moreover, oribi are unusual in that they have a gestation of 7-7.5 mo, are capable of an immediate postpartum estrus, and often produce two young within a period of 14-15 mo (Jongejan et al. 1991). Oribi that do give birth more than once per year will have at least one young outside of the primary peak in births, which coincides with maximum biomass and cover available for young. Thus, it may be that some oribi trade off an increased number of young over their lifetimes at the cost of having a fraction of their young during suboptimal periods with respect to biomass and cover. Oribi, being highly selective mixed feeders, can also choose herbs and young grass leaves within tall grass stands so that high grass biomass may not be a nutritional impediment to them.
Synchrony of births
We predicted that species with precocial newborn should have synchronized births as an antipredator strategy. Four of the five most synchronized species (wildebeest, topi, warthog, and buffalo) fall into this category. However, dikdik, the other synchronized species, is very much a "hider" species. We also predicted that synchronized births should occur in locally abundant species. Wildebeest, topi, and buffalo are abundant but warthog and dikdik are not. Furthermore, both zebra and Thomson's gazelle are locally abundant and neither show greater synchrony than predicted by their food supply. This is all the more curious since zebra and wildebeest migrate together, have close spatial correlation, and overlap in food preferences by [greater than]90% (Maddock 1979, Sinclair 1985). Thus, the predator-satiation hypothesis (Rutberg 1987) is supported by some species but not all of them.
Five of seven species that we predicted from the "predator avoidance" hypothesis (Ims 1990) to show a more even distribution of births than their food supply conformed to this prediction (impala, Thomson's and Grant's gazelle, waterbuck, and oribi). They all had more even birth distributions than that of food biomass and a birth distribution similar to that of food protein content. Waterbuck are a classic example of this category. They live at low density and hide their young. Because this species is closely tied to water and riverine vegetation it may not need to rely on seasonal rainfall for optimal food supplies to give birth. Thus, they can spread their births through the year, a result similar to that found by Spinage (1969) and Wirtz and Kaiser (1988) in Uganda and Kenya. We add the caveat, however, that the data on waterbuck are sparse and dates of birth are the least accurate of all the species.
Giraffe births appear to coincide with food quality and are only slightly more synchronized than this food. This species does not fit easily into the classical groupings of Jarman (1974). Giraffe have a gestation period of 15 months and hence it is difficult for them to produce young at a given time of year without delaying reproduction and reducing fecundity. Dikdik showed more birth synchrony than predicted. Since this is the only species in our sample that could produce two young per year it is possible that they were constrained by their food supply to become synchronized.
The kongoni is perhaps the most anomalous species in our sample. It is a large grazer and has precocial young. It is closely related to topi and wildebeest, both of which show extreme birth synchrony. Yet the kongoni has the most evenly distributed births of all the species. Its main difference in life history from the other two relatives is that it lives at low density in small widely distributed groups. In general, the large difference we observe in birth synchrony among closely related species suggests that ecology rather than phylogeny better explains the differences. A preliminary check on these interspecific comparisons using the method of independent contrasts (Felsenstein 1985) confirms these impressions (J. Brashares, personal communication).
It is tempting to conclude, therefore, that birth synchrony has coevolved with living in large groups, until we see that warthog have extreme birth synchrony and they live as isolated individuals or small families. At present, therefore, we can find no single feature that predicts all species birth distributions. A combination of the phenology of food supply plus antipredator adaptations accounts for most but not all these distributions.
We thank the directors and boards of trustees of the Tanzania National Parks and Serengeti Wildlife Research Institute for their permission to work in the Serengeti. We thank Stan Boutin, Holly Dublin, Ray Hilborn, Ulrike Hilborn, Stephen Makacha, Anne Sinclair, and John Wilmshurst for help in data collection, and J. Brashares for running the independent contrasts. The ideas of Allen Rutberg and RoIf Ims stimulated this analysis. This work is part of the Serengeti Biodiversity Program funded by the Canadian Natural Sciences and Engineering Research Council and the National Geographic Society.
(1.) Centre for Biodiversity Research, 6270 University Boulevard, University of British Columbia, Vancouver, Canada V6T 1Z4
(2.) Department of Wildlife Ecology, 226 Russell Laboratories, 1630 Linden Drive, University of Wisconsin, Madison, Wisconsin 53706 USA
(3.) Present address: Serengeti Wildlife Research Institute, P.O. Box 661, Arusha, Tanzania.
(4.) Present address: Centre for Applied Conservation Biology, Faculty of Forestry, University of British Columbia, Vancouver, Canada V6T 1Z4.
Anderson, J. L. 1979. Reproductive seasonality of the nyala Tragelaphus angasi; the interaction of light, vegetation phenology, feeding style and reproductive physiology. Mammal Review 9:33-46.
Bell, R. H. V. 1970. The use of the herb layer by grazing ungulates in the Serengeti. Pages 111-123 in A. Watson, editor. Animal populations in relation to their food resources. Blackwell Scientific, Oxford, UK.
Bourliere, F, G. Morel, and G. Galat. 1976. Les grand mammiferes de la basse vallee du Senegal et leurs saisons de reproduction. Mammalia 40:401-412.
Bradley, R. M. 1977. Aspects of the ecology of the Thomson's gazelle in the Serengeti National Park, Tanzania. Dissertation. Texas Agricultural and Mechanical University, College Station, Texas, USA.
Braun, H. M. H. 1973. Primary production in the Serengeti: purpose, methods and some results of research. Annals of the University d'Abidjan (E) 6:171-188.
Brokx, P. A. 1972. Ovarian composition and aspects of the reproductive physiology of Venezuelan white-tailed deer (Odocoileus virginianus gymnotis). Journal of Mammalogy 53:760-773.
Bronson, F. H. 1989. Mammalian reproductive biology. University of Chicago Press, Chicago, Illinois, USA.
Bunnell, F 1982. The lambing period of mountain sheep: synthesis, hypotheses, and tests. Canadian Journal of Zoology 60:1-14.
Campbell, K., and M. Borner. 1995. Population trends and distribution of Serengeti herbivores: implications for management. Pages 117-145 in A. R. E. Sinclair and P. Arcese, editors. Serengeti II: dynamics, management and conservation of an ecosystem. University of Chicago Press, Chicago, Illinois, USA.
Clutton-Brock, T. H. and P. H. Harvey. 1983. The functional significance of variation in body size among mammals. Pages 632-663 in J. F Eisenberg and D. G. Kleiman, editors. Advances in the study of mammalian behavior. Special Publication number 7. American Society of Mammalogists, Shippensburg State College, Shippensburg, Pennsylvania, USA.
Dasmann, R. F.; and A. S. Mossman. 1962. Reproduction in some ungulates in Southern Rhodesia. Journal of Mammalogy 43:533-537.
David, J. H. M. 1978. Observations on social organization of springbok, Antidorcos marsupialis, in the Bontebok National Park, Swellendam. Zoologica Africana 13:115-122.
Duncan, P. 1975. Topi and their food supply. Dissertation. University of Nairobi, Nairobi, Kenya.
Estes, R. D. 1966. Behaviour and life history of the wildebeest (Connochaetes taurinus Burchell). Nature 212:999-1000.
Estes, R. D. 1976. The significance of breeding synchrony in the wildebeest. East African Wildlife Journal 14:135-152.
Estes, R. D., and R. K. Estes. 1979. The birth and survival of wildebeest calves. Zeitschrift fur Tierpsychologie 50:45-95.
Fairall, N. 1968. The reproductive seasons of some mammals in the Kruger national Park. Zoologica Africana 3:189-210.
Felsenstein, J. 1985. Phylogenies and the comparative method. American Naturalist 125:1-15.
Field, C. R. 1972. The food habits of wild ungulates in Uganda by analysis of stomach contents. East African Wildlife Journal 10:17-42.
Findlay, C. S., and F. Cooke. 1982. Breeding synchrony of the lesser snow goose (Anser coerulescens) II. the adaptive value of reproductive synchrony. Evolution 36:786-799.
Gwynne, M. D., and R. H. V. Bell. 1968. Selection of vegetation components by grazing ungulates in the Serengeti National Park. Nature 220:390-393.
Hendrichs, H., and U. Hendrichs. 1971. Dikdik und Elefanten. Okologie und Soziologie zweier afrikanischer Huftiere. R. Piper, Munchen, Germany.
Ims, R. A. 1990. On the adaptive value of reproductive synchrony as a predator-swamping strategy. American Naturalist 136:485-498.
Jarman, F J. 1974. The social organization of antelope in relation to their ecology. Behaviour 48:215-266.
Jarman, P. J., and A. R. F. Sinclair. 1979. Feeding strategy and the pattern of resource partitioning in ungulates. Pages 130-163 in A. R. E. Sinclair and M. Norton-Griffiths, editors. Serengeti: dynamics of an ecosystem. University of Chicago Press, Chicago, Illinois, USA.
Jongejan, G., P. Arcese, and A. R. E. Sinclair. 1991. Growth, size and the timing of births in an individually identified population of oribi. African Journal of Ecology 29:340-352.
Kleiber, M. 1947. Body size and metabolic rate. Physiological Reviews 27:511-541.
Klingel, H. 1969. The social organisation and population ecology of the plains zebra (Equus quagga). Zoologica Africana 4:249-263.
Krebs, C. J. 1989. Ecological methodology. Harper and Row, New York, New York, USA.
Kreulen, D. 1975. Wildebeest habitat selection on the Serengeti plains, Tanzania, in relation to calcium and lactation: a preliminary report. East African Wildlife Journal 13:297-304.
Kruuk, H. 1964. Predators and antipredation behaviour of the black-headed gull (Larus ridibundus). Behaviour, Supplement 11. Brill, Leiden, The Netherlands.
Linnell, J. D., and R. Andersen. 1998. Timing and synchrony of birth in a hider species, the roe deer Capreolus capreolus. Journal of Zoology 244:497-504.
McNaughton, S. J. 1979. Grassland-herbivore dynamics. Pages 46-81 in A. R. E. Sinclair and M. Norton-Griffiths, editors. Serengeti: dynamics of an ecosystem. University of Chicago Press, Chicago, Illinois, USA.
McNaughton, S. J. 1983. Serengeti grassland ecology: the role of composite environmental factors and contingency in community organization. Ecological Monographs 53:291-320.
Mduma, S. A. R. 1991. Population ecology of oribi in Serengeti National Park, Tanzania. Thesis. University of Dar es Salaam, Dar es Salaam, Tanzania.
Mduma, S. A. R. 1995. Distribution and abundance of oribi, a small antelope. Pages 220-230 in A. R. E. Sinclair and P. Arcese, editors. Serengeti II: dynamics, management and conservation of an ecosystem. University of Chicago Press, Chicago, Illinois, USA.
Mduma, S. A. R., R. Hilborn, and A. R. E. Sinclair. 1998. Limits to exploitation of Serengeti wildebeest and implications for its management. Pages 243-265 in D. M. Newbery, H. B. Underwood, and H. H. T. Prins, editors. Dynamics of tropical communities. Blackwell Scientific, Oxford, UK.
Mduma, S. A. R., and A. R. E. Sinclair. 1994. The function of habitat selection by oribi in Serengeti, Tanzania. African Journal of Ecology 32:16-29.
Murray, M. G. 1982. The rut of impala: aspects of seasonal mating under tropical conditions. Zeitschrift fur Tierpsychologie 59:319-337.
Murray, M. G. 1993. Comparative nutrition of wildebeest, hartebeest, and topi in the Serengeti. African Journal of Ecology 31:172-177.
Murray, M. G. and D. Brown. 1993. Niche separation of grazing ungulates in the Serengeti: an experimental test. Journal of Animal Ecology 62:380-389.
Pellew, R. A. 1980. The production and consumption of Acacia browse and its potential for animal protein production. International symposium on browse in Africa (Addis Ababa, 1980). International Livestock Centre for Africa, Addis Ababa, Ethiopia.
Rutberg, A. T. 1987. Adaptive hypotheses of birth synchrony in ruminants: an interspecific test. American Naturalist 130: 692-710.
Sekulic, R. 1978. Seasonality of reproduction in the sable antelope. East African Wildlife Journal 16:177-182.
Sinclair, A. R. E. 1975. The resource limitation of trophic levels in tropical grassland ecosystems. Journal of Animal Ecology 44:497-520.
Sinclair, A. R. E. 1977a. The African buffalo. University of Chicago Press, Chicago, Illinois, USA.
Sinclair, A. R. E. 1977b. Lunar cycle and timing of mating in Serengeti wildebeest. Nature 267:832-833.
Sinclair, A. R. E. 1979a. The Serengeti environment. Pages 31-45 in A. R. E. Sinclair and M. Norton-Griffiths, editors. Serengeti: dynamics of an ecosystem. University of Chicago Press, Chicago, Illinois, USA.
Sinclair, A. R. E. 1979b. The eruption of the ruminants. Pages 82-103 in A. R. E. Sinclair and M. Norton-Griffiths, editors. Serengeti: dynamics of an ecosystem. University of Chicago Press, Chicago, Illinois, USA.
Sinclair, A. R. E. 1985. Does interspecific competition or predation shape the African ungulate community? Journal of Animal Ecology 54:899-918.
Skinner, J. D., J. H. M. Van Zyl, and L. G. Oates. 1974. The effect of season on the breeding cycle of plains antelope of the western Transvaal highveld. Journal of the South African Wildlife Management Association 4:15-23.
Skinner, J. D., J. H. M. Van Zyl, and J. A. H. Van Heerden. 1973. The effect of season on reproduction in the black wildebeest and red hartebeest in South Africa. Journal of Reproduction and Fertility, Supplement 19:101-110.
Spinage, C. A. 1969. Reproduction of the Uganda defassa waterbuck, Kobus defassa ugandae Neumann. Journal of Reproduction and Fertility 18:445-457.
Spinage, C. A. 1973. The role of photoperiodism in the seasonal breeding of tropical African ungulates. Mammal Review 3:71-106.
Spinage, C. A. 1982. A territorial antelope: the Uganda waterbuck. Academic Press, New York, New York, USA.
Stanley-Price, M. R. 1978. The nutritional ecology of Coke's hartebeest (Alcelaphus buselaphus cokei) in Kenya. Journal of Applied Ecology 15:33-49.
Underwood, R. 1978. Aspects of kudu ecology at Loskop Dam Nature Reserve, Eastern Transvaal. South African Journal of Wildlife Research 8:43-47.
Walther, F. R. 1972. Social grouping in Grant's gazelle (Gazella grand Brooke ) in the Serengeti National Park. Zeitschrift for Tierpsychologie 31:348-403.
Watson, R. M. 1969. Reproduction of wildebeest, Connochaetes taurinus albojubatus Thomas, in the Serengeti region, and its significance to conservation. Journal of Reproduction and Fertility. Supplement 6:287-310.
Wirtz, P., and P. Kaiser. 1988. Sex differences and seasonal variation in habitat choice in a high density population of Waterbuck, Kobus ellipsiprymnus (Bovidae). Zeitschrift fur Saugetierkunde 53:162-169.
Numbers and distribution of ungulate species in the Serengeti ecosystem considered in this study. Body Population mass Species (no. inds.) (kg) Distribution Wildebeest 1 300 000 163 Migratory Thomson's gazelle 450 000 16 Migratory Zebra 200 000 220 Migratory Topi 95 000 110 West, North Impala 75 000 42 All woods Buffalo 45 000 450 All woods Grant's gazelle 26 000 43 Plains, East Kongoni 14 000 126 East Giraffe 9 000 800 All woods Oribi 7 000 19 North Warthog 6 000 53 Widespread Waterbuck 1 600 175 Riverine Dikdik ... 5 All woods Note: Data from Campbell and Borner (1995) and Mduma (1995). Total numbers of newborn and adult female Serengeti ungulates (all adults for zebra, warthog) recorded in each month, summed over1977-1997. Month Wildebeest Topi Warthog Impala Total numbers of newborn 1 244 9 254 44 2 930 4 936 11 3 4143 2 4148 10 4 74 1 79 1 5 5 0 10 4 6 5 0 11 74 7 1 1 9 116 8 0 12 20 75 9 0 270 279 19 10 0 866 876 22 11 0 436 447 74 12 0 515 527 77 Total 5402 2116 498 527 Total numbers of adult (or adults of both sexes [warthog, zebra]) 1 8861 1099 503 2156 2 4862 2782 317 2891 3 9518 3189 208 3705 4 4399 1444 144 1145 5 4456 1732 232 3218 6 10514 5295 614 6234 7 8391 6969 990 7756 8 3006 2978 767 4330 9 3456 2510 570 3584 10 5925 1840 520 2053 11 3571 2383 686 3935 12 1709 4097 1126 6356 Total 68668 36318 6677 47363 Month Waterbuck [+] Giraffe Zebra Total numbers of newborn 1 8 9 59 2 1 3 33 3 3 10 8 4 15 0 22 5 12 1 63 6 5 12 19 7 2 14 39 8 7 16 23 9 9 27 24 10 7 4 10 11 9 10 41 12 6 30 107 Total 84 136 448 Total numbers of adult (or adults of both sexes [warthog, zebra]) 1 83 254 3143 2 104 276 5541 3 39 230 1083 4 88 52 3066 5 135 203 9476 6 204 378 4437 7 113 633 10113 8 87 377 6297 9 52 340 6405 10 66 200 3126 11 96 306 4556 12 170 520 5990 Total 1237 3769 63231 Month Kongoni Total numbers of newborn 1 15 2 13 3 10 4 10 5 26 6 40 7 47 8 44 9 63 10 23 11 19 12 63 Total 373 Total numbers of adult (or adults of both sexes [warthog, zebra]) 1 239 2 320 3 384 4 268 5 532 6 1077 7 1059 8 442 9 560 10 340 11 389 12 687 Total 6297 (+.)Data for waterbuck include small calves, back-calculated to month of birth. Kolmogorov-Smirnov values for monthly birth distribution relative to even distribution ([K-S.sub.even]) and relative to that predicted by foot-plant biomass ([K-S.sub.biomass]) and distributions of protein content ([K-S.sub.protein]) for respective areas of the Serengeti. Comparisons vs. observed birth distributions Species Area [K-S.sub.even] P [K-S.sub.biomass] Wildebeest Plains 0.72 0.000 0.76 Topi North 0.65 0.000 0.92 Warthog Northcenter 0.58 0.000 0.88 Buffalo West 0.46 0.000 0.22 Grant's gazelle Plains 0.41 0.000 0.48 Oribi North 0.21 0.001 0.15 Dikdik East 0.19 0.738 0.36 Thomson's Plains 0.20 0.063 0.59 gazelle Impala Northcenter 0.17 0.000 0.25 Giraffe Northcenter 0.19 0.016 0.32 Waterbuck Northcenter 0.11 0.763 0.38 Kongoni East 0.17 0.000 0.51 Zebra Plains 0.13 0.002 0.39 Maximum value 0.92 0.92 Species P [K-S.sub.protein] P Wildebeest 0.000 0.63 0.000 Topi 0.000 0.57 0.000 Warthog 0.000 0.50 0.000 Buffalo 0.169 0.48 0.000 Grant's gazelle 0.000 0.32 0.000 Oribi 0.044 0.29 0.000 Dikdik 0.026 0.23 0.373 Thomson's 0.000 0.22 0.000 gazelle Impala 0.000 0.16 0.000 Giraffe 0.000 0.15 0.108 Waterbuck 0.000 0.12 0.673 Kongoni 0.000 0.10 0.037 Zebra 0.000 0.05 0.540 Maximum value 0.92 Note: Distributions were different if P [less than] 0.004 (Bonferroni correction). Evenness (J') values for monthly birth distribution of Serengeti ungulate species relative to evenness for rainfall, food biomass, and food percentage protein, and the ratios of birth J' to rain and food J'. [J'.sub.births]/ Species Area [J'.sub.births] [J'.sub.rain] [J'.sub.rain] Wildebeest Plains 0.354 0.910 0.389 Topi North 0.515 0.965 0.534 Warthog Northcenter 0.629 0.961 0.665 Dikdik East 0.680 0.951 0.715 Buffalo West 0.695 0.935 0.743 Grant's gazelle Plains 0.771 0.910 0.847 Giraffe Northcenter 0.894 0.961 0.930 Impala Northcenter 0.894 0.961 0.930 Oribi North 0.933 0.965 0.967 Waterbuck Northcenter 0.933 0.961 0.971 Zebra plains 0.924 0.910 1.015 Thomson's gazelle plains 0.924 0.910 1.015 Kongoni East 0.962 0.951 1.012 Maximum value 1.000 1.000 [J'.sub.births] Species [J'.sub.biomass] [J'.sub.biomass] [J'.sub.protein] Wildebeest 0.736 0.481 0.938 Topi 0.835 0.617 0.976 Warthog 0.835 0.753 0.975 Dikdik 0.831 0.818 0.976 Buffalo 0.818 0.850 0.977 Grant's gazelle 0.736 1.048 0.938 Giraffe 0.962 0.929 0.997 Impala 0.835 1.071 0.975 Oribi 0.835 1.117 0.976 Waterbuck 0.835 1.255 0.975 Zebra 0.736 0.938 Thomson's gazelle 0.736 1.255 0.938 Kongoni 0.831 1.158 0.976 Maximum value [J'.sub.births] Species [J'.sub.protein] Wildebeest 0.377 Topi 0.528 Warthog 0.645 Dikdik 0.697 Buffalo 0.711 Grant's gazelle 0.822 Giraffe 0.897 Impala 0.917 Oribi 0.956 Waterbuck 0.957 Zebra 0.985 Thomson's gazelle 0.985 Kongoni 0.986 Maximum value The lag in months of the birth peak before (-) or after (+) the peak of food biomass and food percentage protein using the Pearson correlation coefficient (r) to identify the lag. Biomass Protein content (%) Species Lag (mo) r P Lag (mo) r p Kongoni 5 0.632 0.02 -3 0.757 0.01 Warthog 4 0.642 0.02 -2 0.920 0.001 Topi 5 0.707 0.01 -1 0.935 0.001 Thomson's gazelle 6 0.846 0.001 -2 0.601 0.05 Zebra -5 0.791 0.001 0 0.877 0.001 Impala 4 0.614 0.02 0 0.440 NS Wildebeest -3 0.772 0.01 2 0.718 0.01 Buffalo -1 0.864 0.001 4 0.824 0.001 Waterbuck -2 0.215 NS 5 0.675 0.01 Giraffe -5 0.516 NS 0 0.570 0.05 Grant's gazelle -4 0.876 0.001 1 0.883 0.001 Dikdik 1 0.245 NS 1 0.650 0.02 Oribi 0 0.861 0.001 5 0.774 0.01
|Printer friendly Cite/link Email Feedback|
|Author:||SINCLAIR, A. R. E.; MDUMA, SIMON A. R.; ARCESE, PETER|
|Date:||Aug 1, 2000|
|Previous Article:||CHARACTER DISPLACEMENT AND RELEASE IN THE SMALL INDIAN MONGOOSE, HERPESTES JAVANICUS.|
|Next Article:||SPATIAL SYNCHRONY IN POPULATIONS OF BIRDS: EFFECTS OF HABITAT, POPULATION TREND, AND SPATIAL SCALE.|