The term life history broadly refers to a host of anatomical, physiological, and behavioral characteristics of a species, although it is often restricted to demographic traits, such as age-specific birth and death rates, or developmental traits, such as growth rates and body size. While aspects of reproductive behavior, social organization, or feeding ecology in elephants have been recognized as the outcome of natural selection, there has been little attempt to discuss demographic traits in populations in relation to selective forces. A possible exception is the discussion of changes in birth rates with density, for which the evidence in any case is equivocal. There has been insufficient appreciation of the possible influence of habitat and its characteristics in the evolution of (demographic) life history phenomena in elephant populations. Elephants after all inhabit regions from near-desert conditions to tropical lowland rain forests and montane forests. Adaptations needed to deal with variations in resource abundance (forage type, quality, and dispersion), climate (incidence of drought), or even energetics (physiological costs of moving over steep terrain or long distances across arid landscapes) could surely be expected to be reflected in demographic traits. The problem of interpreting life history traits does not lie only with elephants, but extends to other mammals. In the words of Mark Boyce, "There certainly is a wealth of theory, [but] much of it fails to adequately link demography with environment and it is mostly untested" (1988b, p. 358). I therefore only attempt a rudimentary discussion of this topic as it relates to elephant demography.
The concept of r (intrinsic growth rate) and K (carrying capacity) selection, originally formulated as a theory of density-dependent natural selection, is one framework in which to consider the evolution of life history traits. Unfortunately, the rather indiscriminate application of this theory has led to its virtual abandonment by the theorists, even though the r-K terminology still prevails in the ecological literature.
Basically, this theory suggests that there are trade-offs between the ability to do well under conditions of low density (r selection) and high density (K selection), with the comparisons made in relation to the carrying capacity of the habitat. The prediction is, therefore, that at low population density, those genotypes with high r (reproductive ability) will have a selective advantage, while at high population density, this is reversed in favor of genotypes with high K (say, adult survival). Among mammals, a suggested response to increased crowding is a sequence of (1) increased juvenile mortality, (2) increased age of sexual maturity, (3) decreased birth rates, and (4) increased adult mortality. Even when real populations do show short-term demographic responses to changing density, it is important to differentiate the proximate response of a variable such as age of sexual maturity from the process of natural selection.
The evidence for density-dependent changes in reproduction among elephants is generally weak. The only population for which clear evidence is available to the best of my knowledge is that in Uganda. When J. S. Perry examined a sample of culled elephants there during 1946-1950, he observed that a cow "often begins to breed at an age of 10 years or less" (1953, p. 104) (rather vague and not referring to mean age of conception), and that the intercalving interval (based on more objective criteria) was just under 4 years. By the time Richard Laws and his team examined the high-density population of Murchison Falls during 1966-1967, these fecundity measures had increased considerably, but following a steep reduction in density through culling and poaching had again swung back when Rob Malpas conducted his study during 1974 (see table 7.1).
There have been attempts to derive the relationship between density and fecundity from data across African savanna elephant populations (for instance, by Charles Fowler and Tim Smith, see section 7.2.2). Malan Lindeque's more critical examination of the data showed that the relationship between elephant density and age at first conception is not significant for southern Africa and is weakly significant for eastern African populations only after removal of the anomalous data for Lake Manyara.
It of course might be argued that elephant density per se is not the appropriate variable given a certain variation in habitat productivity even across savanna woodland habitat, but the density as a proportion of the carrying capacity (notoriously difficult to estimate) is the appropriate variable. The strongest argument against density dependence in reproduction comes from Addo, where the elephants have consistently maintained a phenomenal 7% annual popula tion growth rate, obviously spurred by high fecundity, over half a century to attain a very high density (see chapter 9). Obviously, a slowing in growth eventually will have to occur in this fenced-in population, but no one is sure when the tide will turn. Asian elephants in southern Indian habitats, now reaching densities comparable to the highest known values in Africa, also do not as yet show any slow down in fecundity.
We know much less about how mortality rates change with increasing density in elephant populations. There are hints of lower survivorship among young elephants, but no evidence that adults suffer similarly under conditions of high density, except during a drought.
A more interesting framework for life history evolution in elephants is, in my opinion, that of fluctuating versus stable environments. This is quite distinct from density-dependent selection; the predictions of r-K selection theory do not directly apply to selection under fluctuating environments, although they may share certain common features.
Environmental fluctuation can be a powerful force in evolution, but is not easy to unravel. To quote Mark Boyce again, "Environmental variability is a virtual Pandora's box of selective forces which can influence the evolution of life histories, and there is still much that we do not understand about the nature of selection in fluctuating environments" (1988a, p. 16). All environments fluctuate to a certain extent, both seasonally and from one year to another; what is of interest is the relative degree of fluctuation across environments. A well-established climatic pattern is the inverse relationship between annual rainfall and its variability. Thus, a region with low rainfall experiences high variability (or "coefficient of variation" in statistical terms), while one with high rainfall has low variability. The semiarid savanna is thus a highly fluctuating environment not only on a seasonal scale, but also on an interannual scale, while the equatorial rain forest is a relatively stable environment, with rainfall being more evenly distributed over the year and more predictable from one year to another.
Let us consider how elephant life histories should be molded to deal with a highly fluctuating environment, such as in the semiarid savannas. When high-quality resources are available only seasonally, traits promoting rapid (metabolic) growth and high fecundity to take advantage of this distinct seasonal flush would be favored. Larger body size could be one consequence in the more fluctuating habitat, and this could be reflected in other traits correlated with size (see chapter 5 and appendix 2). Likewise, higher fecundity would be expressed in female elephants through an early age of sexual maturity and shorter intercalving interval. This would enable a more rapid population growth when favorable conditions return after a drought and a population decline.
I am not arguing for "group selection," but rather that selection operating on individual elephants in semiarid habitats over the centuries has given us populations with individuals that have reproductive traits that promote rapid growth. Obviously, it may also be true that individuals should delay reproduction (late age at maturity or longer calving intervals) when conditions are unfa vorable. Thus, one could expect to see greater flexibility in reproductive traits in elephants of semiarid regions. The opposite can be expected in elephants that have evolved in more stable rain forests. A later age of sexual maturity and longer intercalving intervals, with less plasticity in these fecundity traits, would be sufficient as long as these are close to or above replacement-level rates. Fecundity traits would also be influenced by lower-quality forage in rain forests. Even if they were to exist at much lower densities, elephant populations in rain forests would themselves have fluctuated little and persisted at close to the carrying capacity for longer periods. I would therefore argue that it is more likely for rain forest elephants to have been selected for K-related rather than r-related traits. This could provide a link, however crude, between models of selection under density dependence and selection under environmental variability.
What is the evidence that elephant life histories conform to the above expectations? The earliest ages at first conception in female elephants, in the range of 9-13 years, are known only from arid and semiarid habitats such as the southern and eastern African savannas. These include Etosha, Kruger, Hwange, Gonarezhou, Tsavo, Mkomazi, and Kabalega. With an average annual rainfall from as low as 10 cm to not more than about 75 cm, an exception being Kabalega (more than about 100 cm), these are also the most variable environments across the elephant's range. Mean intercalving intervals are also shorter (in the range of 3.5-4.5 years) in semiarid Africa. The elephant population at Kabalega has also shown considerable plasticity in reproductive traits.
What about the elephants in regions of high rainfall? A late age of first conception and a long intercalving interval is reported for the moist Budongo forest, where annual rainfall averages over 150 cm. It is also interesting that J. S Perry, quoting H. Hedigar, observes that captive African forest elephants in the erstwhile Belgian Congo "do not reach puberty until about 15 years of age" (1953, p. 104). This is a clear hint that African forest elephants inhabiting the high rainfall zone may have a relatively late age of sexual maturity. Asian elephants in the regions of medium-to-high rainfall (100-200 cm) of southern India and Myanmar also show a relatively late age of sexual maturity (13-20 years).
These patterns broadly support my contention of life history variation across habitats. We need to know more, however, of the rain forest elephant both in Africa and in Asia. Most biologists study elephants in dry habitats, in which they are plentiful and far more visible. Study of the rain forest elephant is a much more challenging task. The ongoing studies in Central African rain forests may perhaps provide some of the much-needed answers.
Was this article helpful?