Speciation in the Cape

History

The Pliocene epoch, starting around 6 Mya, saw the cooling and aridification of the Cape region, as well as the advent of the winter rainfall regime.¹ This is thought to have led to a change in vegetation from subtropical forest to the fire-adapted sclerophyllous shrub- and heathlands and drought-adapted succulent shrublands characteristic of the Cape flora.²

During the subsequent Pleistocene epoch, the Cape experienced a stable, moderate climate. The absence of catastrophic climatic and geological events may have led to relatively low extinction rates, partially accounting for the high plant diversity in the region compared to the Northern Hemisphere, where glaciation, for example, caused the removal of many taxa from the local flora.^1,3

In the formerly dominant view that speciation occurs mostly by vicariant allopatry, the Cape flora was thought to consist of three components: Antarctic, African, and Eurasian. However, Linder 2005 has inferred several additional introductions of clades by long-distance dispersal since the break-up of Gondwana.⁴ Thus, the onset of radiation of different groups varies from 35 Mya, through the Miocene, to the Pleistocene, and there was no single environmental trigger for the massive speciation that has occurred. He attributes current species richness to the diversity of lineages recruited into the flora and their subsequent evolution of key innovations adapting them to the heterogeneous environment (see below).⁴

Abiotic factors

Topography

The CFR is a topographically diverse area, being dissected by a series of parallel mountain ranges — the Cape Fold Mountains. The landscape thus consists of steep altitudinal gradients, with change in aspect over short distances. This also causes the terrain to be fractured by temperature and rainfall differences.¹ It has been suggested that this environmental heterogeneity accounts for the higher species richness in the Cape, compared to southwestern Australia, which has also had a stable geological and climatic history, but which is more topographically uniform.^2,4

A heterogeneous landscape offers a variety of different ecological niches in close proximity.³ Evolutionary divergence is predicted to take place, as it is relatively easy for propagules to be spread to new patches, but establishment requires adaptation to the different conditions. Disruptive selection would result in genetic and morphological differences accruing in adjacent subpopulations as individuals better adapted to the different physical environments would have greater reproductive success on average.

These differences may result in reproductive isolation in various scenarios. For example, water, light, or nutrient availability may lead to non-overlapping flowering seasons; gametes may be incompatible for biochemical reasons, like preferred pH, or reinforcement mechanisms like sperm competition may be at play; hybrids resulting from outcrossing may be inviable due to having intermediate characters that make them ill-adapted to either parental habitat, or they may be sterile; epistatic interactions in the new genetic combination may also cause death, developmental problems, or reproductive incompatibility with individuals from the parental populations.⁵

In a highly fractured environment, the constituent patches will be relatively small, and able to support lower numbers of plants. Levin 1988 argues that small subpopulations respond more rapidly to disruptive selection.⁶ If patches of similar habitat are discontinuous, there will be little gene flow between them, as populations are not adapted to colonize the intervening habitats, and even in the absence of selection, divergence will take place due to the founder effect.⁶ This is stronger in smaller subpopulations, as a small number of individuals will tend to be less genetically representative of the ancestral gene pool.⁵

Soil

Various soil types of different geological origins are found in the CFR.¹ Mountains typically expose nutrient-poor sandstone, while richer clayey soils are found in valleys,³ thus forming part of the environmental heterogeneity discussed above.

It is thought that this mosaic of soil types could lead to parapatric speciation as subpopulations on either side of the dividing line adapt to differences in soil chemistry and nutrient and water content.¹ Van der Niet & Johnson 2009, however, claim that soil is not diverse in the Cape,³ and it has been suggested by Cowling et al. 1996 that different edaphic habitats may, in fact, be a result of plant speciation rather than a driver.²

Nevertheless, the generally low level of nutrients in soils in the CFR has been implicated in a disturbance regime that may be of importance in understanding speciation, as explained in the following section.

Fire

One of the consequences of the shift in the Pliocene to a dry-summer climate is the establishment of a regime of frequently recurring fire. The low water retention of sandy soils and their low nutrient status increase the probability of fires sweeping through the vegetation.⁷

Fire events have been postulated to increase the rate of speciation by increasing generation turnover.^1,7 If generation times are short, combination of parental genes can occur faster, allowing individuals with new features to spread into habitats that have been cleared of competition from the parental generation.

This disturbance regime also increases patchiness in the vegetation by causing local extinctions.⁷ This promotes isolation between subpopulations and allows founders to colonize these newly opened habitats. Many equivalent shrubs and grasses can thus coexist within the landscape, as the frequent disturbance precludes domination and competitive displacement by any one (or a few) species.²

However, it was predicted by Van der Niet & Johnson 2009 that, if the physical environment was the main driver of selection, morphological adaptations to different habitats would be mainly vegetative, as opposed to floral.³ Since the Cape flora is considered florally diverse but vegetatively uniform (Johnson 1996a, quoted in Van der Niet & Johnson 2009³), biotic factors may, in fact, have greater importance.

Biotic factors

Pollination

That plant-pollinator coevolution has been significant in the Cape flora is evident from the large number of genera that show wide variation in flower morphology, with many different pollination systems often occurring in the same closely related group of species.³ Fynbos plants have become adapted to pollination by animals few of which function as pollinators outside the CFR, including rodents, sunbirds, moths and butterflies, hopliine beetles, and long-proboscid flies.¹

These pollinators show a wide range of sizes, behaviours, and nutritional requirements. By being adapted to a particular pollinator, e.g. by being large, white, near the ground, with a sweet scent, and having suitable proportions of protein and sugar, a flower has a higher probability of being visited by that animal, e.g. a nocturnal rodent, in this case. This raises the probability of the flower receiving pollen or having its pollen deposited on a similar flower of the same species.

Flowers that attract several pollinators weakly will have lower average reproductive output than ones that specialize on a single pollinator. This ensures that, if a shift in pollinator takes place within a subpopulation, it will become reproductively isolated from the original population and divergence can occur (either through selection or simply genetic drift).

Van der Niet & Johnson 2009 found that there was often a shift in flowering time between sister species if they had overlapping distribution ranges.³ In allopatric species pairs, this could be explained by ecological differences, such as peak time of pollinator activity or water and light availability.

In sympatry, however, this indicates reproductive isolation by character displacement: By being receptive to pollen at different times of the year, plants are assured of not being pollinated by members of the other species. This could occur if hybrids of the two species were less fit than offspring of conspecifics.⁵ (Since the two species are closely related and sympatric, this would most likely be due to deleterious interactions between the genes from parents of different species, rather than for ecological reasons.)

Seed dispersal

Few fynbos species have their seeds dispersed by exochory, as the vegetation is not nutritious enough to support large numbers of mammals. Bird- or bat-dispersed species are also rare, possibly due to low-nutrient soils making fleshy fruits too expensive to produce. Wind dispersal is also uncommon in the endemic groups. In other words, seeds are not dispersed over long distances.¹

Instead, most (non-serotinous) species make use of water or ants (myrmecochory), resulting in seed vagility of less than five metres.¹ Thus, different subpopulations of the same species are relatively isolated from one another, allowing divergence to take place by selection or drift, since gene flow is too limited to ensure genetic uniformity of the population.⁷

Most authors cite a variety of the hypotheses explained above when attempting to explain the plant species diversity of the CFR.

A relatively stable climate over the last 6 My has allowed species to accumulate (possibly including significant influx from other continents) without major extinction events decimating the plant communities.

The complex topography and geomorphology of the region give rise to a heterogeneous landscape containing many potential niches in close proximity, but with sufficiently steep environmental gradients (e.g. soil nutrients, rainfall, temperature) to limit gene flow between patches.

Frequent fires may contribute to this mosaic vegetation structure by causing local extinctions and opening up the habitats for genetic novelties to manifest themselves without being competitively displaced by a few dominant forms.

Finally, adaptation to different pollinators has allowed radiation in flower morphology, while the homogenizing effect of gene flow is constrained by short seed dispersal distances.