General introduction: In which I attempt to briefly put the present thesis in a historical, global, social and scientific context.

Background:

Life on earth may have started as early as 3.5 billion years ago (Brasier et al. 2002, Schopf et al. 2002, Furnes et al. 2004, Kerr 2004) and since that time many new life forms have evolved, flourished and gone extinct. Insects appeared about 300 million years ago, and the Lepidoptera (the order in which butterflies are placed) some 100 million years ago (Gaunt & Miles 2002), whilst our species, Homo sapiens sapiens, appeared only about 150,000 years ago (Shields 2000). We are a species with an intellectual capacity and cultural development that has added a new dimension to evolution since we can consciously strive to manage our own and other species to persist for many generations to come.

Man has had a dramatic influence on the ecosystems on earth and will continue to do so (Houghton et al. 2001). To which degree and how this will influence the human population itself is still unclear, but an important negative impact can be expected as ecosystems and the biodiversity therein provide invaluable services (Myers 1996, Gitay et al. 2001). One disputable point may be the intrinsic value of biodiversity: perhaps extinction is likely to be followed by a new round of diversification with evolution of life forms that would be equally valuable and interesting.

With this in mind and considering the intellectual capacities and cultural development noted in man, it may seem surprising that most people are more concerned with the short term (Lagerspetz 1999). In most cases however, poverty or suppressive regimes can easily explain this, since they form a more imminent life threat. Perhaps more surprisingly, many well-off people in developed countries also prefer a short-term approach and occasionally attempt to back this up with corrupted data presented as real science (Pimm & Harvey 2001, Wilson et al. 2001).

If natural ecosystems and biodiversity are to be maintained, serious conservation measures have to be taken. This requires awareness of the problem and funding for conservation efforts, as well as knowledge on the structure and spatial distribution of this biodiversity to allocate funding and effort efficiently (Peuhkuri & Jokinen 1999). Therefore, studies that enhance our understanding of ecosystem functioning and biodiversity that potentially lead to predictions on how ecosystems will react to expected changes or management efforts are of critical importance for conservation. However, in the current age of artificial intelligence, human genomics, perceptual robotics, flow visualisation, advanced astrophysics and other high-tech wizardry, it is both disturbing and ironic that our understanding of vital ecological processes is rudimentary (Harvey 2001).

Tropical forests

Tropical forests have been identified as major hotspots of biodiversity (Myers et al. 2000), and especially the canopy is thought to harbour a diverse and poorly known fauna and flora (Basset 2001, Mitchell 2001, Stork 2001). However, the area of tropical forest has decreased rapidly over recent decades and most forests are subject to disturbance. Undisturbed forest will soon be confined to isolated reserves surrounded by cultivated land (FAO 1999).

Most tropical forests are situated in developing countries where the awareness of the value and vulnerability of biodiversity is generally low, government budgets are limited, and human pressure on the land is high. According to the Intergovernmental Panel on Climate Change (McCarthy et al. 2001), third world countries are especially vulnerable to global change due to their poor economic development and stability, and a generally low level of knowledge. However, sacrificing the National Parks in Uganda to agriculture, for example, will only compensate for 1.5 years of population growth, whilst at the same time diminishing the possibilities for eco-tourism, not to mention the loss of ecosystem services.

The governments of most countries are to some degree aware of the value of wilderness areas and do protect them. However, such protected areas cost more than the income they generate and aid from western donors who appreciate their more global value will be necessary (Balmford et al. 2000, Balmford et al. 2002, Balmford et al. 2003, Balmford & Whitten 2003, Williams et al. 2003). Law enforcement is of critical importance to the effectiveness of protective measures, even though some local use may well be sustainable (Gordon & Ayiemba 2003). Sadly, political instability and corruption can easily lead to periods of anarchy in which poaching and habitat degradation flourish (Draulans & Van Krunkelsven 2002). So, conservation of biodiversity can not be viewed independently of global policies that affect peace, good governance and development (Kahn & McDonald 1995, Pimentel et al. 1997, Avery 1998, Haila 1999, Swanson 1999, Kremen et al. 2000, Campbell & Vainio-Mattila 2003).

Sustainable development is a key concept in this discussion. This is defined as: “development that meets the needs of the present without compromising the ability of future generations to meet their own needs”, noting that: “even the narrow notion of physical sustainability implies a concern for social equity between generations, a concern that must logically be extended to equity within each generation” (Brundtland Commission; WCED, 1987). The goal of sustainable development is a stable human environmental system in which available resources are sufficient to meet the needs of society in perpetuity. Questions have been asked about whether “needs,” as conceived in the Brundtland Commission report, should be limited to basic necessities of food, clothing, shelter, and health or should include more qualitative aspects such as comfort, convenience, or other “quality of life” measures. There is no consensus in the literature regarding what constitutes the limits of “needs” in this context.

Extinction and rarity:

Insights from both evolutionary biology and ecology are necessary to understand biodiversity. The number of species in the world is the outcome of both speciation and extinction. Speciation is mostly viewed as an evolutionary process which takes place in an ecological arena, whilst extinction is usually viewed from an ecological perspective where evolved traits clearly play an important role.

Palaeontology has demonstrated that diversification and extinction rates have varied considerably through time, with mass extinction’s followed by diversification. The study of biogeography has made major contributions to our appreciation of how migration between continents and from continents to islands shaped biodiversity on a taxonomic level. To explain biodiversity on a local scale, immigration can be added to the key-factors of speciation and (local) extinction. One successful approach in describing the outcome of immigration and local extinction is found in island biology. The distance of an ‘island’ to a source area can largely explain immigration rate whilst the size of the island can predict the extinction rate. These ‘islands’ can be habitats of different types and scales, including forest patches and lakes.

Extinction rate can be viewed as a function of size and stability (amount of fluctuation) of populations (Hanski 2003). The size of a population is dependent on the area of the habitat, and also on the population density, which again depends on the ecology and life history of the species. The stability of a population can depend on a host of biotic and abiotic factors, and also on the life history. Moreover, genetic diversity plays a role in the capacity of a species to adapt to a changing environment and genetic health in general (Frankham et al. 2004).

Natural selection results in adaptation of a species to the environment. However, when the environment changes, the amount of pre-adaptation of the present species to this new environment can depend largely on chance. Generalist species are more likely to be pre-adapted to a new environment than are specialists. Therefore, from a long-term evolutionary perspective, specialisation can be viewed as a dead end, even though selection can temporarily favour specialisation in a particular location.

From a conservationist’s perspective, it is important to know why certain species have low abundances and limited geographical ranges, and thus why they have a more threatened status. Surprisingly few studies address this question systematically. These studies include those in mosses (Cleavitt 2002, Heinlen & Vitt 2003), higher plants (Bevill & Louda 1999, Hegde & Ellstrand 1999, Gitzendanner & Soltis 2000, Kelly et al. 2001, Cadotte & Lovett-Doust 2002, Lloyd, Lee & Wilson 2002b, a, Rogers & Walker 2002), insects (Didham et al. 1998, Malmqvist 2000, Lewis 2001), fish (Dulvy, Sadovy & Reynolds 2003), and primates (Harcourt, Coppeto & Parks 2002). Plant species with narrow geographical distributions were found to produce significantly fewer seeds (per unit measurement) than common species (in four of six studies), but did not differ with respect to breeding system (five of five studies). The majority of traits (including seed size, competitive ability, growth form and dispersal mode) were related to rarity in different ways from one study to the next.

Studies on butterflies and aquatic insects have suggested that species with a restricted distribution tended to have limited dispersal behaviour or abilities (Hill et al. 1995, Lewis, Wilson & Harper 1998, Malmqvist 2000). However, data on leaf litter beetles showed that rarer species are predicted to be better dispersers, and more likely to persist in a given habitat. In this case, rarity and population variability (in undisturbed forests) were significant predictors of susceptibility to fragmentation. Common species were significantly more likely to become locally extinct in small fragments than rarer species, lending empirical support to models of multi-species coexistence under disturbance that suggest competitively dominant but poorly dispersing species are the first to become extinct due to habitat destruction (Didham et al. 1998). Studies on invertebrate herbivore communities in a forest in Papua New Guinea indicated that host-plant specialization did not affect rarity (Novotny & Basset 2000). In primates however, specialization was the only trait that correlated with rarity (Harcourt et al. 2002).

The highly context-dependent nature of most trait relationships with rarity implies that application of knowledge concerning rare-common differences and similarities to management plans will vary substantially for different organisms, vegetation types, and possibly among different continents. This is a serious problem for management decisions in the light of scant data.

Evolution

The process of diversification and speciation, the source of biodiversity, has received ample attention from biologists. Evolutionary biologists are currently very successful in the application of new molecular and population genetical techniques that have become available. These studies investigate how natural selection together with constraints affects adaptive evolution and speciation.

An important field within evolutionary biology is life history evolution. Among species, growth and reproduction occur at different rates with different timing, and result in, for example, differences in adult size. Central to life history theory is the distribution of energy, time and nutrients to maintenance, storage, growth and reproduction (Stearns 1992). The possible allocation of resources obviously depends on these resources themselves, and diet is, therefore, an important factor in life history. A diet shift is, therefore, likely to be accompanied by life history evolution. To elucidate such relations, one can compare different clades in which a similar diet shift has occurred. For example, it is striking that filter-feeding in marine vertebrates is found in the largest species of sharks, rays and whales. Possibly, plankton is a rich food-source that facilitated the evolution of large adult sizes. Additionally, the strength and nature of nutritional constraints within a species can be studied by manipulation of the diet and exploring plastic responses.

Fruit-feeding in butterflies provides an example of diet shifts that are accompanied by changes in nutritional ecology, since the quality and spatial and temporal availability of fruits differs from that of nectar. In this thesis, data are presented on nutritional ecology of fruit-feeding butterflies in the field (chapters 3, 4 and 6), whilst most of the life history data still await analysis. Additionally, we performed three experiments in which we manipulated the protein (chapter 5) and sodium content (chapter 7) of the adult diet of the small fruit-feeding tropical butterfly Bicyclus anynana and measured lifespan and egg-production.

Within research in evolutionary biology, sexual selection takes an important place. However, sexual differences can also have an ecological background. Sexual differences give an extra dimension to diversity and often diversify the habitat use of a species. A fine example can be found in hummingbirds where the sexes can have different bill morphologies, each adapted to a certain flower type (Temeles et al. 2000). If this species had split into two species with their own flower specialisation, both populations would presumably be of about one-half the size of the current population and thus be more prone to extinction.

Butterflies provide excellent examples of sexual dimorphism, and in the field it can be difficult to match males and females of the same species. The feeding behaviour of males and females can differ too. Generally, male butterflies are more often trapped on fruit baits, possibly because they feed more frequently (Fermon, Waltert & Muhlenberg 2003), and in many species, males are almost exclusively the only sex to puddle (= feeding on mud, dung or carrion). In this thesis, I will discuss sexual size dimorphism and mating system in relation to possible functions of puddling behaviour in a community of fruit-feeding butterflies (Chapter 6).

Ecology

Despite extreme geographical variation in species composition, some similarities in patterns of biodiversity can be found in similar habitats on different continents. This indicates that ecological processes have substantial effects on patterns in biodiversity. However, it is still poorly understood to what extent and how fundamental biotic interactions, such as competition, facilitation and trophic interactions, determine community patterns (Peres-Neto 2004). Such information is essential for understanding the effects of disturbance, and with massive anthropogenic global disturbance, highly relevant for conservation (Harvey 2001). The complexity of ecosystems is so high that science has only begun to understand how the food webs, nutrient cycles and physical traits of an ecosystem interact. Two approaches can be distinguished: investigating community patterns in time and space for complete guilds, and case studies of focal taxa. In this thesis, vertical and temporal patterns in abundance and biodiversity of fruit-feeding butterflies are described (Chapter 2). These data will later be used by Prof. Russel Lande (San Diego) and Prof. Steinar Engen (Oslo) to estimate to what degree the fruit-feeding butterfly community is structured in time and space.

The description of temporal and spatial diversity patterns of tropical Lepidoptera has now reached a stage where the search for causal explanations should be intensified. Some phenomena that have been recognised as important to understanding patterns in biodiversity will be elaborated on below in relation to this thesis.

Specialisation is expected to be accompanied by efficient resource use and high competitive abilities on the specific resource, as well as by superior defence against natural enemies. However, generalists need less time to find suitable resources and may adapt more easily to a changing environment. It has long been believed that the high diversity of insects in tropical forest is due to high floral diversity and high levels of specialisation (Owen 1966). However, recent studies have shown that catholic food choice is common amongst herbivorous insects in tropical forests, and tropical Lepidoptera are no more host-specific than temperate species (Basset & Burckhardt 1992, Fiedler 1998, Novotny et al. 2002).

In the case of butterflies, host-plant specialisation is most likely to be associated more with natural enemies than with competition, since caterpillar densities are usually low. Although it was not a prime goal of the project, we found a host-plant specialist, Gnophodes chelys (Satyrinae), that was particularly well camouflaged on the stripy leaves of its host-plant (Setaria poiretiana). I monitored the reproduction of this species and two students, Mechteld van Dijk and Peter Boons, have made more detailed surveys of parasitoid loads in eggs and larvae. The results showed that almost all eggs are laid in a four-week period at the beginning of the rainy season. By the time the parasitoids can develop a second generation, very few caterpillars are available, and these few late caterpillars are typically parasitised. These findings are in accordance with the idea that parasitoids can be a selective force in the evolution of seasonal reproduction in environments where the seasonality of plant availability is not very marked. In this butterfly species, other life history traits facilitate the evolution of seasonal reproduction, including the large egg-batches that reduce the host-plant searching time (only one host-plant is needed), and the large host-plant that allows for large egg-batches. This could potentially be an example of how life history and trophic interactions together affect the phenology of individual species and thus community dynamics. It would be interesting to know to what extent predator and parasitoid faunas’ overlap between butterfly species, plants and habitats, and thus how these natural enemies connect butterfly species ecologically.