Volume 217, Issue 3 p. 1035-1041
Tansley insight
Free Access

Convergent and divergent evolution in carnivorous pitcher plant traps

Chris J. Thorogood

Corresponding Author

Chris J. Thorogood

Botanic Garden, University of Oxford, Rose Lane, Oxford, OX1 4AZ UK

Department of Plant Sciences, University of Oxford, South Parks Road, Oxford, OX1 3RB UK

Author for correspondence:

Chris J. Thorogood

Tel: +44 01865 610301

Email: [email protected]

Search for more papers by this author
Ulrike Bauer

Ulrike Bauer

School of Biological Sciences, University of Bristol, 24 Tyndall Avenue, Bristol, BS8 1TQ UK

Search for more papers by this author
Simon J. Hiscock

Simon J. Hiscock

Botanic Garden, University of Oxford, Rose Lane, Oxford, OX1 4AZ UK

Department of Plant Sciences, University of Oxford, South Parks Road, Oxford, OX1 3RB UK

Search for more papers by this author
First published: 13 November 2017
Citations: 37


Summary 1035
I. Introduction 1035
II. Evolution of the pitcher 1036
III. Convergent evolution 1036
IV. Divergent evolution 1038
V. Adaptive radiation and speciation 1040
VI. Conclusions and perspectives 1040
Acknowledgements 1040
References 1040


The pitcher trap is a striking example of convergent evolution across unrelated carnivorous plant lineages. Convergent traits that have evolved across pitcher plant lineages are essential for trap function, suggesting that key selective pressures are in action. Recent studies have also revealed patterns of divergent evolution in functional pitcher morphology within genera. Adaptations to differences in local prey assemblages may drive such divergence and, ultimately, speciation. Here, we review recent research on convergent and divergent evolution in pitcher plant traps, with a focus on the genus Nepenthes, which we propose as a new model for research into adaptive radiation and speciation.

I. Introduction

Carnivorous plants have inspired generations of scientists since Darwin's early experiments (Darwin, 1875). Strikingly similar trap morphologies, attractive features and trapping mechanisms have evolved convergently across unrelated carnivorous plant lineages, which caused confusion amongst early taxonomists (Ellison & Gotelli, 2001). A prime example is the pitcher trap which has evolved independently at least six times (Givnish, 2015). Here, we focus on the three best known pitcher plant genera: the palaeotropical Nepenthes (c. 150 species), the neotropical Sarracenia, Darlingtonia and Heliamphora (c. 30 species) and the monotypic Australian Cephalotus follicularis (Juniper et al., 1989). Pitfall traps have evolved a further three times independently in the Bromeliaceae and Eriocaulaceae in the form of tightly bound leaf ‘tanks’ that capture water and nutrients (Schulte et al., 2009; Givnish, 2015).

Nepenthes, Cephalotus and the Sarraceniaceae show a remarkable convergence in trap morphology (Fig. 1). All three groups produce highly specialized leaf-derived pitfall traps to attract, capture, retain, kill and digest animal prey. Nectar is widely used as a primary attractant, often aided by volatiles and contrasting colour patterns in the visible or UV spectrum. Specialized slippery surfaces, often with strikingly similar micromorphology, lead arthropods to slip and fall into a pool of digestive liquid at the base of the pitcher. Notwithstanding this, recent research highlights multiple examples of divergent trap morphology, especially in the genus Nepenthes (Moran & Clarke, 2010; Bauer et al., 2012). This diversity of pitcher morphology is mirrored by a diversification in utilized nutrient sources which may include arthropods, other animal prey, faeces and organic matter, such as leaf litter. Adaptation to any of these nutrient sources may lead to divergent selective pressures which drive a diversification in pitcher morphology (Pavlovic, 2012). Here, we highlight recent research into the convergent and divergent evolution in trap morphology and nutrient acquisition strategies across the pitfall-trapping carnivorous plants, with a focus on the genus Nepenthes, which has received significant attention in recent years.

Details are in the caption following the image
Pitcher plants show a remarkable convergence of traits associated with carnivory. (a) Nepenthes pitchers comprise a pitcher chamber containing digestive fluid, a rim (peristome) and roof-like lid. The pitcher lid shows interspecific variation, but is believed to have an attractive function and, in some species, to prevent dilution of the digestive fluid with rainwater. Attractive nectar glands are numerous on the underside of the pitcher lid and along the inner edge of the peristome, which also plays a key role in prey capture. (b) Cephalotus pitchers resemble those of Nepenthes in their ovoid shape, ridged peristome and lid. The trapping surfaces feature parallel ridges on the peristome teeth and downward-pointing imbricate cells in the funnel-shaped region. (c) Sarracenia pitchers have slippery surfaces that resemble those of Nepenthes, in which contact area-reducing microtopography and directional features contribute to the anti-adhesive properties of the trap. L, lid; p, peristome; PC, pitcher chamber; T, tendril. Images not to scale.

II. Evolution of the pitcher

Many of the structures and compounds used by carnivorous plants to trap and digest prey are shared with non-carnivorous plants in association with defence against pathogens (Renner & Specht, 2013). Phylogenetic data indicate that Nepenthes evolved from a Drosera-like progenitor (Meimberg et al., 2000; Thorogood, 2010). Fusion of the leaf margins led to the tubular, so-called ‘epiascidiate’, leaf in which the upper leaf surface became the inner surface of the tube. Spontaneous mutations can lead to epiascidiate leaves in non-carnivorous plants, such as Codiaeum variegatum (family Euphorbiaceae) (Juniper et al., 1989; Fig. 2a). Heubl et al. (2006) suggested that the spontaneous occurrence of tubular leaves could have conveyed a selective advantage by facilitating water and nutrient storage, analogous to the evolution of tight leaf rosettes in bromeliads (Givnish, 2015). Digestive glands secrete digestive enzymes and absorb nutrients in pitcher plants (Juniper et al., 1989). Glands are anatomically similar across carnivorous sundews Drosera (Fig. 2b) and Drosophyllum (Fig. 2c), Nepenthes pitcher plants (Fig. 2d) and their non-carnivorous relatives, such as Plumbago (Fig. 2e). The glue secreted by Drosera glands contains viscoelastic polysaccharides, as does the pitcher fluid of many Nepenthes species, corroborating a plesiomorphic origin of this characteristic (Bonhomme et al., 2011).

Details are in the caption following the image
(a) Mutant leaf of Codiaeum variegatum showing extension of the midrib from the leaf blade, homologous to the Nepenthes tendril (white arrow), and epiascidiate leaf (black arrow); scanning electron microscope (SEM) images of stalked digestive glands of (b) Drosera and (c) Drosophyllum, which secrete digestive enzymes, absorb nutrients and trap prey via a localized tentacle-bending reaction which, together with sticky mucilage, aids retention. Glands homologous with these are embedded on the inner wall of Nepenthes pitchers, such as (d) N. inermis; calyx glands of non-carnivorous Plumbago (e) are anatomically similar to the mucilage glands of Drosera and Drosophyllum pointing to a common ancestral gland structure; (f) pitchers of N. albomarginata showing the white band of lichen-mimicking tissue (arrow) which attracts termites; (g) the closely packed pitchers of N. ampullaria on the forest floor; (h) tree shrew (Tupaia montana) feeding on a pitcher lid of N. lowi; (i) pitcher of N. rajah showing the reflexed lid (arrow); (j) T. montana faeces inside a pitcher of N. macrophylla (arrow); (k) roosting bat (Kerivoula hardwickii) inside a pitcher of N. hemsleyana; (l) pitcher of N. bicalcarata with nectar-producing thorns (arrow) on which mutualistic ants (Colobopsis schmitzi) (inset) feed.

Pitcher evolution in the Sarraceniaceae remains largely unstudied, but recent work has established that tissue-specific changes in the orientation of cell divisions establish the pitcher shape in Sarracenia purpurea (Fukushima et al., 2015). Cephalotus produces pitchers and non-carnivorous flat leaves in synchrony. Recently, Fukushima et al. (2017) have sequenced the transcriptome of both leaf types and have shown that the genes involved in adaxial–abaxial polarity are upregulated in pitcher-bearing shoots. The authors also identified an upregulation of genes involved in sucrose, wax and cutin biosynthesis. The co-existence of carnivorous and non-carnivorous leaves on the same plant in Cephalotus promises to yield more mechanistic insights into the evolution of the pitcher plant trap (Fukushima et al., 2017).

III. Convergent evolution

Convergent traits associated with carnivory can be seen in both trap morphology and physiology. In Nepenthes, a hierarchical microtopography of the rim (peristome) renders its surface slippery when wet, causing prey to ‘aquaplane’ on a lubricating fluid film (Bohn & Federle, 2004; Bauer et al., 2008). In many species, epicuticular wax crystals reduce the contact area and performance of the adhesive pads of insects (Gaume et al., 2004; Scholz et al., 2010). Directional features on the peristome and on the inner pitcher wall provide a grip for insects entering the pitcher, but do not facilitate their exit. Finally, the digestive fluid of many Nepenthes species is viscoelastic, facilitating the retention of arthropod prey (Gaume & Forterre, 2007).

Pitchers in the Sarraceniaceae also produce nectar to attract insects, and have slippery surfaces for prey capture and digestive fluids. The functional morphology of the slippery surfaces resembles that of Nepenthes, with contact area-reducing microtopography and directional features contributing to the anti-adhesive properties. Inward-pointing trichomes and imbricate cells are common across the family, whereas epicuticular wax layers have only been observed in a few species. Remarkably, the hairy inner pitcher wall of Heliamphora nutans is wettable and traps insects via an ‘aquaplaning’ mechanism, similar to that of the Nepenthes peristome (Bauer et al., 2013). Interestingly, the trichomes have a parallel ridge structure, resembling the microscopic peristome ridges of Nepenthes, which may be crucial for the wettability of both surfaces (D. Labonte, unpublished data). The evolution of two morphologically different superhydrophilic trap surfaces in unrelated plant orders is a striking example of functional convergence, which suggests that biomechanical constraints may have driven the evolution of trap microstructures. Studies on the trapping mechanism of Cephalotus are scarce, but the similarities in functional morphology suggest that it may be similar to that of Nepenthes.

Strong patterns of convergence are also emerging for both attractive and digestive syndromes. Transparent fenestrations are another striking convergent feature in the attractive syndromes of some Sarraceniaceae (Fig. 3a), Nepenthes (Fig. 3b) and the pitcher lid of Cephalotus (Fig. 3c). Recent work, however, indicates that fenestrations in Saraccenia minor play a more important role in prey attraction (Schaefer & Ruxton, 2014). Studies of the digestive syndromes of Nepenthes, Cephalotus, Drosera and Dionaea suggest that orthologous pathogen defence proteins have been repeatedly co-opted for digestion and as antimicrobial agents in the digestive fluid of carnivorous plants (Bemm et al., 2016). Fukushima et al. (2017) have shown recently that Arabidopsis genes related to the genes coding for digestive fluid proteins in carnivorous plants are upregulated under biotic and abiotic stresses. The authors suggest that the co-option of stress response proteins may be a widespread pattern in the evolution of carnivorous plant enzymes. Functional convergence is also observed amongst the predators, filter feeders and detritivores that inhabit digestive pitcher fluids (Bittleston et al., 2016).

Details are in the caption following the image
Striking examples of convergence in morphological adaptations to the pitfall trap include (1) domed pitchers with fenestrations which operate as light traps in which ‘false exits’ disorient flying prey in (a) Sarracenia psittacina, (b) Nepenthes aristolochioides and (c) the lid of Cephalotus follicularis, as well as (2) the remarkably similar ridged peristome structures of (c) Cfollicularis and (d) Nepenthes villosa.

IV. Divergent evolution

Morphological trap characteristics, such as size, shape, peristome geometry, and the presence and location of wax crystal layers, vary in Nepenthes (Bauer et al., 2012). Physiological traits such as attractive volatiles, nectar and pitcher fluid composition are also variable across the genus. Recent studies have shown that this diversity is mirrored by a range of nutrient acquisition strategies linked to habitat characteristics (Pavlovic, 2012). Nepenthes spp. occur from sea level to 3000 m elevation across the palaeotropics, and prey availability varies between habitats. Ants comprise the bulk of prey in many lowland species (Moran et al., 2001), whereas flying insects predominate in montane environments. Viscoelastic pitcher fluids are more effective for the retention of flying prey (Di Giusto et al., 2008) and are more common in montane species (Bonhomme et al., 2011).

Epicuticular wax crystals are effective for the retention of ants, but are absent or reduced in approximately one-third of examined species. Bauer et al. (2012) performed a comparative study of trap morphology which identified wax crystal presence as the ancestral state and showed that wax layers have been lost several times independently. Pitchers without wax crystals frequently have larger and more inward-sloping peristomes than those with wax blooms on the inner wall. Bonhomme et al. (2011) demonstrated that wax loss is associated with montane habitats and with the occurrence of viscoelastic fluids. Moran et al. (2013) linked divergent trapping syndromes to climate, in which viscoelastic fluids are common in perhumid (ever-wet) regions and epicuticular wax crystals are common in both perhumid and seasonal areas (Moran et al., 2013). Climatic conditions as well as the faunal composition of the plants’ habitats may exert selective pressures that favour one trapping strategy at the expense of another.

Carnivorous plants typically rely on prey for optimal growth and reproduction (Moran & Moran, 1998), and strong selection pressures should act on traps to maximize their prey intake. Modifications in trap geometry may enable the utilization of novel nutrient sources, analogous to well-known examples in animals, such as the diverse beak shapes of Darwin's finches and the various adaptations of cichlid fish in the African Great Lakes. For example, Bauer et al. (2015a) demonstrated that the pitcher lid of N. gracilis is adapted to exploit the impact of rain drops for capturing insect prey. The lid of this species functions as a rain-driven torsion spring, flicking insects into the pitcher during heavy rain. The pitchers of N. albomarginata (Figs 2f; 4a) produce a white ring of lichen-mimicking tissue that is specifically attractive to termites (Hospitalitermes), which account for > 50% of its foliar nitrogen (N) (Moran et al., 2001). Nepenthes ampullaria (Fig. 2g) grows in closed-canopy forests and derives up to 40% of its foliar N from leaf litter (Moran et al., 2003; Pavlovic, 2012). The formation of dense pitcher ‘carpets’ with reflexed lids maximizes the intake of leaf litter fall (Moran et al., 2003).

Details are in the caption following the image
Divergence in the carnivorous syndrome in (a–f) Nepenthes and (g–j) Sarraceniaceae. Several species of Nepenthes have diverged (at least partially) from the carnivorous syndrome and coevolved mutualistic relationships with animal partners: (a) N. albomarginata has a white band of hairs which attracts termites (Hospitalitermes bicolor); (b) N. bicalcarata produces swollen tendrils in which mutualistic ants (Camponotus schmitzi) rear broods; (c) N. hemsleyana has a modified pitcher shape and fluid level to facilitate bat roosting (Lim et al., 2015) and an elongated inner pitcher wall to reflect the ultrasound calls of bats, enabling them to locate and identify plants in dense vegetation (Schöner et al., 2015); montane species which associate mainly with tree shrews (Tupaia montana) include (d) N. macrophylla, (e) N. lowii and (f) N. rajah; examples of divergence of pitcher morphologies across the Sarraceniaceae lineage include hooded pitchers with fenestrations of (g) Sarracenia minor and (h) Darlingonia californica, as well as the open pitfall traps of (i) Sarracenia and (j) Heliamphora. Images not to scale.

Four Bornean species (Figs 2h–k, 4c–f) produce pitchers which trap the faeces of mammals (Chin et al., 2010), which can contribute 57–100% of their foliar N (Clarke et al., 2009). Chin et al. (2010) demonstrated that the size and geometry of the pitcher orifices in N. lowii, N. rajah and N. macrophylla are closely correlated with the body size of tree shrews. All three species have pitchers with large orifices and concave, reflexed lids oriented to optimally position the animal for faeces capture. Nepenthes rajah and N. macrophylla are closely related to each other, but not to N. lowii, according to phylogenetic data (Meimberg et al., 2001), suggesting that tree shrew associations evolved independently at least twice. Records of birds and tree shrews feeding on the nectar of Nepenthes that are not associated with mammals (Bauer et al., 2016) suggest that pitchers may be preadapted to attract vertebrates. Broad pitcher orifices could initially have been an adaptation for enhanced water or leaf litter capture. Faeces capture may have conferred a strong selective advantage, facilitating and consolidating further changes in pitcher geometry. Nepenthes hemsleyana (Figs 2k, 4c) produces slender pitchers which capture few insects, but provide a roosting site for bats (Kerivoula hardwickii), which provide the plants with a third of their foliar N (Grafe et al., 2011). The pale-coloured tubular pitchers resemble whitish, suspended, tubular flowers that are typically bat-pollinated. Bat-adapted pitcher morphology may be analogous to shifts in floral spur length as an adaptation to pollinators with greater tongue length (Whittall & Hodges, 2007).

Nepenthes bicalcarata (Figs 2l, 4b) has evolved a unique mutualistic association with ants (Colobopsis schmitzi), which defend the plant against herbivores in exchange for nectar and nesting sites (Merbach et al., 2007; Bonhomme et al., 2010), similar to other myrmecophytes. Analogies can be drawn with sea anemones providing anemonefish with shelter from predators in return for defence against cnidarian predators (Godwin & Fautin, 1992). Nepenthes bicalcarata may benefit from colony waste for nutrition (Bazile et al., 2012), the prevention of nutrient export by pitcher-dwelling Diptera (Scharmann et al., 2013) and the maintenance of the slipperiness of the trap by cleaning the peristome (Thornham et al., 2012).

Divergent trap forms have also evolved across the Sarraceniaceae and may be linked to the ecological niche in Sarracenia. Most species produce trumpet-shaped traps (Fig. 4i), but those of S. minor (Fig. 4g) and S. psittacina (Fig. 3a) are hooded with concealed entrances. Forms of Sarracenia minor show specificity towards either flying insects (tall pitchers) or ground-dwelling prey (shorter pitchers with dense trichomes) depending on habitat differences (Stephens et al., 2015).

V. Adaptive radiation and speciation

Adaptive radiations are characterized by rapid speciation, recent common ancestry and a strong link between phenotype and environment (Schluter, 2000). The genus Nepenthes meets these criteria because it has undergone rapid speciation (Meimberg & Heubl, 2006), and functional pitcher morphology is linked to prey availability and climate (Clarke & Moran, 2016). Many Nepenthes species have overlapping distributions and produce hybrids. Nepenthes rajah and N. villosa, both confined to the Mount Kinabalu National Park in Borneo, overlap and hybridize. Nepenthes rajah is established to capture faeces, but nothing is known about the nutrient sources of N. villosa or the hybrid. The more tubular pitcher, smaller orifice and horizontal lid suggest that faecal capture is unlikely in N. villosa. Divergence, despite ongoing hybridization at intermediate altitudes, is also seen in ragworts on Mount Etna (Senecio, Asteraceae) (Filatov et al., 2016), where strong ecological and phenotypic differences between the high- and low-altitude species evolved as a result of strong selection at just a few key genes (Chapman et al., 2016). Selection at a few key genes may also feature in Nepenthes, but this is yet to be explored.

Co-occurring Nepenthes species are established to capture different combinations of prey in a given location (Chin et al., 2014). Hybridization between co-occurring Nepenthes species is widespread (M. Scharmann et al., unpublished data); therefore, if hybrids arise from parents with different pitcher morphologies, novel pitcher traits may arise and facilitate the utilization of novel nutrient sources. On the other hand, Pavlovic (2012) hypothesized that intermediate hybrid phenotypes of insect-trapping N. rafflesiana and detritivorous N. ampullaria may perform poorly and be outcompeted. Indeed, preliminary data suggest that the quantity and variety of prey caught by lowland hybrids may be less than that of their parent taxa in some cases. This could be caused by a loss of specialized structures in hybrids of parent species that target specific types of prey (Peng & Clarke, 2015). Further work should explore this hypothesis by investigating the fitness of hybrids of species with divergent nutrient acquisition strategies, for example N. hemsleyana and N. rafflesiana which rely on mutualistic bats (Grafe et al., 2011) and insects (Bauer et al., 2015b), respectively. Nepenthes rafflesiana, N. hemsleyana, N. ampullaria, N. bicalcarata and N. albomarginata all have overlapping distributional ranges and have evolved distinct specializations linked to nutrient sources. Given that sympatric Nepenthes species appear to target different prey taxa (Chin et al., 2014), interspecific competition for prey may have been a driver of disruptive selection and speciation in the genus (Thorogood, 2010).

VI. Conclusions and perspectives

Pitcher plants are one of the most striking examples of convergent evolution in the plant kingdom. Repeated patterns of adaptation for prey capture across unrelated groups include similar surface micromorphology and gland anatomy. A recent surge in research is now providing growing evidence that adaptive radiations may be driven by nutrient competition. New Nepenthes species are described every year and little or nothing is known about their trapping strategies. Mammalian faeces capture in pitcher plants was discovered only recently from extensive field observations on well-known species in North Borneo. Ecological studies are now required across the genus, including the many poorly known species described recently from the Philippines. This will build a platform for the exploration of adaptive radiation, for example through the examination of competitive exclusion and niche partitioning in situ. The increasing availability of genomic data also offers new and exciting opportunities for the study of the molecular basis of carnivorous plant evolution, in particular, the quantification of genetic divergence and positive selection in relation to phenotypic divergence. A combined approach exploring these ecological and molecular aspects will greatly enhance our understanding of adaptive radiations in carnivorous pitcher plants in the future.


We thank Hugh Dickinson for the SEM preparations, Kate Pritchard for supplying plant material and Bruno Nevado for constructive comments on the manuscript.