New evidence on the origin of carnivorous plants.

Carnivorous plants have fascinated scientists and the general public since the pioneering studies of Charles Darwin (1). No doubt part of their wide appeal is that carnivorous plants have turned the evolutionary tables on animals, consuming them as prey, with the green predators often equipped with remarkable lures, traps, stomachs, and—in a few cases—extraordinary speed of movement. To be considered carnivorous, a plant must be able to absorb nutrients from dead bodies adjacent to its surfaces, obtain some advantage in growth or reproduction, and have unequivocal adaptations for active prey attraction, capture, and digestion (2, 3). Some carnivorous species [e.g., Pinguicula (butterworts), Philcoxia] lack obvious attractants; some rely on passive pitfalls [e.g., Cephalotus (Australian pitcher plant), Sarracenia (American pitcher plants)] rather than active traps based on sticky tentacles [e.g., Byblis, Drosera (sundews)] or snap traps [e.g., Dionaea (Venus fly-trap), Utricularia (bladderworts)]; and some lack digestive enzymes and instead depend on commensal microbes or insect larvae to break down prey (e.g., Brocchinia, Darlingtonia, some species of Sarracenia). Based on these criteria, today we recognize at least 583 species of carnivorous plants in 20 genera, 12 families, and 5 orders of flowering plants (Table 1). Based on DNA sequence phylogenies, these species represent at least nine independent origins of the carnivorous habit per se, and at least six independent origins of pitfall traps, five of sticky traps, two of snap traps, and one of lobster-pot traps. To the extent to which molecular phylogenies have been calibrated against the ages of fossils of other plants, these origins of carnivory appear to have occurred between roughly 8 and 72 million years ago (Mya). In PNAS, Sadowski et al. (4) contribute to our understanding of the origins of plant carnivory by describing the first fossilized trap of a carnivorous plant, a fragment of a tentacled leaf preserved in Baltic amber from 35 to 47 Mya, and allied to modern-day Roridula of monogeneric Roridulaceae (Ericales) from South Africa. As with most carnivorous plants, the two living species of Roridula today grow on open, extremely infertile, moist sites. The occurrence of carnivorous plants on nutrientpoor substrates has been understood since Darwin showed that such plants augment their supply of mineral nutrients through prey capture. The restriction of carnivorous plants to open, infertile, moist sites, however, remained unexplained until modern costbenefit models showed that carnivores are likely to obtain an advantage in growth relative to noncarnivores only on such sites, where nutrients and nutrients alone limit plant growth, and where carnivory can accelerate photosynthesis and the conversion of photosynthate to new leaf tissue while decreasing allocation to root tissue (2, 3, 5, 6). Wet soils and fire can favor carnivorous plants, by making N more limiting for growth while making light and water less limiting (3). The wet, sandy, fireswept sites in fynbos occupied by Roridula (6) should thus favor carnivory, and indeedRoridula often grows in association with large numbers of carnivorous sundews. Roridula, however, is in other respects highly unusual as a carnivorous plant. Although its glistening, glandular tentacles do trap large numbers of insects, the secretions are resinous rather than aqueous, and so cannot support the activities of digestive enzymes. It does not secrete proteolytic enzymes; several authors thus argued that Roridula could not be carnivorous because it could not digest prey or absorb the minerals released (7, 8). The resinous nature of Roridula secretions may be an adaptation to the summer drought in the Mediterranean climate it now occupies, in that they do not lose volume or stickiness during long periods of drought; the secretions also do not dissolve during winter rains (9). It turns out that certain hemipterans (Pameridea) are capable of negotiating the glandular leaves of Roridula without becoming entangled; they eat the prey immobilized by the plant, and then N from their excretions is absorbed by Roridula (Fig. 1). This process substantially augments the N supply to the plants, with the plants obtaining 70% or more of their nitrogen supply in this fashion (7, 10). The mutualism appears stabilized by nonlinear interactions: excess densities of Pameridea turn counterproductive as the bugs switch to sap-sucking in the absence of prey, leading to negative impacts on Roridula and, ultimately, on the bugs themselves (11). Table 1. Currently recognized groups of carnivorous plants