SEARCH

SEARCH BY CITATION

Keywords:

  • carnivorous plant;
  • Nepenthes;
  • plant–insect interface;
  • slippery surfaces;
  • adhesive pads;
  • locomotion;
  • attachment;
  • SEM

Summary

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgements
  9. References
  • •  
    Several epidermal microstructures characterize surfaces of pitcher plants and are presumably involved in their trapping function. Here we report the effects of Nepenthes alata surfaces on insect locomotion and trapping efficiency.
  • •  
    The architectural designs of pitcher surfaces were characterized using scanning electron microscopy. Two insect species – fruitfly (Drosophila melanogaster) and ant (Iridomyrmex humilis) – were tested for their ability to remain and walk on them. The relative contributions of various epidermal structures to trapping ability were quantified.
  • •  
    Pitchers were very effective traps for both insect species. They were slightly more efficient in capturing the ants, but slightly more effective in retaining captured flies. Trapping efficiency was attributed to the combined effects of several surfaces displaying different functions. The waxy zone played a key role in the slippery syndrome: in addition to the wax itself, the subjacent layer of convex lunate cells interfered considerably with insect locomotion. The unsubmersed glandular zone displayed an important retentive effect and secretions of the digestive glands are suspected to be adhesive.
  • •  
    Pad performances of the hairy and smooth system of attachment are discussed to explain the differences between the two insect species. This study aims to encourage biomechanical studies of plant–insect surface mechanisms.

Introduction

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgements
  9. References

Carnivorous plants grow in mostly nutrient-poor habitats and rely greatly on insect-derived nitrogen (Juniper et al., 1989; Elison & Gotelli, 2001). This is especially the case for pitcher plants such as Nepenthes (Schultze et al., 1997). Their ecological success strongly depends on their trapping and digestive ability. Different macro-morphological features associated with the trapping function involve a range of specialized forms, including the familiar pitcher and flypaper traps, the snap-traps of Dionaea and the suction bladders of Utricularia. Several micromorphological features, including trichomes, wax and glands, that characterize the surfaces of carnivorous plants, are also involved in trapping and digestion (Darwin, 1875; Juniper et al., 1989). However, the exact functions of such microstructures and their physical interactions with insects remain largely unknown.

In pitcher plants such as Nepenthes, there is a range of surface architectures, which differ widely in terms of the geometry and surface features (Adams & Smith, 1977; Owen & Lennon, 1999). The thick layer of crystalline wax characterizing the ‘slippery’ surface of Nepenthes pitcher has received most of the attention focused on its trapping mechanism (Lloyd, 1942; Juniper & Burras, 1962; Juniper et al., 1989). However, other microstructures such as trichomes, lunate cells and digestive glands are responsible for a microrelief susceptible to interact closely with insect pads (Juniper & Southwood, 1986) and may therefore contribute to prey capture. Moreover, with one conspicuous exception (Merbach et al., 2002), prey specialization is rare in Nepenthes and captured prey is taxonomically varied (Moran, 1996; Moran et al., 1999). Since attachment pads vary among insect groups (Beutel & Gorb, 2001; Gorb, 2001), the deployment of diverse plant surfaces able to trap a wide variety of arthropods and thus maximize the number of prey captured may be a selective advantage.

The principal aim of this study was quantitatively to investigate effects of different pitcher surfaces on locomotion of insects with different attachment pads, in particular smooth vs hairy types of attachment. We addressed the following questions. (1) Which surfaces are responsible for the initial fall and retention of insects in Nepenthes? (2) How do the different pitcher surfaces influence insect attachment and locomotion?

To answer these questions, a series of laboratory experiments was designed with Nepenthes alata and two insect species (Drosophila melanogaster and Iridomyrmex humilis). These experiments were complemented by the structural study of surfaces using scanning electron microscopy (SEM).

Materials and Methods

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgements
  9. References

Organisms and surface characterization

Pitcher plants (Nepenthes alata Blanco) originating from tropical mountain forests of Indonesia, were obtained from a nursery in February 2000 and grown in controlled shaded conditions (18–24°C, 70–90% humidity) in Sphagnum moss without fertilizer.

Two insects species were chosen according to three criteria: (1) they belong to the two insect orders the most frequently captured by the plant (Juniper et al., 1989; Moran et al., 1999); (2) they possess two main types of attachment systems in arthropods (Gorb, 2001); (3) They are of similar size and weight. Insect species, selected for this study, included Iridomyrmex humilis (Formicidae, Dolichoderinae), bearing smooth arolia, and Drosophila melanogaster, bearing hairy pulvilli. As our principal aim was to test attachment abilities of insects on different surfaces, we decided to test the flightless ‘curly’ mutant of D. melanogaster, as this would ‘factor out’ the influence of flight in comparisons with the flightless ant.

Fresh pitcher surfaces were observed untreated in the scanning electron microscope. Specimens were dissected and then mounted on stubs, uncoated; they were examined for a period of 5–10 min after dissection using a JSM-6300F (JEOL, Peabody, MA, USA) scanning electron microscope at 5 kV.

Experimental design

Two experiments were conducted on 10 fully developed pitchers of approximately the same age (corresponding to the seventh leaf produced ± 1) and belonging to 10 different plants. Three repetitions were carried out for each pitcher (corresponding to a total of 30 trials for each species of insects). In the first experiment, the capture efficiency of entire pitchers was estimated for both insect species. In the second experiment, the walking speed of ants was compared on different excised surfaces of the pitchers. Statistical analyses were done using the software package sas v. 6.2 (SAS, 1996).

Experiments on entire pitchers

An insect was placed on the peristome (toothed rim of the pitcher) and observed 5 min. We recorded for each trial (1) the time the insect spent before falling down, (2) the surface from which it slipped, (3) the time it spent on each surface and (4) its success of exit if it fell in the pitcher. Thirty workers of I. humilis and 30 individuals of D. melanogaster were used, comprising three repetitions for each pitcher plant.

Multiple Fisher's exact test was used for frequency data comparisons. We performed a two-way anova to test for a plant effect (variation between individuals) or an insect effect (variation between both insect species) on the time of trapping. We also performed a three-way anova to test for an effect of the plant, the kind of surface and the insect species on the proportion of time spent by an individual on a plant surface during the experiment. Arcsin (square root) transformations were carried out to normalize the data. The anova were performed using backward selections of variables and type III sas-tests.

Experiment on excised pitcher cylinders

Pitchers were cut transversely both at the lower limit of the waxy zone, and at the lower limit of the digestive zone. Two types of cylindrical surfaces, waxy and glandular, were thus obtained. Ants were confined within the cylinders and their time of exit or the maximal height reached in 1 min if they failed to exit, were then measured. We then compared their vertical displacement and speed on different surfaces during 1 min. Flies were not used in this kind of experiment because they spent considerable periods of time grooming and it was not practical to calculate a speed of vertical displacement.

For waxy surfaces, three kinds of treatments were carried out. (1) The wax was removed by rinsing in chloroform for 5 min; this time was necessary to dissolve most of the wax without affecting significantly the plant tissue in terms of observed physical changes such as surface wrinkling. (2) The cylinders with removed wax were inverted to reveal any effect of cell geometry on ant locomotion. (3) The waxy surface was sprayed with a fine aerosol of water from 30 cm in order to test for an effect of water on ant locomotion on waxy surface.

For glandular surfaces, (1) the cylinders were tested fresh and dried for 10 min after dissection in order to reveal a possible effect of the glandular secretions on ant locomotion and (2) the dried cylinders were tested in normal and inverted positions.

The same ant was used for different treatments of each cylinder so that paired comparisons could be made. Before and between different treatments, ants were forced to walk on a piece of cartridge paper for 5 min to ensure that their pads were cleaned of residual material. This period appeared to be sufficient for the ants to stop grooming themselves and recover their walking ability. Thirty ants were tested on a total of 10 cylinders from 10 plants. The same cylinder was used for three different ants, which were tested in rapid succession. The data were analysed statistically using nonparametric tests for paired samples (Wilcoxon's signed-ranks test). This paired design allowed us to factor out natural uncontrolled variation between velocities of individual ants or biomechanical properties between pitchers.

Results

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgements
  9. References

Microtopography of pitcher surfaces

Figure 1(a) shows the different pitcher surfaces studied. In our N. alata sample (n = 22), the mean pitcher length was 68.9 mm ± 14.7. The waxy zone constituted about half the pitcher height (51.3 ± 2.8%), the transitional zone 5.1 ± 0.9% and the digestive zone comprised the remaining part. The lid is approximated as a disc with mean diameter of 17.4 ± 3.7 mm and the peristome was 3.3 ± 0.7 mm thick.

image

Figure 1. Surfaces of Nepenthes alata pitcher. (a) Longitudinal section of the pitcher showing the different zones studied. (b–l) scanning electron microscopy photomicrographs. (b) Upper face of the lid: u. tr., unbranched trichome; f. s. tr., flattened stellate trichome; st., stoma. (c) Lower face of the lid; s. tr., stellate trichome; nec., nectary. (d) Peristome with downward-pointing teeth above the waxy zone. (e) Striae of the peristome, which correspond to lines of overlapping elongated cells, forming series of steps towards the trap. (f) End of a tooth forming a system of overlapping protuberances. (g) Waxy zone with downward-directed lunate cells. (h) Lunate cell covered with wax crystals. (i) Limit of the waxy zone: w. cr., wax crystals on the transitional zone that have detached from waxy zone. (j) Smooth transitional zone between waxy zone (upper) and digestive zone (lower). (k) Glandular surface of digestive zone. (l) Multi-cellular digestive gland.

Download figure to PowerPoint

The upper face of the lid is characterized by an epidermis of flattened hexagonal cells with scattered upward-pointing unbranched trichomes, smaller flattened and stellate trichomes and sparsely distributed stomata (Fig. 1b). The inner face of the lid is characterized by an epidermis of domed cells creating a microrelief of very smooth ridges and depressions (Fig. 1c). On this surface, flattened and stellate trichomes are more abundant than unbranched ones. Stomata and extra-floral nectary glands are also visible, and nectaries are particularly concentrated in the bottleneck just above the pitcher. As in Nepenthes rafflesiana (Adams & Smith, 1977), they are sunk below the level of the epidermis (Fig. 1c, inset).

The peristome, although radially ridged, has a smooth and glassy appearance. A first level of striae or ridges traverses the peristome terminating in teeth whose points overhang the trap, and point towards the conductive zone below (Fig. 1d). Nectary glands are located in recessed areas between teeth. A second series of striae is visible along each tooth, consisting of lines of overlapping elongated cells forming a series of steps towards the trap (Fig. 1e). The end of each tooth (Fig. 1f) is composed of rounded points, forming a system of overlapping protuberances, each corresponding to the end of a stria of elongated cells.

The waxy zone has been described in detail previously (Juniper & Burras, 1962; Owen & Lennon, 1999). It is composed of an epidermis of flattened hexagonal cells scattered with modified stomata forming downward directed lunate cells (Fig. 1g). Each lunate cell corresponds to a single, enlarged and overlapping guard cell, forming a crescent outline with a convex shaped surface (Fig. 1g). The lower margin is slightly raised and overhangs the pitcher surface (Fig. 1h). The epidermis is covered by wax crystals (Fig. 1h–i), which are more densely distributed just beneath the peristome.

A smooth transitional zone 1.5–5 mm wide separates the waxy zone from the digestive zone (Fig. 1j). This zone has not been described previously. Some isolated scale-like wax crystals were located in the upper part of this zone (Fig. 1i).

The digestive zone bears glands secreting a viscous substance that can be seen in the SEM microphotograph made on fresh material (Fig. 1k). Each gland is multicellular (Fig. 1L) and surrounded by a hood, forming a downward-directed crescent, similar to that of lunate cells.

Roles of the different pitcher surfaces in trapping efficiency

Experimental data show that N. alata pitchers constitute an efficient trapping system for both insect species. The percentage of insects trapped during the 5 min of observation in the first experiment is high for ants (29 of the 30 ants fell off into the pitcher, Table 1a) and lower but still considerable for flies (22 of the 30 flies fell off). Moreover, the plant is extremely effective in retaining insects that fell down, since only two ants and no flies succeeded in escaping from the trap.

Table 1.  Comparison of plant trapping efficiency for ants (Iridomyrmex humilis) and flies (Drosophila melanogaster). Numerical results of the first experiment concerning (a) capture and retention of insects (b) falls from different parts of the pitcher (c) falls onto different parts of the pitcher (d) failures of first attempts to climb from different parts of the pitcher. In their first attempts to climb up the inner face of the pitcher, all insects fell down again inside
(a)Number placedFallenEscaped   
Ants3029  2   
Flies3022  0   
(b)Fallen from Lid (inner face)PeristomeWaxy wallTotalP (Fisher exact test) 
Ants  032629  0.175 
Flies  321722  
(c)Fallen onto Transitional wallGlandular wallDigestive liquidTotalP (Fisher exact test) 
Ants1012  729  0.0005 
Flies  1  41722  
(d)When first trying to climb up, were stopped in/on Digestive liquidGlandular wallTransitional wallWaxy wallTotal P (Fisher exact test)
Ants  1  51310290.00001
Flies1010  2  022 
Those that fell in the liquid were stopped where?
Ants  1  1  5  0  70.005
Flies10  6  1  017 
Those that fell on the glandular wall were stopped where?
Ants   4  5  3120.081
Flies   4  0  0  4 
Those that fell on the transitional wall were stopped where?
Ants    3  7100.364
Flies    1  0  1 

The surfaces responsible for the initial fall of insects The principal surface from which insects fell was the waxy zone (26 out of 29 fallen ants and 17 outof 22 fallen flies fell from this surface; Table 1b). The peristome accounted for three ants and two flies. The inner face of the lid accounted for three flies and no ants (but only two ants walked on this surface during the experiment). The proportions of fallen insects between these three surfaces were not significantly different between insect species (Fisher's exact test, P = 0.175). Observations on individual insects indicate that both flies and ants could remain on the margin of the waxy zone even with up to five legs on the waxy surface and at least one on the peristome. However, as soon as all six feet were placed on the waxy surface, the insect fell. Neither insect could walk much more than one step on the waxy surface: only one fly and six ants could make a few steps before falling. Insects that placed fewer than six feet on the wax and then succeeded in climbing back on to the peristome, immediately began to clean their legs and pads.

The main zone on which insects fell include, in order of frequency: the digestive liquid, the glandular wall and the transitional wall. A greater proportion of ants than flies landed on the pitcher walls than the digestive liquid (Fisher's exact test, P < 0.001; Table 1c).

The surfaces responsible for the retention of insects The digestive liquid contributes to the retention of the insects. Once inside it, a little more than half of the insects succeeded in emerging from the liquid and these were mostly ants (χ2 = 3.96, P = 0.047; Table 1d). Insects that did not succeed in escaping from the liquid died quickly.

Insects that succeeded in getting out of the liquid, along with those that landed on the unimmersed surfaces, tried to climb up the wall of the pitcher. Almost all of them fell off again into the liquid after one or more attempts. Only two ants escaped from the trap on their second attempt. All the other insects were dead after 10 min of observation. Ants made one to six attempts of climbing (mean = 2, n = 29), while the flies made only one to four attempts (mean = 1.4, n = 22).

During the first attempts to climb out, no insect reached the peristome. They were all stopped, retained or made to fall by one of the surfaces below (Table 1d).

Of the 34 insects that first began to climb up from the liquid, none could reach the waxy surface (Table 1d). Those that reached the end of the transitional zone could not put more than three feet on the wax, without slipping. Insects were stopped with different frequencies in the liquid, on the glandular zone or on the transitional zone, and these frequencies were significantly dependent on the insect species (Fisher's exact test of independence, P = 0.005; Table 1d). More flies than ants were stopped in the liquid or on the glandular surface than on the transitional area. This narrow transitional zone, however, plays a partly retentive role by inducing the fall of one-quarter of insects that were trying to climb up (Table 1d). The unsubmersed surface of the glandular wall was the most important one in retaining both insects and was more effective against flies than ants. Flies had difficulty in detaching their feet from the surface and both insects cleaned themselves either by rubbing their feet against antennae (ants) or against wings (flies). While cleaning, insects orientated themselves vertically, head down. Ants appeared to further stabilize their position by pressing their abdomen to the pitcher wall.

Among the 16 insects that began their first attempt to escape from the unsubmersed glandular surface, only three ants and no flies reached the waxy wall and none of the flies got further than the glandular zone. However, despite a statistical trend (P = 0.08), insects were not stopped in significantly different proportions on each surface during escape attempts (Fisher exact test, P = 0.08; Table 1d).

Among the 11 insects that first attempted to escape from the transitional wall, seven ants reached the waxy wall (Table 1d), which is more than escape attempts from the liquid or the unsubmersed glandular wall. This indicates that the transitional wall and the waxy wall appeared to be more slippery for insects that had been in contact with the glandular wall or digestive liquid.

Effects of different surfaces on the time spent by insects The two-way anova on time of trapping (period elapsing before first falling into the pitcher) was carried out on raw data since they follow a normal distribution (Shapiro test, w = 0.97, P = 0.34). As shown by the model (r2 = 0.60), plant variation did not significantly influence the time of trapping (F9,31 = 1.63, P = 0.15). After backward selection, the interaction between plant and insect, despite a statistical trend, was not significant (F18,31 = 1.90, P = 0.06). Only the insect species was revealed to be significant (F1,49 = 9.23, P = 0.004): ants were trapped significantly more quickly (mean time of trapping 40.7 ± 47.3 s) than flies (102.8 ± 95.8 s). Observations showed that they were more strongly attracted to extra-floral nectaries situated between lobes of the peristome and were thus more inclined towards the trap.

According to the three-way anova results (Table 2), surface type had the most significant effect on the proportion of time spent by an individual on a surface once visited. According to the multiple comparisons of means (Table 3), the surface on which individuals spent the least time was the waxy zone, followed by the peristome and the internal face of the lid. The surface on which insects spent the most time was the glandular zone. The interaction between the type of surface and the insect species was significant (Table 2), resulting in ants and flies spending different proportions of time on the same surfaces (Fig. 2). Separate analyses carried out for each surface (two-way anovas testing for the effects of plant and insect species) revealed this to be especially true for the waxy zone, where ants stayed a significantly greater proportion of time than flies (F1,36 = 7.91, P = 0.008). They also spent a greater proportion of time than ants in the digestive liquid (F1,28 = 5.99, P = 0.02). On the external surface of the pitcher, flies spent a significantly greater proportion of time than ants (F1,28 = 4.51, P = 0.04). In the global analysis (three-way anova; Table 2), the interaction between plant and surface was also significant: the proportion of time spent by an insect on a surface varied between plants but not in the same way for the different surfaces. According to the separate analyses, the plant effect was significant for the time spent on the peristome (F9,50 = 2.25, P = 0.03) and that spent on the wax (F9,36 = 2,30, P = 0.03), therefore some plants induced shorter insect visits (or earlier falls) on these surfaces than others.

Table 2.  Effect of the plant, the insect species and the type of pitcher surface on the visit time
SourcedfSSMSFP
  1. Results of the three-way anova: r2 = 0.72, F128,136 = 2.73, P = 0.0001. Dependent variable: arcsine [square root (T/Ttotal)] where T is the time spent by an individual on a surface visited and Ttotal corresponds to the 5-min delay of the experiment. SS: sum of squares, MS: mean square.

Plant    9  0.510.06  0.770.6398
Surface    7  8.051.1515.570.0001
Insect    1  0.060.06  0.800.3735
Plant × surface  59  7.820.13  1.790.0029
Plant × insect    9  0.770.09  1.160.3231
Insect × surface    7  1.110.16  2.150.0429
Plant × insect × surface  36  2.940.08  1.110.3310
Error13610.04   
Table 3.  Pitcher surfaces listed in decreasing order of time visited by insects
Type of pitcher surfaceMeann
  1. Means are the arcsine (square root) transformations of time proportions. Means followed by different letters are significantly different from each other, as determined by the multiple comparisons of means (Tukey tests with Kramer corrections for unequal sample sizes) carried out in the anova.

Glandular surface0.707a37
Digestive liquid0.674ab30
Transitional surface0.668ab24
External surface0.485bc30
Outer lid0.367c23
Inner lid0.361c14
Peristome0.289cd60
Waxy surface0.130d47
image

Figure 2. Results of 5 min experiment. Mean time spent by an insect (closed columns, ants; open columns, flies) on the different pitcher zones encountered (30 trials for each insect species). Comparison between two insect species: ext. s., external surface; u. s., upper surface; l. s., lower surface; perist., peristome; z., zone; trans., transitional; gland., glandular.

Download figure to PowerPoint

Effects of different surfaces on ant displacement

The second group of experiments investigated the locomotion of ants on specific tissue surfaces represented by excised ‘tubes’ dissected from fresh living pitchers. It allowed us to show the effects of the microstructures and the plant secretions on ant displacement. Each experiment lasted 1 min.

The waxy zone

Effect of the wax None of the 30 ants escaped from cylinders with the waxy conductive zone intact. They did not manage to traverse a distance of 10 mm without slipping. After removal of the wax by chloroform, six ants succeeded in escaping. Ants that did not reach the top of the cylinder in one minute managed to climb significantly higher than those on the intact waxy surface (T = 1, Z = 4.76, P < 0.001, Wilcoxon's signed-ranks test for paired samples, Fig. 3a). The fact that only six ants succeeded in escaping from wax-free cylinders suggested that the wax is not the only microstructure that prevented escape.

image

Figure 3. Results of paired experiments on dissected cylinders of waxy zone. Comparison of ant vertical speed inside cylinder with different treatments of cylinder surfaces: (a) with wax (closed columns; no escapes) and without wax (open columns; six escapes); (b) normal position (closed columns, no wax; six escapes) and inverted position (open columns, no wax; 29 escapes); (c) untreated wax (closed columns; no escapes) and water aerosol-treated wax (open columns; no escapes).

Download figure to PowerPoint

Effect of pitcher orientation By contrast, 29 of the 30 ants managed to emerge from dewaxed cylinders that had been inverted. Moreover, ant speed was significantly higher for inverted cylinders (Wilcoxon's signed-ranks test, T = 33, Z = 4.1, P < 0.001; Fig. 3b).

Effect of water aerosol on the wax surface None of the 30 ants succeeded in escaping from the waxy cylinders sprayed with a fine aerosol of water. This treatment also appeared to decrease ant mobility and adhesion to the waxy surfaces. While ants could generally walk a small distance (< 10 mm) on untreated surfaces, on aerosol-treated surfaces they generally started to fall as soon as all six legs were placed on the wax. Distances covered were thus significantly less than those on dry surfaces (Wilcoxon's signed-ranks test, T = 80, Z = 2, P = 0.045; Fig. 3c).

The glandular surface

Effect of the glandular surface secretions Fifteen of the 30 ants escaped from freshly cut cylinders within the glandular zone. Twenty-eight escaped from the cylinders left to dry at room conditions for 10 min. On fresh cylinders, ant movements were for longer periods and individuals appeared to have difficulty in detaching their feet from the surface. Ants also cleaned their feet against their antennae. Speed of ant displacement was significantly lower on fresh cylinders than on dry ones (Wilcoxon's signed-ranks test, T = 35, Z = 4.06, P < 0.001, Fig. 4).

image

Figure 4. Results of paired experiments on dissected cylinders of digestive zone. Comparison of ant vertical speed on freshly cut cylinders (closed columns; 15 escapes) and after 10 min (open columns; 28 escapes).

Download figure to PowerPoint

Effect of pitcher orientation Twenty-nine ants escaped from partly dried inverted cylinders and 28 escaped from naturally oriented cylinders. The speed of displacement between the two treatments was not significantly different (Wilcoxon's signed-ranks test, T = 143, Z = 1.84, P = 0.07).

Discussion

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgements
  9. References

Our experiments show that N. alata is very efficient in both initial entrapment and retaining of insects. This results from a complex design of the trap, combining morphological, microstructural and chemical features. Pitchers of N. alata display a wide range of surfaces with different functions. Among all factors tested, the type of surface had the most significant incidence on the length of insect visit: each one displays different ‘slippery’ or ‘adhesive’ properties responsible for the fall or retention of insects within pitchers. This wide diversity of functional surfaces appeared to be effective on insects with different pad designs. Because Nepenthes spp. greatly depend on insect derived-nitrogen (Schultze et al., 1997), this contingency effect of different surfaces might represent an adaptation of the plant for maximizing insect prey.

Our results show that the surface primarily responsible for the initial fall of insects is the waxy zone. This finding is consistent with many observers (Lloyd, 1942; Juniper & Burras, 1962). The waxy zone is also the surface on which insects are able to remain the least amount of time. Three other pitcher surfaces cause insects to fall. The peristome, which has been reported as presenting precarious footholds for larger insects or as an obstacle to escape (Juniper & Burras, 1962), is also involved in the falling of small insects. The lower surface of the lid and the transitional zone between the waxy and transitional surface also induced falling of insects. All three surfaces have different morphological features but all are composed of relatively smooth epidermal cells. Because of a scarcity of footholds, such as hairs and trichomes, insects cannot use their claws and are obliged to use their attachment pads (Lloyd, 1942). Attachment pads, particularly among flies (Wigglesworth, 1987), enable insects to walk on smooth surfaces (Gorb, 1998; Gorb, 2001) even at high inclinations, such as in the pitcher lid or parts of the peristome. Thus, smoothness alone cannot explain loss of attachment on these surfaces.

Surface relief and insect attachment

The surface relief combined with overall slope angle may contribute to preventing attachment. For example, smooth corrugations on the lower face of the lid and those arranged vertically in a system of overlapping protuberances on each terminal tooth of the peristome, presumably reduce the contact area of insect pads with the substrate. In addition, geometry and orientation of epidermal cells might influence anisotropy in the attachment abilities of insects.

The potential role of lunate cells, with convex downward directed surfaces, was suspected by Knoll (1914, cited by Lloyd, 1942) but his experiments could not reveal any effect on ant locomotion. Our study demonstrates effects of anisotropic microstructure of the lunate cells on insect locomotion. The lunate cells may promote insect fall as well as impeding escape. They are likely to be responsible for anisotropic frictional properties of the surface, according to their orientation on the pitcher. Compared with the wax itself, the lunate cells here were found to play an important role in the trapping system. Modification of stomata was probably evolved under selective pressures linked to the carnivory habit and thereby constitutes a striking example of cooption from one function to another.

Wax crystals and insect attachment

The waxy surface is clearly responsible for initial falling of insects tested and hence plays a strategic role for the pitcher plant. As shown by our wax removal experiment, the wax also plays a consistent role in secondary falls and insect retention. In this experiment, the chloroform might have created some artifact causing difficulty in interpreting the results on ant attachment. It might have induced cell desiccation and wrinkling, which influence insect attachment in addition to the wax removal itself. However, our observation of the treated surfaces using SEM, revealed that cell wrinkling was not significant: the lunate cells and surrounding surfaces were intact and retain the shapes of untreated surfaces. Even if there is an effect of desiccation, this effect is probably less influential than wax removal.

Waxy surfaces were reported to play a key role for a number of other angiosperms, among which evolution has often been influenced by close interactions with insects. In effect, such ‘slippery’ surfaces were reported to act as a barrier against phytophagous insects (Eigenbrode & Espelie, 1995), as an ecological filter against some ants, such as nectar-robbing ones (Harley, 1991), or parasites of ant–plant mutualisms (Federle et al., 1997). Strategically positioned, such as on pedicels of reproductive organs or on inflated twigs nesting protective ants, waxy surfaces represent an important measure of protection in what Juniper (1995) coined the ‘greasy pole syndrome’.

Here, the exact mechanism of the wax is not completely resolved. Two hypotheses are commonly put forward: (1) the wax may reduce pad contact area and thus reduce insect adhesion because of its crystalline structure, which is responsible for the surface microroughness (Stork, 1980); (2) the waxy layer may induce falling of insects because of its fragility – wax crystals are easily detachable plates – and possibly prevent attachment because of a contamination effect where wax plates become detached and adhere to the pads (Juniper & Burras, 1962; Juniper et al., 1989). Two further hypotheses can also be suggested: (3) wax may have chemical properties responsible for its slipperiness and (4) the wax may chemically interfere with insect adhesive secretions. Our SEM observations support the results of Juniper & Burras (1962) by visualizing wax crystal as detachable structures (Fig. 1i). Moreover, the fact that insects very frequently cleaned their pads after having walked on the wax is also coherent with the second hypothesis. Wax contamination has also been reported in other plants (Eigenbrode et al., 1999). This implies that wax crystals have rather strong adhesive properties and would lead us to reject hypothesis (3) but this needs to be confirmed by biomechanical tests. We are unable to reject hypotheses (1) and (4), but they are not exclusive.

The microroughness of the wax may furthermore explain why we found that an aerosol of water reduced insect adhesion to it. Water repellence by wax is not only due to the hydrophobic constituents of the wax, but also to the microroughness of the surface. It diminishes the surface angle of contact of water droplets, which run off the surface quickly, removing contaminating particles, because particles also have a weak contact area with substrate. This self-cleaning ability of structured hydrophobic surfaces is called the ‘Lotus effect’ (Barthlott & Neinhuis, 1997). In our experiments, ants encountered greater difficulty with adhering to a water-sprayed waxy surface. This result might be explained by the fact that the pads were detached from the surface by moving water droplets in a similar way to foreign particles. This may be of ecological importance for the tropical plant if it enhances trapping efficiency during the rainy season.

Glandular secretions and insect attachment

Glandular secretions comprising the digestive liquid may chemically influence insect behavior by secreting paralyzing substances before digestive compounds. However, the presence of paralyzing substances has only been demonstrated in Sarracenia flava (Sarraceniaceae) (Juniper et al., 1989). Our observations showed that glandular secretions may also act mechanically like a glue and impede locomotion. Experimental data indicate that insects spent the longest time on the glandular surface. Furthermore, ants walk significantly faster on partly dried glandular surfaces than on fresh ones. In addition, both ants and flies spent much time on this glandular zone cleaning their pads. The glandular surface appears to constitute an adhesive zone in Nepenthes for retaining insects. In another family of pitcher plants lacking digestive glands (the Sarraceniaceae), the retaining function is fulfilled by microstructures, such as downward-pointed retentive hairs (Adams & Smith, 1977). Our findings suggesting an adhesive role of glandular secretions needs to be confirmed by more detailed biomechanical tests. Nevertheless, they offer a potentially fascinating insight into common, possibly homologous functional similarities between Nepenthes and flypaper-trapping plants such as Drosera and Drosophyllum, which are phylogenetically more closely related to Nepenthes than other pitcher plants, such as the Sarraceniaceae and Cephalotus (Albert et al., 1992; Williams et al., 1994).

Many insects use adhesive secretions to adhere to surfaces (Gorb, 1998, 2001; Orivel et al., 2001). The present study shows that insects coming into contact with glandular secretions (especially flies), were unable to reach further than the adhesive zone. Presumably, the gland secretion may contaminate insect pads and interfere with their own adhesive secretion. These observations suggest the importance of contingency in the trapping function. Some surfaces, such as the transition zone, become even more effective if the insect's pads have previously been in contact with other surfaces.

Trapping efficiency and different insect prey

High trapping ability results from combined contingencies of different surfaces with different ‘strategies’ in relation to the insect prey. Comparative data show different effects of various pitcher surfaces on different insect species. One substantial difference between flies and ants was in the duration of stay on the vertical waxy surface. Some ants were able to make a few steps on the wax before falling down, both during the initial fall or the escape attempts. Furthermore, two ants succeeded in crossing the waxy zone and emerged from the trap. Ants were thus able to stay significantly longer on wax than flies. During the capture phase, flies, once on this waxy surface, more often promptly fell while trying to straddle the peristome. As the peristome overhangs the pitcher walls, this explains why a significantly greater proportion of ants than flies settled on the pitcher walls rather than falling directly into the liquid. This prompt and direct fall of flies, once on the waxy surface, has also been recorded by other observers (Lloyd, 1942) and reveals that flies were unable to fly or maneuver in flight sufficiently quickly to prevent their fall. This suggests that the presence of functional wings in flies would not drastically change their capture rate.

Why do ants show a greater ability to walk on the vertical waxy surface than flies? Because claws are not used on this surface (Juniper et al., 1989), the type of pads may explain the observed difference. Iridomyrmex humilis is a terrestrial ant and terrestrial ants, in contrast to arboreal ones, often have small or less-effective arolia (Orivel et al., 2001; Federle et al., 2000). Furthermore, they were trapped in a greater proportion than flies. Finally, flies are known to possess effective pulvilli, especially on smooth surfaces (Wigglesworth, 1987; Gorb, 1998). Here, flies are thus likely to have better organs of adhesion than ants. Why then, are they less able to stay on the wax? This paradoxical situation is similar to that discussed by Federle et al. (2000) in relation to waxy Macaranga twigs where they note that ‘better wax runners have a poorer attachment to a smooth surface’. In Nepenthes, where fragile platelets of wax differ from the filamentous wax of Macaranga, it is possible that ‘worse’ wax runners (flies) having ‘better’ organs of adhesion, adhere better to wax crystals, which detach more readily from the plant to contaminate their pads. They are thus impeded more than ants when walking on the wax and consequently fall sooner from the waxy surface.

Conclusion

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgements
  9. References

By focusing on the plant–insect interface, this study underlines the functional importance of epidermal surfaces in the trapping efficiency of N. alata. The experimental design, by quantifying the contribution of surface microstructures, such as modified stomata or wax crystals, in the trapping system, indirectly demonstrated the selective advantage they provide to the pitcher plant. This study therefore shows that microstructural adaptations represent an important aspect of the carnivory syndrome.

This study points to exciting new directions both in the study of the biomechanical mechanisms of the plant surfaces and in the study of the evolutionary pattern of structural features of carnivorous plants. Potentially measurable variables, such as friction and adhesion of insects (Gorb & Scherge, 2000; Jiao et al., 2000) on plant surfaces will undoubtedly help us to understand how such complex interfaces operate. Moreover, further comparative studies will allow us to test whether such epidermal microstructures and their function will represent the key characters for understanding the evolutionary patterns of these remarkably adapted carnivorous plants (Albert et al., 1992).

Acknowledgements

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgements
  9. References

This research was funded by a grant of the Société de Secours des Amis des Sciences, founded by members of the French Academy of Sciences. M. Poissonnet is thanked for his help during the experiments. The manuscript was improved by helpful comments of Carine Brouat, Marianne Elias, Olena Gorb, Christopher Neinhuis and an anonymous reviewer.

References

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgements
  9. References
  • Adams RM, Smith GW. 1977. An S.E.M. survey of the five carnivorous pitcher plant genera. American Journal of Botany 64: 265272.
  • Albert VA, Williams SE, Case MW. 1992. Carnivorous plants: phylogeny and structural evolution. Science 257: 14911495.
  • Barthlott W, Neinhuis C. 1997. Purity of the sacred lotus, or escape from contamination in biological surfaces. Planta 202: 18.
  • Beutel RG, Gorb SN. 2001. Ultrastructure of attachment specializations of hexapods (Arthropoda): evolutionary patterns inferred from a revised ordinal phylogeny. Journal of Zoological Systematics and Evolutionary Research 39: 177207.
  • Darwin C. 1875. Insectivorous plants. London, UK: John Murray.
  • Eigenbrode SD, Castognola T, Roux M-B, Steljes L. 1999. Mobility of three generalist predators is greater on cabbage with glossy leaf wax than on cabbage with a wax bloom. Entomologia Experimentalis et Applicata 81: 335343.
  • Eigenbrode SD, Espelie KE. 1995. Effect of plant epicuticular lipids on insect herbivores. Annual Review of Entomology 40: 171194.
  • Elison AE, Gotelli NJ. 2001. Evolutionary ecology of carnivorous plants. Trends in Ecology and Evolution 16: 623629.
  • Federle W, Maschwitz U, Fiala B, Riederer M, Hölldobler B. 1997. Slippery ant-plants and skilful climbers: selection and protection of specific ant partners by epicuticular wax blooms in Macaranga (Euphorbiaceae). Oecologia 112: 217224.
  • Federle W, Rohrseitz K, Hölldobler B. 2000. Attachment forces of ants measured with a centrifuge: better ‘wax runners’ have a poorer attachment to a smooth surface. Journal of Experimental Biology 203: 505512.
  • Gorb SN. 1998. The design of the fly adhesive pad: distal tenent setae are adapted to the delivery of an adhesive secretion. Proceedings of the Royal Society of London B 265: 747752.
  • Gorb SN. 2001. Attachment devices of insect cuticle. Dordrecht, The Netherlands: Kluwer Academic Publishers.
  • Gorb SN, Scherge M. 2000. Biological microtribology: anisotropy in frictional forces of orthopteran attachment pads reflects the ultrastructure of a highly deformable material. Proceedings of the Royal Society of London B 267: 12391244.
  • Harley R. 1991. The greasy pole syndrome. In: HuxleyCR, CutlerDF, eds. Ant–plant interactions. Oxford, UK: Oxford University Press, 310330.
  • Jiao Y, Gorb S, Scherge M. 2000. Adhesion measured on the attachment pads of Tettigonia viridissima (Orthoptera, Insecta). Journal of Experimental Biology 203: 18871895.
  • Juniper BE. 1995. Waxes on plant surfaces and their interactions with insects. In: HamiltonRJ, ed. Waxes: chemistry, molecular biology and functions. Dundee, UK: The Oily Press, 157174.
  • Juniper BE, Burras JK. 1962. How pitcher plants trap insects. New Scientist 13: 7577.
  • Juniper BE, Robins RJ, Joel D. 1989. The carnivorous plants. London, UK: Academic Press.
  • Juniper BE, Southwood R. 1986. Insects and the plant surface. London, UK: E. Arnold.
  • Knoll F. 1914. Über die Ursache des Ausgpeitems der Insettenbeine an Wachsbedeckten Pflanzentheilen. Jahrbücher für Wissenschaftliche Botanik 54: 488497.
  • Lloyd FE. 1942. The carnivorous plants. Waltham, MA, USA: Chronica Botanica Co.
  • Merbach MA, Merbach DJ, Maschwitz U, Booth WE, Fiala B, Zizka G. 2002. Mass march of termites into the deadly trap. Nature 415: 3637.
  • Moran JA. 1996. Pitcher dimorphism, prey composition and the mechanism of prey attraction in the pitcher plant Nepenthes rafflesiana. Borneo. Journal of Ecology 84: 515525.
  • Moran JA, Booth WE, Charles JK. 1999. Aspects of pitcher morphology and spectral characteristics of six Bornean Nepenthes pitcher plant species: implications for prey capture. Annals of Botany 83: 521528.
  • Orivel J, Malherbe MC, Dejean A. 2001. Relationships between pretarsus morphology and arboreal life in ponerine ants of the genus Pachycondyla (Formicinae: Ponerinae). Entomological Society of America 94: 18.
  • Owen TP, Lennon KA. 1999. Structure and development of the pitchers from the carnivorous plant Nepenthes alata (Nepenthaceae). American Journal of Botany 86: 13821390.
  • SAS. 1996. SAS/STAT user's guide, release 6.12 edn. Cary, NC, USA: SAS Institute.
  • Schultze W, Schultze ED, Pate JS, Gillison AN. 1997. The nitrogen supply from soils and insects during growth of the pitcher plants Nepenthes mirabilis, Cephalotus follicularis and Darlingtonia californica. Oecologia 112: 464471.
  • Stork NE. 1980. Role of waxblooms in preventing attachment to Brassicas by the mustard beetle, Phaedon cochleariae. Entomologia Experimentalis et Applicata 28: 100107.
  • Wigglesworth VB. 1987. How does a fly cling to the under surface of a glass sheet. Journal of Experimental Biology 129: 363367.
  • Williams SE, Albert VA, Chase MW. 1994. Relationships of Droseraceae: a cladistic analysis of rbcL sequence and morphological data. American Journal of Botany 81: 10271037.