Abstract

Dionaea is a highly specialized carnivorous plant species with a unique mechanism for insect capture. The leaf is converted into an osmotically driven trap that closes when an insect triggers sensory trichomes. This study investigates the significance of insect capture for growth of Dionaea at different successional stages after a fire, under conditions where the prey is highly variable in its isotope signature. The contribution of insect‐derived nitrogen (N) was estimated using the natural abundance of 15N. In contrast to previous 15N studies on carnivorous plants, the problem emerges that δ15N values of prey insects ranged between −4.47‰ (grasshoppers) and +7.21‰ (ants), a range that exceeds the δ15N values of non carnivorous reference plants (−4.2‰) and soils (+3‰). Thus, the isotope‐mixing model used by Shearer and Kohl to estimate the amount of insect‐derived N is not applicable. In a novel approach, the relationships of δ15N values of different organs with δ15N of trapping leaves were used to estimate N partitioning within the plant. It is estimated that soon after fire approximately 75% of the nitrogen is obtained from insects, regardless of plant size or developmental stage. The estimates are verified by calculating the average isotope signatures of insects from an isotope mass balance and comparing this with the average measured δ15N values of insects. It appears that for Dionaea to survive and reach the flowering stage, seedlings must first reach the 6th‐leaf rosette stage, in which trap surface area nearly doubles and facilitates the capture of large insects. Large amounts of nitrogen thus made available to plants may facilitate an enhanced growth rate and the progressive production of additional large traps. Dionaea reaches a maximum abundance after fire when growth of the competing vegetation is suppressed. About 10 years after fire, when grasses and shrubs recover, Dionaea becomes overtopped by other species. This would not only reduce carbon assimilation but also the probability of catching larger prey. The amount of insect‐derived nitrogen decreases to 46%, and Dionaea becomes increasingly dependent on N‐supply from the soil. Competition for both light and N may cause the near disappearance of Dionaea in older stages of the fire succession.

Introduction

The venus fly trap (Dionaea muscipula) is an endemic carnivorous plant growing on sandy soils in the central south‐eastern coastal plain of North America. Among carnivorous plants, Dionaea has a unique mechanism for capturing insects (Juniper et al., 1989). The leaf is modified to form two moving lobes that will close on prey when the trap is triggered via touch‐sensitive trichomes (Juniper et al., 1989). The trap dies after it closes on one prey. The petiole of the Dionaea leaf is broadened to support photosynthesis.

Since its first description (Ellis, 1770), Dionaea has been of great interest to botanists. The insect‐trapping mechanism of two moving leaf lobes has been particularly well studied (Robins and Juniper, 1980; Fagerberg and Howe, 1996; Trebacz et al., 1996; Trebacz and Sievers, 1998). In contrast, very little is known about the nutrient relationship of Dionaea, and the contribution of N from insect capture to total N content remains unknown. It has been demonstrated that Dionaea takes up 15N‐ or 13C‐labelled amino acids via the traps, but transport and metabolism of these labelled amino acids was only followed to the petioles of fed trap leaves (Greenaway et al., 1990). Kinetic analysis of amino acid uptake in Dionaea leaves (Rea, 1984), and the localization of mRNA encoding for an ammonium transporter in glands of Nepenthes (Schulze et al., 1999), shows that uptake of amino acids and nutrients by carnivorous plants may be regulated by specific transporters. The transporter studies confirm that insect nitrogen is indeed utilized, a process questioned in another species, Roridula, that also has sticky glands (Ellis and Midgley, 1996).

The utilization of insect N by various carnivorous plants in their habitat has previously been quantified using the relative 15N‐enrichment of insects as a natural isotopic marker for insect‐derived N and following an approach described previously (Shearer and Kohl, 1988) for N2‐fixation in plants. In this approach, the fraction of insect‐derived N is calculated from the ratio of the difference in δ15N values of the carnivorous plant and non‐carnivorous reference plants versus the difference between δ15N values of insect nitrogen and reference plants. Using this method, the contribution of insect N to total plant N in their natural habitat was determined to range from 20–54% in a variety of Drosera species (Schulze et al., 1991), and from 26% and 79% in pitcher plants (Schulze et al., 1997).

The growth form of Dionaea as a rosette with leaves close to the ground makes this plant less competitive against grasses and shrubs. Following fire, Dionaea may be highly abundant for 3–5 years until shrubs and monocotyledonous plants overtop the small rosette plants. Thus this species is restricted to early successional stages after fire (Roberts and Oosting, 1958). The succulent cell structure suggests that not only the hypocotyl, but also the root can serve as a storage organ.

The aim of this study was (i) to analyse the dependency of N nutrition in Dionaea on insect capture, (ii) to investigate plant internal partitioning of nitrogen to different organs, and (iii) to understand growth and reproduction in successional stages after fire.

Materials and methods

Dionaea is a small rosette plant with green leaves that serve as an insect‐capturing organ, a leaf‐like photosynthesizing petiole (phyllode; further on referred to as ‘leaf’), a thick, succulent hypocotyl serving as a main storage organ, and thick, succulent non‐mycorrhizal roots (Roberts and Oosting, 1958) which also appear to serve partly as storage organs. The apex forms new rosette leaves which remain non‐capturing until it has reached its final size.

Site description

Plant material of Dionaea muscipula, as well as reference plants and insects, was collected in situ at three neighboring sites in North Carolina, USA. Site I was an open Pinus longifolia heathland on nutrient‐deprived coastal quartz sands which had burned by a surface fire about 2 years before this study. The understorey consisted of herbaceous and graminoid species, no shrubs were present. Dionaea plants formed a dense lawn together with Aristida stricta, Ctenium aromaticum, Lachnocaulon ancheps, Fimbristylis sp. and herbaceous species of the genera Polygala, Calopogon, Dichromena, Lobelia, and Tephrosia. Site II, directly adjacent to site I, was an open heath/woodland representing an intermediate successional stage after fire (about 10 years previously), as indicated by small shrubs (Ilex glabra, Magnolia virginiana, Gaylussacia fondosa). Site III was an open woodland, with 2–3 m high shrubs of Ilex glabra, Magnolia virginiana and Gaylussacia fondosa, and Pteridium aquilinum and Lycopodium indicating that this site had burned about 20–30 years previously.

KCl‐extractable soil ammonium concentrations were low at all three sites (site I: 0.08 μmol g−1 DW±0.045 sd, site II: 0.14 μmol g−1 DW±0.048 sd, site III: 0.06 μmol g−1 DW± 0.017 sd), nitrate could not be detected. Differences between sites were not significant (P>0.05) indicating that in all three sites the investigation was not affected by the short‐term mineralization effects of fire. The low fertility of the coastal sands of North Carolina in the Dionaea habitat is characterized by a lack of calcium, manganese, and nitrate and low levels of ammonia, iron, magnesium, and phosphate in recently burnt soils (Roberts and Oostings, 1958).

Sampling procedures and analyses

At site I, site II, and site III 25, 13, and 10 plants were sampled, respectively. The plants ranged from very small seedlings with few leaves to mature flowering plants, representing the range in plant size found at each site. Plants were separated into open traps (with separate sampling of traps with and without insects) and closed, still developing leaf traps, as well as leaf petioles, roots, central storage organ, and flower stalks. All plant organs were weighed and 15N and 13C isotopic composition and N concentrations were determined in an on‐line system connecting an elemental analyser to an isotopic mass spectrometer (deltaS, FinneganMAT, Bremen, Germany) with a precision±0.1‰. In addition, samples from the non‐carnivorous neighbouring plants and the trapped insects were collected for δ15N and δ13C analysis. Non‐carnivorous plants were all VA‐mycorrhizal, which leads to an uncertainty in the analysis. In a separate set of samples, leaves containing fresh prey were collected from different plants, and the δ15N and δ13C values of insects and leaves and their weight were determined. All regression analyses were performed using SPSS statistical software.

Calculation of the insect‐derived nitrogen

The amounts of insect‐derived nitrogen were calculated as follows:  

1
formula
Based on the biomass measurements, R%, S%, YL%, L%, and F% are the percentages of roots, storage organs, young leaves, capturing leaves, and flowers of total weight, respectively. The coefficients r, s, l, and f can be obtained as the slopes of the regressions between the δ15N‐values of roots, storage organs, young leaves and flowers, respectively and leaf δ15N values. The slopes are thus used as a measure of the dependence of the nitrogen sink organs (new leaves, storage organ, roots, flowers) on nitrogen from the source organ (trapping leaf). 15N‐discrimination during assimilation and distribution, as well as different source of nitrogen for the sink organ may affect this relationship. Therefore, as an independent validation of the calculated insect‐derived nitrogen, an isotope mass balance was calculated as following:  
2
formula
where δ15Nc is total δ15N in the capturing plant, δ15Ni is the δ15N of captured insects, δ15Nnc is the δ15N value of non‐capturing Dionaea plants, fi is the fraction of insect‐derived nitrogen, and (1−fi) is the fraction of soil‐derived nitrogen. Solving the equation for d15Ni results in:  
3
formula
This calculated value of δ15Ni can then be compared with the average δ15N of the insects captured by Dionaea. A large difference between calculated and measured δ15Ni would indicate that an important source of nitrogen was not considered in this approach.

Results

N concentration and content in Dionaea at different successional stages after fire

Nitrogen concentration was independent of plant size at sites I and II, but large differences in N concentration existed among plant organs (Table 1), with leaf nitrogen concentrations being higher than in roots. Only the leaf and trap nitrogen concentration differed among sites—plants at site I with the largest Dionaea population had half the nitrogen concentration in leaves compared to plants from site III where Dionaea was outcompeted by other species. Site III was the only one in which small plants (less than 2 mg N plant−1) had significantly higher (15 mg g−1) whole‐plant nitrogen concentrations compared to large plants (9 mg g−1).

Nitrogen content varied with the number of leaves per plant (Fig. 1.), reflecting plant size. At site I, small seedlings and plants with fewer than six leaves had N contents of less than 4 mg plant−1, averaging 0.77 mg plant−1 (±0.42 sd). However, at the stage of six leaves or more, N content increased 14‐fold to an average of 9.69 mg plant−1 (±3.9 sd). There is an additional group of plants, which had flowers but few or no leaves. N content of these plants was 3.69 mg plant−1 (±3.51 sd). Plants belonging to this latter group at sites II and III were of comparable size in site I.

The change in biomass composition during development of Dionaea is summarized in Table 2 . In small plants with fewer than six leaves, the highest proportion of biomass is leaf biomass, which decreases significantly when plants start flowering. The fraction of root biomass is smallest in large plants.

The strong increase in nitrogen content (Fig. 1) and biomass was accompanied by a strong increase in the average size of individual traps per plant. Only plants with more than six leaves had large traps (Fig. 2A). The association between trap size and number of leaves produced an almost exponential increase in plant nitrogen content with the average trap size (Fig. 2B).

Fig. 1.

(A) Nitrogen content of individual Dionaea plants at the three investigated sites reflecting the biomass composition. Plants are listed according to their number of leaves following their development from seedlings with few leaves to mature plants with many leaves or flowering plants with no leaves.

Fig. 1.

(A) Nitrogen content of individual Dionaea plants at the three investigated sites reflecting the biomass composition. Plants are listed according to their number of leaves following their development from seedlings with few leaves to mature plants with many leaves or flowering plants with no leaves.

Fig. 2.

(A) The average trap size per plant depending on leaf number for plants at site I. (B) Plant nitrogen content increased exponentially with the average trap size per plant. Trap size was measured as the area of the trapping leaf lobes.

Fig. 2.

(A) The average trap size per plant depending on leaf number for plants at site I. (B) Plant nitrogen content increased exponentially with the average trap size per plant. Trap size was measured as the area of the trapping leaf lobes.

Table 1.

Nitrogen concentrations in various organs of plants collected at the different sites

Values are expressed as means of all plants analysed ±standard error. Letters indicate significant differences between plant organs and experimental sites (P<0.05).

 Nitrogen concentration (mg g−1)
 

 

 

 
Site I
 
Site II
 
Site III
 
Root 9.0±1.3 ab 7.9±0.6 b 9.7±0.5 b 
Storage organ 7.8±1.1 bc 9.8±1.5 b 7.8±0.8 bc 
Leaf+trap 5.7±0.7 c 9.4±0.7 b 10.6±2.0 ab 
Closed leaf 14.7±2.3 a 12.4±0.6 ab – 
Flower 10.5±1.2 ab 9.7±1.0 b 8.6±0.5 b 
Total (small plants) 8.1±1.7 ab 12.2±1.1 ab 15.0±3.3 a 
 Nitrogen concentration (mg g−1)
 

 

 

 
Site I
 
Site II
 
Site III
 
Root 9.0±1.3 ab 7.9±0.6 b 9.7±0.5 b 
Storage organ 7.8±1.1 bc 9.8±1.5 b 7.8±0.8 bc 
Leaf+trap 5.7±0.7 c 9.4±0.7 b 10.6±2.0 ab 
Closed leaf 14.7±2.3 a 12.4±0.6 ab – 
Flower 10.5±1.2 ab 9.7±1.0 b 8.6±0.5 b 
Total (small plants) 8.1±1.7 ab 12.2±1.1 ab 15.0±3.3 a 

Table 2.

Biomass composition as a percentage of total biomass for three stages of development in Dionaea

A small plant represents plants with up to six leaves, a large plant has 7–18 leaves and may be flowering, flowering plants have no leaves and a large flower. Values are listed as means ±standard deviation. Letters indicate significant differences between differently sized plants for each organ.

 Biomass compsotion (%)
 

 

 

 

 

 
Flower
 
Root
 
Storage organ
 
Closed leaves
 
Capturing leaves
 
Small plant – 18.8±4.3 a 25.5±3.5 a 2.2±0.2 a 53.5±11.3 a 
Large plant 1.8±0.8 a 8.3±1.5 b 24.9±7.6 a 9.5±0.5 b 55.5±9.0 a 
Flowering plant 65.2±0.5 b 11.9±0.7 c 20.1±8.3 a – 2.8±0.6 b 
 Biomass compsotion (%)
 

 

 

 

 

 
Flower
 
Root
 
Storage organ
 
Closed leaves
 
Capturing leaves
 
Small plant – 18.8±4.3 a 25.5±3.5 a 2.2±0.2 a 53.5±11.3 a 
Large plant 1.8±0.8 a 8.3±1.5 b 24.9±7.6 a 9.5±0.5 b 55.5±9.0 a 
Flowering plant 65.2±0.5 b 11.9±0.7 c 20.1±8.3 a – 2.8±0.6 b 

Variation in δ15N and δ13C values in the prey of Dionaea

The size of insect prey ranged between 0.3 mg for ants and 20–30 mg for grasshoppers and spiders. The δ15N values of these preys is mainly determined by the ratio of chitin (δ15N=−10‰) to protein (δ15N=+12‰, G van Klinken, personal communication) and by differences in the diet of the insects (Lee‐Thorp et al., 1989; Nelson et al., 1998). There is no consistent relationship between δ15N of the insect and insect dry weight. However, more importantly for this study, plants with small traps preferentially caught small ants with δ15N values of about +6‰, while the prey in larger traps included large grasshoppers with δ15N values of −4‰ (r2=0.46, P<0.05: Fig. 3A). δ15N values of non‐carnivorous reference plants were in the range of −2‰ to −4‰. A few of the large traps (4%) also captured ants, but no grasshopper was found in a small trap. Soil to 10 cm depth at all three sites had more positive δ15N values (+1 to 3‰) than non‐carnivorous reference plants but lower δ15N values than ants. The overlapping ranges of prey and trap δ15N values and of soil and reference plants make an isotope‐mixing model according to Shearer and Kohl (Shearer and Kohl, 1988) unsuitable for estimating insect N in Dionaea. Thus, a more detailed analysis of the partitioning of insect nitrogen within the plant during the uptake process was performed.

The δ15N value of Dionaea leaves correlated positively (slope of 0.3) with the δ15N value of the prey (r2=0.53, P<0.017: Fig. 3B). Leaves on the same plant may exhibit different δ15N values depending on the prey. In all cases, δ15N values of Dionaea leaves were more positive than δ15N values of non‐carnivorous reference plants. The reference plants were VA‐mycorrhizal, and thus probably dependent primarily on inorganic nitrogen, while non‐mycorrhizal roots of Dionaea may also take up free amino acids (Marschner, 1988). The following estimation of insect‐derived nitrogen is not based on reference plants, and is therefore independent of the potential error introduced by different sources of soil nitrogen.

In addition to the large variation in δ15N values, insects also showed large variation in δ13C values (Fig. 4A). On average, ants (−23‰) had more negative δ13C values than grasshoppers (−19‰). However, grasshoppers showed a large variation in δ13C values ranging from –30‰ to –12‰, reflecting the δ13C values of the respective diet plant (C3 plant or C4 plant). The high variation in δ13C of grasshoppers and the difference from δ13C values of ants affected the correlation between δ15N and δ13C values of the capturing leaf: Depending on the type of prey, there was a tendency for leaves to have low δ13C values and high δ15N values when ants were prey, and low δ15N values and high δ13C values when C4 plants‐fed grasshoppers were prey (Fig. 4B). However, there was also a group of leaves with low δ13C values that seemingly had captured grasshoppers feeding on C3 plants with low δ13C values.

Due to the variation in δ13C values there was no simple relationship between trap biomass and δ13C values of insects. Because Dionaea leaves perform C3 photosynthesis, the relationship between δ13C values of leaves and captured insects was additionally affected by the contribution of photosynthesis by Dionaea (data not shown).

Fig. 3.

(A) δ15N values of insects captured by traps of different size (dry weight). The type of insect captured is listed next to their δ15N values and δ15N values of soil and reference plants are shown as dotted lines. The inverse relationship between δ15N values of captured insects and trap size is significant (P<0.05). (B) Relationship between δ15N values of Dionaea traps and δ15N values of insects captured in these traps. The regression is significant (P<0.017). Long‐dashed lines represent the 95% confidence interval.

Fig. 3.

(A) δ15N values of insects captured by traps of different size (dry weight). The type of insect captured is listed next to their δ15N values and δ15N values of soil and reference plants are shown as dotted lines. The inverse relationship between δ15N values of captured insects and trap size is significant (P<0.05). (B) Relationship between δ15N values of Dionaea traps and δ15N values of insects captured in these traps. The regression is significant (P<0.017). Long‐dashed lines represent the 95% confidence interval.

Fig. 4.

(A) δ15N and δ13C values of different types of insects. Open symbols represent averages of three to five insects, and filled symbols show individual data points. (B) δ15N and δ13C values of leaves having captured an insect. Circles indicate leaves that had similar prey. Data points in the thick circle were not used for the regression analysis. Long‐dashed lines represent the 95% confidence interval.

Fig. 4.

(A) δ15N and δ13C values of different types of insects. Open symbols represent averages of three to five insects, and filled symbols show individual data points. (B) δ15N and δ13C values of leaves having captured an insect. Circles indicate leaves that had similar prey. Data points in the thick circle were not used for the regression analysis. Long‐dashed lines represent the 95% confidence interval.

The relationship between δ15N values of different tissues of Dionaea

The δ15N values of developing traps that were still closed (i.e. never caught an insect) correlated positively with the δ15N values of capturing traps (r2=0.84, P<0.001, Fig. 5A). The slope of the regression (0.95) was not significantly different from 1, indicating a rapid and indiscriminant transport of nitrogen from traps to developing leaves.

The positive relationship between δ15N values of the storage organ (P<0.01, Fig. 5B) or roots (P<0.01, Fig. 5C) and δ15N values of traps not only showed higher variation (r2=0.44 and r2=0.43, respectively), but the slopes of the relationship decreased compared to the slope with closed traps (Fig. 5A). The correlation between δ15N values of flowers and trapping leaves showed a slope near 0.8 with little deviation for plants with leaves (Fig. 5D), and a higher deviation for flowering plants without leaves (Fig. 5E). The regressions of Fig. 5 are based on plants growing at site I and site II, while the relationships for plants at site III were shallower, as shown by the short‐dashed lines in Fig. 5B and C. This may indicate that plants at site III have some access to additional nitrogen sources. The slope of the regression between δ15N values of traps and storage organ at site III was 0.20, and decreased to a slope of 0 for the relationship between δ15N values of traps and roots at this site (short dashed lines in Fig. 5B, C).

Using the slopes in Fig. 5 as an indication for the partitioning of nitrogen between source and sink organs, it was estimated that about 95% (slope in Fig. 5A) of the nitrogen in young leaves, 54% (Fig. 4B) of nitrogen in storage organs, 46% (Fig. 5C) of nitrogen in roots, and 76% (Fig. 5D, E) of the nitrogen in flowers originated from the trapping source leaf in plants growing at site I and site II. Following equation 1 and taking the differences in dry weight composition (Table 2) into account, up to 75% of the nitrogen in young seedlings originates from insects at sites I and II. In large capturing plants the proportion of insect nitrogen increases to about 80% due to the higher proportion of capturing leaf biomass and lower proportion of root biomass. In large flowering plants without leaves, the fraction of insect nitrogen decreases to 68%, due to an increased fraction of root biomass and decrease in that of capturing leaf biomass.

The amounts of insect‐derived nitrogen for large plants (more than 2 mg total N) at the three different sites are summarized in Table 3. Highest amounts of insect‐derived nitrogen were found at site I (80%), whereas at sites II and III, the amount of insect‐derived nitrogen decreased significantly to 70% and 46%, respectively. Not only the percentage of insect‐derived nitrogen changed between sites, but also the total δ15N values of the plants (Table 3). Large plants at site I had lower δ15N values than plants at sites II and III. This may indicate a change in the diet of large plants from a dependence on mainly grasshoppers at site I to a dependence on more ants and spiders at sites II and III. This did not change the N‐concentrations at the whole‐plant level. Estimate of the mean δ15N‐value of insects using equation 3 revealed a δ15N‐value for insects of 1.11‰ at site I which equals the actual mean δ15N value of insects captured by Dionaea on that site. The estimated values of insect δ15N increased for sites II and III up to 1.71‰, consistent with the expectation based on the hypothesis that at these sites nitrogen not originating from the soil is obtained mostly from smaller insects.

Fig. 5.

Relationship of δ15N values of closed leaves (A), storage organ (B), and roots (C) with δ15N values of traps of plants growing at sites I and II. Short‐dashed lines represent regressions for site III. Relationship of δ15N values of traps (D) and storage organ (E) with δ15N values of flowers stalks for plants collected at site I and II. All regressions are significant (P<0.01). Long‐dashed lines represent the 95% confidence interval.

Fig. 5.

Relationship of δ15N values of closed leaves (A), storage organ (B), and roots (C) with δ15N values of traps of plants growing at sites I and II. Short‐dashed lines represent regressions for site III. Relationship of δ15N values of traps (D) and storage organ (E) with δ15N values of flowers stalks for plants collected at site I and II. All regressions are significant (P<0.01). Long‐dashed lines represent the 95% confidence interval.

Table 3.

Summary of δ15N values, percentage of insect‐derived nitrogen, and calculated δ15N values for captured insects for large plants (more than 2 mg total N content) at the three different sites

Values are listed as means±standard deviation. Letters indicate significant differences between sites.


 
Site I
 
Site II
 
Site III
 
% Insect N 80±3.8 a 70±6.8 b 46±11.3 c 
Total δ150.71±0.94 a 2.69±0.91 b 2.14±0.64 b 
Calculated δ15N (‰) 1.11 1.65 1.71 
Measured δ15Ninsect (‰) 1.20±4.1 n.d. n.d. 

 
Site I
 
Site II
 
Site III
 
% Insect N 80±3.8 a 70±6.8 b 46±11.3 c 
Total δ150.71±0.94 a 2.69±0.91 b 2.14±0.64 b 
Calculated δ15N (‰) 1.11 1.65 1.71 
Measured δ15Ninsect (‰) 1.20±4.1 n.d. n.d. 

Discussion

The available methods for calculating the fraction of insect‐derived N (Shearer and Kohl, 1988) are not applicable under the conditions in which insect δ15N values exceed the range found in soils and reference plants. This study presents a new approach for the analysis of pools containing nitrogen obtained from different sources with different isotopic composition.

The analysis is based on the assumption that the slope of linear regressions between sources and sinks can be used as transfer coefficient. However, this approach is highly sensitive to fractionation that might occur between sources and sinks, and thus the parameters of this transfer must be carefully inspected. In the case of insect capture by Dionaea, a very close relationship between the captured insect and the δ15N value of the leaf (which is in this case the petiole) was observed, but the slope of this relationship was only 0.3 (Fig. 3B). Low scatter in the relationship indicates an active uptake process, but the low slope shows that fractionation may occur during uptake (i.e. not the whole prey is being digested) and that the δ15N value of the leaf is not only determined by the prey but also by transfer of nitrogen from the plant during leaf development, and possibly by nitrogen uptake from the soil. Leaves that captured insects with high δ15N values (mainly ants) had a systematically lower δ15N value than the prey. In contrast, in leaves with low δ15N values of the prey, the leaf had higher δ15N values (Fig. 3B). Furthermore, the relationship between leaf δ15N and insect δ15N values is weakened because the leaf is not the digesting organ but the trap assimilates insect N. Relationships with δ15N of leaves (petioles) were used instead of traps due to contamination of the traps by captured insects and the traps disintegrated after digestion of the prey.

Unity in the relationship between the δ15N value of old and new leaves (Fig. 5A) indicates that the source of nitrogen for the construction of new traps is the insect‐catching leaves. The 1:1 relationship indicates that no isotope fractionation exists during phloem transport (Yoneyama et al., 1997), and that possible fractionations during the formation of certain metabolites cancel when analysing the whole leaf. The lack of fractionation during partitioning is also caused by the fact that very little N is transported out of a sink leaf (Marschner, 1998), and that in Dionaea a trap catches an insect only once (Juniper et al., 1989). The trap and the leaf are shed after digestion and the nitrogen of the source leaf is almost totally reallocated to the plant. The slope of the relationship between δ15N values of source leaves and storage organ (Fig. 5B) and roots (Fig. 5C) becomes increasingly shallower, and in this case, it is assumed that N from other sources (i.e. from the soil) is added to the nitrogen pool. Fractionation cannot be excluded in the allocation to storage nitrogen, which is a variable pool of nitrogen that is never exploited. In this case the estimate of insect N would probably be too low rather than too high. In roots, the situation is similar to that in leaves, i.e. it appears to be a one way flux of N during construction and then an additional influx from soil, which would result in the observed difference between δ15N values of source leaves and roots. Given the uncertainty of such a model prediction, verification of the estimates is necessary.

Such verification relies on the assumption that if the predicted contribution of insect nitrogen is used as partitioning coefficient, it is possible, as a bulk estimate, to back‐calculate (equation 3) the average δ15N value of the captured insects. The result shows a non‐significant difference between measured and calculated values at site I (where most data are available). Thus it appears that the chosen approach is useful under conditions where nitrogen pools are composed of a nitrogen mixture from sources with different isotopic composition. The calculated δ15N values of average prey increased at sites II and III, as would be expected from the likely change in diet towards smaller insects of Dionaea growing below a canopy of competing plants.

Based on this analysis, it is shown that the amount of insect‐derived N (Table 3) decreases from site I (c. 2 years after burning) to site III (c. 30 years after burning) despite the fact that nitrogen concentration in capturing leaves increases (Table 1). It appears that Dionaea, a rosette plant, is likely to capture large prey primarily in open habitats, where the trap is unobstructed. These results suggest that, as succession proceeds after fire, neighbouring plants, especially graminoids, overtop the rosette of Dionaea. This decreases the likelihood of capturing large insects, such as grasshoppers. A change in the diet of Dionaea at different stages following fire has been reported (Gibson, 1983), with shaded plants capturing mostly small insects of high δ15N (ants and spiders), while plants in open habitats capturing mostly grasshoppers are characterized by low δ15N. The differences in diet of unshaded and shaded plants is clearly reflected in the total δ15N value between sites (Table 3). Also the carbon assimilation by the traps and petioles of Dionaea is strongly reduced in the more shady environment of late successional stages, as demonstrated by a cost–benefit analysis of carnivory in sunny versus shady habitats (Givnish et al., 1984). The high leaf nitrogen concentration at site III, especially in small plants, can be interpreted as a result of carbohydrate limitation (Schulze et al., 1994; Schulze and Stitt, 1994).

In addition to the effects of competition with the surrounding vegetation, there appears to be a threshold of nitrogen supply during plant development. Small plants have higher δ15N values than large plants, which in turn indicates a difference in diet of small plants (ants) versus large plants (grasshoppers). These small plants tend to have a higher δ15N value than the soil, which indicates that small plants are indeed dependent on nitrogen from insect capture and not from the soil. The total amount of nitrogen that can be gathered from small insects, such as ants, is low. Plants remain small until they have grown 5 to 6 leaves (site I), at which stage they exhibit a large increase in size. It was found that the average trap size per plant increases greatly once the plant has more than six leaves, leading to a change in the plant's diet. This threshold of size is interpreted as the stage where traps become large and strong enough to catch heavy insects such as grasshoppers that, when captured, suddenly provide a large amount of N to the plant. This may increase the growth rate of plants, leading to a large increase in size and the production of subsequent large traps. Because nitrogen concentration in leaves does not decrease with increasing plant size beyond this stage, more leaves of the larger trap size result in an increase in both plant biomass and nitrogen content. At the same time, the total δ15N value of these plants decreases, reflecting the impact of capturing larger insects. It is unclear whether the increase in trap size is genetically or environmentally controlled, or influenced by the interaction between genetic and environmental factors.

Investigations designed at individual plant level might not infer population responses, but can be used to generate hypothetical relationships as subjects for further investigations. The following hypothetical sequence of events is proposed as being consistent with the results. Although the supply of ammonium or nitrate may increase immediately after fire, leaching and mycorrhizal competitors deplete soil nitrogen rapidly, and soil nitrogen supply remains low. Emerging seedlings of Dionaea with small traps capture small ground dwelling insects (i.e. ants). Their survival and success of reaching flowering stage appears to depend on capturing larger insects with larger traps, which occur at approximately the 6th leaf stage. This supplies sufficient nitrogen to produce a burst in growth of yet another large trap, until flowering occurs. Formation of a large flower stalk and seeds may require most of the nitrogen in Dionaea plants. Nitrogen is withdrawn from leaves, which are subsequently shed (Roberts and Oosting, 1958). With the seeds not present in flowering plants, their nitrogen pools are reduced relative to other large plants. The growth of new leaves after seed formation is supported from the storage organ (Fig. 4D), and progressively more from soil N at later stages of succession. As succession proceeds, the rosette becomes increasingly shaded and the limitation to carbohydrate availability from photosynthesis increases (Givnish et al., 1984) relative to nitrogen availability from the capture of ground insects and uptake of soil nitrogen. Ultimately, carbohydrates may be insufficient to produce large traps, resulting in a complete halt to the capturing of large insects, and despite nitrogen supply from the soil, a combined carbon–nitrogen limitation makes Dionaea seedlings unable to compete.

Among insectivorous plants, Dionaea appears to be the most dependent on insect N (maximum 92%; for other species see Schulze and Schulze, 1990; Schulze et al., 1991, 1997). The modified leaves of Dionaea for insect capture allow a large population increase shortly after fire. However, the growth form of a rosette precludes a successful competition with graminoids, which in time establish an overtopping canopy thereby reducing both light and the likelihood of Dionaea capturing large insects.

4

To whom correspondence should be addressed. Fax: +49 7071 293287. E‐mail: Waltraud.Schulze@zmpb.uni‐tuebingen.de

We thank Willi Brand for excellent δ15N and δ13C analysis of insects and plant material. WS thanks the Studienstiftung des deutschen Volkes for traveling support.

References

Ellis AG, Midgley JJ.
1996
. A new plant–animal mutualism involving a plant with sticky leaves and a resident hemipteran insect.
Oecologia
 
106
,
478
–481.
Ellis J.
1770
. Directions for bringing over seeds and plants from the East Indies, and other distant countries in a state of vegetation. To which is added, the figure and botanical description of a new plant, Dionaea muscipula or Venus's Flytrap. London: L. Davis.
Fagerberg WR, Howe DG.
1996
. A quantitative study of tissue dynamics in Venus's flytrap Dionaea muscipula (Droseraceae). 2. Trap reopening.
American Journal of Botany
 
83
,
836
–842.
Gibson TC.
1983
. Competition, disturbance and the carnivorous plant community in south eastern US. PhD thesis, Department of Biology, University of Utah.
Givnish TJ, Burkhardt EL, Happel RE, Weintraub JD.
1984
. Carnivory in the bromeliad Brocchinia reducta, with a cost/benefit model for the general restriction of carnivorous plants to sunny, moist, nutrient‐poor habitats.
The American Naturalist
 
124
,
479
–497.
Greenaway W, May J, English S, Whatley FR.
1990
. Metabolism of 15N‐glycine and 13C‐glycine by Dionaea muscipula studies by gas chromatography‐mass spectrometry.
New Phytologist
 
114
,
581
–588.
Juniper BE, Robins RJ, Joel DM.
1989
.
The carnivorous plants
 . London: Academic Press.
Lee‐Thorp JA, Sealy JC, van der Merwe NJ.
1989
. Stable carbon isotope ratio differences between bone collagen and bone apatite, and their relationship to diet.
Journal of Archaeological Science
 
16
,
585
–599.
Marschner H.
1998
.
Mineral nutrition of higher plants
 . London: Academic Press.
Nelson R, Hamilton D, Wishart J, Marriott CA, Boag B, Handley L, Scrimgeour CM, McNicol JW, Robinson D.
1998
. Stable isotope natural abundances of soil, plants and soil invertebrates in an upland pasture.
Soil Biology and Biochemistry
 
30
,
1773
–1782.
Rea PA.
1984
. Evidence for the H+‐co‐transport of d‐alanine by the digestive glands of Dionaea muscipula Ellis.
Plant, Cell and Environment
 
7
,
363
–366.
Roberts PR, Oosting HJ.
1958
. Responses of venus fly trap (Dionaea muscipula) to factors involved in its endemism.
Ecological Monographs
 
28
,
193
–218.
Robins RJ, Juniper BE.
1980
. The secretory cycles of Dionaea muscipula Ellis. I. The fine structure and the effect of stimulation on the fine structure of the digestive gland cells.
New Phytologist
 
86
,
279
–296.
Schulze ED, Gebauer G, Schulze W, Pate JS.
1991
. The utilization of nitrogen from insect capture by different growth forms of Drosera from Southwest Australia.
Oecologia
 
87
,
240
–246.
Schulze ED, Stitt M.
1994
. Plant growth, storage and resource allocation: from flux control in a metabolic chain to the whole‐plant level. Schulze ED, ed.
Flux control in biological systems
 . San Diego, Academic Press:
57
–118.
Schulze W, Frommer WB, Ward JM.
1999
. Transporters for ammonium, amino acids and peptides are expressed in pitchers of the carnivorous plant Nepenthes.
The Plant Journal
 
17
,
637
–646.
Schulze W, Schulze ED.
1990
. Insect capture and growth of the insectivorous Drosera rotundifolia.
Oecologia
 
82
,
427
–429.
Schulze W, Schulze 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
,
464
–471.
Schulze W, Schulze ED, Stader J, Heilmeier H, Stitt M, Mooney HA.
1994
. Growth and reproduction of Arabidopsis thaliana in relation to storage of starch and nitrate in the wild‐type and in starch‐deficient and nitrate‐uptake‐deficient mutants.
Plant, Cell and Environment
 
17
,
795
–809.
Shearer G, Kohl DH.
1988
. Estimates of N2 fixation in ecosystems: the need for and basis of the 15N natural abundance method. Rundel PW, Ehleringer JR, Nagy KA, eds.
Stable isotopes in ecological research
 . Berlin: Springer Verlag,
68
,
342
–375.
Trebacz K, Busch MB, Hejnowicz Z, Sievers A.
1996
. Cyclopiazonic acid disturbs the regulation of cytosolic calcium when repetitive action potentials are evoked in Dionaea traps.
Planta
 
198
,
623
–626.
Trebacz K, Sievers A.
1998
. Action potentials evoked by light in traps of Dionaea muscipula Ellis.
Plant and Cell Physiology
 
39
,
369
–372.
Yoneyama T, Handley LL, Scrimgeour CM, Fisher DB, Raven JA.
1997
. Variations of the natural abundances of nitrogen and carbon isotopes in Triticum aestivum L. with special references to phloem and xylem exudates.
New Phytologist
 
137
,
205
–213.

Comments

0 Comments