The Effects of Soil Nitrogen Content on the Caloric Investments of the Carnivorous Plant Drosera capenis

Part of the Young Naturalist Awards Curriculum Collection.

by Melody, Grade 11, Massachusetts - 2013 YNA Winner

Background Information

Since their discovery, carnivorous plants have fascinated amateurs and scientists alike. In the late 1800s, Charles Darwin was a pioneer in identifying and studying these unique plants, which he studied for many years before publishing a book of his findings, Insectivorous Plants (Darwin 1875). Carnivory in plants is generally accepted as an adaptation to a specific environment in which nutrients are scarce but sunlight and moisture are abundant (Givnish 1984). There are approximately 630 species of known carnivorous plants, most of which are found in this type of habitat (Barthlott 2007).

Among the best-known varieties of carnivorous plants are the sundews, which belong to the genus Drosera. Most species are perennial herbs and share a unique method of prey capture. The leaves possess specialized structures, referred to as glandular trichomes, which consist of an elongated stalk with a glandular tip that secretes a sticky, polysaccharide-rich mucilage. (Barthlott 2007). These trichomes have a red pigmentation that attracts insects, which are caught in the mucilage when they land on the leaf. As the insect struggles to escape, the trichomes respond by releasing more mucilage until the insect is consumed by it. Then the trichomes turn inward, pushing the insect to the horizontal center of the leaf. This response by the trichomes is activated by mechanical stimulus and is relatively rapid, starting as soon as a few seconds after stimulation and enduring for a few minutes. The response is considered to be nastic because the direction of the response is always toward the center of the leaf, independent of the location of the stimulation (Williams 1976).

Figure 1. A photograph of a leaf typically found on Drosera capensis. The flat surface is cover by glandular trichomes, each topped a small droplet mucilage.

Once the prey is suffocated and secure in the center of the leaf, slower and more dramatic movements take place. These movements are characterized as chemotrophic because they are triggered by the presence of certain chemicals, and their direction depends on the location of the stimulus. Essentially, those trichomes surrounding the stimulated trichomes respond to the chemical stimulus by moving toward the stimulated trichomes. The sessile glands of the leaf that surround the prey secrete a mixture of digestive enzymes that digest the prey’s soft organs. Once the prey is digested, the chemical stimulus ceases, and the trichomes return to their original positions and are poised for the next capture (Williams 1976).

It was Darwin who originally proposed that the relatively immediate response of Drosera was mediated by action potentials (Darwin 1875). Although his idea was met with skepticism in his time, contemporary research supports his theory. This mechanism, found in Drosera and other carnivorous plants, resembles nerve impulses in animals. This system of action potentials and their receptors, however, is slow and rudimentary compared to even the most primitive of animals (Williams 1976).

Nutrient-poor habitats with sufficient water and sunlight, such as regions of Australia and southwestern Africa, are the epicenter of carnivorous plant life, which is present throughout the world except for Antarctica. The Cape sundew, or Drosera capensis, is named after its place of origin on the Cape of Good Hope. The sandy, moist soils that border the rivers in this area are a common habitat for these plants (Barthlott 2007).

Carnivory in plants is defined by Givnish (1984) as having two main qualities. First, the plant must be able to absorb nutrients from dead animals. Second, the plant must have one or more structural adaptations that clearly evolved to serve the purpose of prey attraction, capture or digestion. Carnivory is generally accepted as an adaptation to an environment in which conditions are harsh for traditional plants (Ellison 2001, Givnish 1984, Méndez 1999). Nutrient acquisition is the main benefit of carnivory for plants. Most of these unique species grow in habitats with low nutrient content in the soil. This lack of nutrients would inhibit the growth and function of most species of plants. To compensate, the carnivorous plants obtain nutrients such as nitrogen and phosphorus from insect prey (Givnish 1984, Méndez 1999). Carnivorous plants utilize this adaptation to different extents; some species of Drosera obtain as little as 20% of their nitrogen from insect prey, while other species derive up to 90% (Ellison 2001).

Carnivory as a method of nutrient acquisition has major disadvantages, however, because the specialized structures needed for this strategy lower the photosynthetic efficiency of the plant and create a higher demand for carbohydrates (Millet 2012). In fact, a study by Karlsson (1991) estimated that 4% to 6% of the total carbon budget of the plant is used to produce mucilage. The caloric stress on the carnivorous plant is furthered by its relatively low photosynthetic rate compared to traditional plants. Specialized structures decrease the surface area of the carnivorous plant available for photosynthetic activity (Méndez 1999). It is in a nutrient-poor environment with abundant sunlight and water that the advantages of carnivory are the greatest because the nutrient uptake from prey can outweigh the photosynthetic costs (Givnish 1984).

Many studies support the theory that carnivorous plants reduce their investment in carnivory when sufficient levels of nitrogen are available in the soil. This environmental flexibility is advantageous for the plant because it prevents the costs of carnivory from outweighing the benefits (Millet 2012). For instance, Thorén (2003) found that Drosera rotundifolia reduced the concentration of polysaccharides in its mucilage when nitrogen was abundant in the soil. Millet’s 2012 studies showed that the percentage of nitrogen derived from prey, out of the total nitrogen content of the plant, declined in response to nitrogen-rich soils. More surprisingly, he also found that the quantity of nitrogen derived from prey declined, from which he inferred that the plant’s investment in carnivory had declined in response to the level of soil nitrogen. Generally, these studies show that carnivorous plants will invest more energy in nutrient uptake from the soil and less energy in prey capture whenever possible. This seems to increase their chance of survival by eliminating the high caloric costs associated with carnivory.

Karlsson (1991) noticed in his studies on the resource allocation of carnivorous plants that this quick response to environmental change is unique to the genus Drosera. In these studies, Drosera rotundifolia showed a response to the addition of soil nutrients, but the other carnivorous plants in his experiment did not. Most species of plants determine their reproductive characteristics for the year at the start of the flowering period. This was not the case with D. rotundifolia, however; this species seemed more versatile, because changes in soil nutrients resulted in adaptation within only one year. S. Poppinga (personal communication 2012), a co-discoverer of the new species Drosera glanduligera, theorizes that this is an adaptation that allows Drosera to co-evolve with their insect prey. This flexibility is an essential aspect of their survival.

Research Question

How does soil nitrogen content affect the morphology and physiology of the carnivorous plant Drosera capensis?


When the amount of nitrogen in the soil is insufficient for supporting D. capensis, the plant will change its morphology and physiology in a way that will invest more resources in carnivory, thus yielding a higher rate of capture and absorption of prey.


Twenty-five D. capensis were acquired from the Carnivorous Plant Nursery in Derwood, Maryland. The plants were potted in groups of four in a standard plastic bin (31.75 cm by 12.7 cm by 22.52 cm) filled with a one-to-one mixture of sand (99% crystalline silica from Quikrete) and sphagnum peat moss (from Miracle-Gro). Nine holes with radius of 0.635 cm were drilled in the bottom of each bin, and the entire setup was placed into a larger plastic bin (36.51 cm by 29.21 cm by 12.06 cm) filled to one-third of its height with distilled water. The water absorbed by the soil was frequently replaced. Two desk lamps containing standard fluorescent bulbs were set over the plants.

The experiment took place in at ambient room temperature of 20°C. The 24 plants were separated into six groups of four, and preliminary data was collected for each. Each group received one liter of a nutrient solution identical to the solution used in Chandler and Anderson’s 1976 study on Drosera. The solution was slightly altered for each group, however, because substances containing nitrogen—Ca(NO3)2 and (NH4)2SO4—were replaced in certain percentages by other substances—CaCl2 and K2SO4. The percentage of added nitrate ranged from 100% to 0%, with an increment of 20% between each group. The control group, which was administered 100% of the original nitrate quantities, was labeled “F.” Each setting was labeled in alphabetical order based on their declining percentage of added nitrates. The exact quantities used for each setting are shown in Table 1.

Group A B C D E F
Inorganic Salt (g) (g) (g) (g) (g) (g)
(NH4)SO4 0.000 0.002 0.004 0.005 0.007 0.009
K2SO4 0.009 0.007 0.006 0.004 0.002 0.000
Ca(NO3)2 0.000 0.013 0.026 0.040 0.053 0.066
CaCl2 0.066 0.053 0.040 0.026 0.013 0.000
Table 1.Quanitities of nitrate forms and substitutes for each setting.


Figure 2. An image of the experimental setup. Six groups of four plants were arranged on a table as shown. Two desk lamps were set up on either side of the groups to provide controlled lighting.

For six weeks, data was collected every seven days to record the changes caused by varying soil nitrogen content. At the end of this period, nylon netting was set up to surround the system by fastening the edges of the netting to the walls with duct tape. The enclosure was sealed to the table with duct tape as well. Fruit flies (Drosophila melanogaster) were then released into the test area. Data collection continued for six days.

Data Collection Procedure

Color of leaves

Using an eight-megapixel HTC Evo digital camera, I took a picture of each plant. The pictures were uploaded to a computer, where they were analyzed using an RGB color analyzer (Image J).

Figure 3. Taking digital images of each plant with an HTC Evo. Photographs were taken from a top down angle and required a certain sharpness of detail.
Percentage of Leaves Producing Mucilage

The number of mature leaves producing mucilage and the total number of mature leaves on the plant were counted and recorded in the logbook. Leaves were not included in the former measurement if the mucilage drops were not visible to the naked eye, or if the entire leaf was not producing mucilage.

Average Diameter of the Mucilage Drop

A picture was taken of two leaves on each plant using the HTC Evo digital camera aligned with a standard magnification lens (2.0 scale). These images were then uploaded to a computer, where the diameters of three random mucilage drops were measured in pixels using Microsoft Paint. These data were recorded in Excel.

Number of Drosophila Caught
Figure 4. Apparatus used to take images of mucilage drops.  A 2.0 magnification lens was secured tot eh HTC Evo with standard masking tape.

The number of Drosophila caught by each plant was measured daily. This number was an accumulation; therefore, it included any flies caught on previous days. All data were recorded in the logbook.


In terms of qualitative observations, a difference in morphology was visibly apparent in the different soil samples. For example, a difference in the diameter of the mucilage drops between Group A and Group F could be perceived without magnification. It was also observed that red pigmentation accumulated in each individual leaf of a plant as it aged, causing the ration of red to green to increase over time.

    Trail 1 Trail 2 Trail 3 AVG Group AVG STDEV
Group A 1a 0.99 0.94 0.90 0.94 0.90 0.04
2a 0.87 0.84 0.94 0.88
3a 0.93 0.89 0.91 0.91
4a 0.91 0.86 0.87 0.88
Group B 1b 0.90 0.90 0.71 0.84 0.88 0.07
2b 0.89 0.92 1.00 0.94
3b 0.88 0.84 0.87 0.87
4b 0.93 0.88 0.88 0.90
Group C 1c 0.90 0.87 0.80 0.86 0.86 0.04
2c 0.85 0.81 0.86 0.84
3c 0.87 0.92 0.82 0.87
4c 0.87 0.82 0.86 0.85
Group D 1d 0.88 0.91 0.82 0.87 0.89 0.04
2d 0.92 0.87 0.85 0.88
3d 0.95 0.88 0.93 0.92
4d 0.86 0.89 0.95 0.90
Group E 1e 0.85 0.85 0.95 0.89 0.88 0.04
2e 0.90 0.87 0.82 0.86
3e 0.90 0.87 0.84 0.87
4e 0.90 0.85 0.90 0.88
Group F 1f 0.89 0.92 0.86 0.89 0.86 0.03
2f 0.85 0.86 0.90 0.87
3f 0.85 0.85 0.85 0.85
4f 0.89 0.81 0.83 0.85
Table 2. Ratio of re RGB value to green RGB blue at one week of maturity.


  21-Dec 31-Dec 6-Jan 13-Jan 20-Jan 27-Jan AVG Δ % STDEV
  (%) (%) (%) (%) (%) (%) (%) (%) (%)
Group A 32.5 30.9 48.8 69.5 71.3 84.8 56.3 52.3 35.4
Group B 35.4 51.1 60.0 75.4 69.2 81.0 62.0 45.5 8.5
Group C 37.4 43.4 56.8 54.4 62.1 62.6 52.8 25.2 22.0
Group D 40.0 42.8 59.5 59.9 62.3 70.4 55.8 30.4 25.1
Group E 23.6 43.6 36.9 44.0 48.9 56.9 42.3 33.3 33.3
Group F 41.0 46.0 47.8 52.7 56.5 61.2 50.9 20.2 11.9
Table 3. Percentage of leaves producing mucilage on each plant at weekly intervals.
Figure 5. Change in the percent of leaves producing mucilage over a six-week period.  This data represents an average value for each setting and is calculated by subtracting the initial percentage from the final percentage.
  Individual AVG Group AVG STDEV
(pixels) (pixels) (pixels)
Group A 1 55.83 61.88 6.84
2 66.17
3 57.33
4 68.17
Group B 1 57 59.92 5.76
2 61.17
3 65.5
4 56
Group C 1 68.5 62.54 7.66
2 65.17
3 56.33
4 60.17
Group D 1 55.5 60.79 6.9
2 59
3 62.83
4 65.83
Group E 1 51.5 50.54 3.8
2 51.17
3 48.67
4 50.83
Group F 1 48.67 46.17 4.88
2 46.83
3 43.33
4 45.83
Tabel 4. Average diameter in pixels of a single droplet of cucilage for each individual.
Figure 6. A scatterplot of the average diameter of mucilage droplet for each group. The data was fit with moderate strength to polynomial trendline.
    diameter of mucilage % of producing mucilage total mucilage
    AVG Group AVG % AVG AVG*% Group AVG STDEV
GROUP (pixels) (pixels) % % (pixels*%) (pixels*%) (pixels*%)
A 1 55.83 61.88 0.77 0.85 42.95 53.02 13.02
2 66.17 0.90 59.55
3 57.33 0.72 41.41
4 68.17 1.00 68.17
B 1 57.00 59.92 0.76 0.81 43.59 48.57 6.95
2 61.17 0.94 57.77
3 65.50 0.76 50.09
4 56.00 0.76 42.82
C 1 68.50 62.54 0.79 0.63 54.08 39.56 12.7
2 65.17 0.52 34.00
3 56.33 0.44 25.04
4 60.17 0.75 45.13
D 1 55.50 60.79 0.72 0.7 40.08 42.55 6.25
2 59.00 0.80 47.20
3 62.83 0.76 48.05
4 65.83 0.53 34.85
E 1 51.50 50.54 0.60 0.57 30.90 28.99 15.52
2 51.17 0.65 33.11
3 48.67 0.15 7.49
4 50.83 0.88 44.48
F 1 48.67 46.17 0.71 0.61 34.76 28.37 5.05
2 46.83 0.63 29.58
3 43.33 0.53 22.94
4 45.83 0.57 26.19
Table 5. Total mucilage produced by each individual.
Figure 7. Linear model of total mucilage produced by an individual at a certain percentage of nitrogen soil content. Total mucilage produced was calculated by multiplying the average diameter of a mucilage drop by the percentage of leaves producing mucilage on the plant. The model is a moderate representation of the data (R2=0.93).
  day 1 2 3 4 5 6 % AVG
Group A 4 5 8 8 10 15 26.3 8.3
Group B 2 3 3 3 5 6 10.5 3.7
Group C 2 2 2 3 2 10 17.5 3.5
Group D 5 5 5 6 9 15 26.3 7.5
Group E 1 2 2 2 2 5 8.8 2.3
Group F 3 3 4 4 4 6 10.5 4.0
Table 6. Total Drosophila caught by each group of individuals, including percentage of total Drosophila caught.
Figure 8. A bar chart comparing the total flies caught over a period of six day by each group.  No correlation is apparent between soil nitrogen content and the number of flies caught in this amount of time.

Data Analysis and Discussion

Some components of the hypothesis were supported by the data, while others were contradicted. No trend was found in the red-to-green pigmentation of the leaves, indicating that there was no correlation between soil nitrogen content and pigmentation of the leaves for this experiment. The average ratio for each group showed no trends related to the soil nitrogen content. Also, a single-factor ANOVA comparing every setting showed that any difference between the group averages was not statistically significant (P = 0.082). This result does not support the hypothesis that the carbon investment of an individual plant will change depending on the amount of nitrogen present in the soil.

The percentage of leaves producing mucilage showed a trend over time for each level of soil nitrogen. As shown in Figure 4, the change in mucilage production is inversely proportional to the nitrogen content in the soil. However, it was found using a single-factor ANOVA that the correlation is not statistically significant (P = 0.291). The data does not reflect any correlation between mucilage production and the nitrogen content of the soil. Ultimately, the hypothesis was not supported by the data.

The average diameter of a mucilage droplet at different levels of soil nitrogen was compared using a single-factor ANOVA. As shown in Figure 5, the average diameter seems to increase as the nitrogen content of the soil goes down until a maximum limit is reached. It was found that the correlation was highly significant (P = 1.038 x 10-20), allowing the null hypothesis to be rejected. Figure 5 displays a scatter plot of the average values for each group fit to a polynomial curve. This trend line models the data with moderate strength (R2 = 0.912) and has logical limits. As the nitrogen content in the soil declines, the plant is stimulated to produce more mucilage up to a certain point, beyond which the severe lack of nutrients reduces the rate of photosynthesis and, consequently, the carbon availability for mucilage production.  

In Table 5, the average diameter of a mucilage drop for each individual plant was multiplied by the percentage of leaves producing mucilage, thus producing the effective total mucilage production of the plant. In Figure 6, these data were graphed on a scatter plot and fit to a linear model of moderate strength (R2 = 0.910). The calculated values were also compared using a single-factor ANOVA test, which indicated a statistical significance to the data (P = 0.021). Because this data is both statistically significant and fits a linear trend, it provides support for the hypothesis that carbon investment is correlated with the nitrogen content of the soil.

Although it was found that the nitrogen content of the soil can have an effect on a plant’s carbon investment in carnivory, no corresponding trend was found in the number of Drosophila caught by each plant. Differences between the prey-catch averages of each group were not found to be significant (P = 0.058), and no clear trend can be seen in the final calculations (Figure 7). This contradicts the second component to my hypothesis; an increased investment in carnivory did not necessarily lead to a higher rate of prey capture in this experiment.

One particular factor in this experiment could have affected the data for the dependent variables. The plant labeled “3C,” meaning that it was the third member of the group that received 40% nitrogen in the soil, started to develop a flowering stalk in the midst of experimentation. Because reproduction is a substantial carbon sink for the plant, it was likely that the data for this plant was highly skewed compared to the vegetative individuals. This factor could explain the slight offset in the values for Group C from any observed trend. For example, the 40% nitrogen setting in Figure 4 breaks the trend by being much lower than the expected value.


Ultimately, the hypothesis was only partially supported by the data. Soil nitrogen content was found to have no effect on the red pigmentation of the plant, regardless of its age. Conversely, the average diameter of the mucilage droplets found on individual plants was found to have a strong correlation with the amount of nitrogen in the soil, both in statistical significance and polynomial trend. The parabolic shape of the model reflects the quality being described very well; a maximum limit is reached at some intensity of nitrogen deficiency, and every quantity after this limit will decline as the plants lose photosynthetic efficiency. This relationship was analyzed further by calculating the effective total mucilage produced by the individual plants at each level of soil nitrogen, a correlation that was also found to be significant. Instead of fitting a polynomial function, however, these data were best modeled in a linear fashion. Despite this clear change in carbon investment for each level of soil nitrogen, no equivalent trends were found in the number of captured Drosophila for any of the groups. The group that received 0% nitrogen was the one to capture the most flies, which is expected based on the hypothesis; nevertheless, the other five settings captured flies in a seemingly random fashion. It is possible that mucilage production has no bearing on the capture of insect prey, and that the abundance of flies captured by Group A was merely a coincidence; however, it is equally likely that the proximity of Group A and Group D to the light fixtures gave them an advantage over differently placed groups. Because of this highly significant source of error, these data cannot be justly compared to the hypothesis.

Limitations and Sources of Error

To ensure that the nitrogen content of the soil was solely responsible for any differences between individuals, numerous variables were controlled in the experiment. Standard factors in the maintenance of the plants, such as light, temperature, and water availability, were controlled. Along with these, the contents of the growth medium were also controlled to verify that only differences in nitrogen availability were present between each group. Various factors in data collection were also controlled, such as the insect prey availability for each plant.

Although attempts were made to keep all variables and procedures consistent, some sources of error and limitations can be identified in this experiment. The duration of the experiment was the major limitation of this experiment because it limited the extent of the changes seen in the characteristics of the plant. Not every plant was of the same size and vitality at the beginning of the experiment, potentially causing some inconsistencies in carbon investment. Plant vitality could also be altered slightly by differing quantities of nutrients added to the soil. For instance, the quantity of each inorganic salt was measured using a digital scale; a more sensitive device would have been used if it were available. Other sources of error include a slight fluctuation in temperature and photoperiod from day to day. Some sources of error were specific to a certain dependent variable. For example, the diameter of the mucilage droplets could have been altered by a slight change in scale of the picture. Also, the ratio of red to green RGB values could have been affected by lighting. Not all photographs were taken in the same weather, thus changing the lighting from week to week of data collection. The most significant source of error in the experiment was the differing proximity of each group to the lights. As seen in Figure 1, Groups A and D were closer to the desk lamps than the other four groups. It is highly likely that the natural phototaxis ofDrosophilaled them to fly toward these lights, which would explain the unusually high quantity of prey captured by these two groups (Figure 7).

Applications and Future Extensions

The adaptive capabilities of Drosera require further investigation and could lead to significant discoveries in the biology of carnivorous plants. A similar experiment, in which the plants are grown from seeds instead of being introduced to the environment as adult plants, could result in more extensive differences between groups. In future studies, a stronger focus could be put on the vitality and success of the plant, which could provide insight into the types of environments carnivorous plants thrive in. To understand this, a competitive variable must be added to the experiment to better mimic natural conditions, and variables such as the total dry mass of the plant could be added. These experiments may also reveal the limits of the carbon/nutrient balance, in which individuals can no longer survive due to overcompetition or nutrient deficiency. In any case, a long-term study would yield more distinct and significant results.

Most importantly, this study supports an important cost/benefit model used to explain the evolution of all carnivorous plants. If these changes in morphology and physiology are truly caused by nitrogen availability, then a balance maintained between nutrient acquisition and carbon investment is supported. Understanding the adaptive capabilities of Drosera may also provide insights into the detailed mechanisms of phenotypic plasticity, which is highly prominent in the genus. Analysis of the Drosera genome may reveal the cause of this unique characteristic and help scientists identify these same traits in traditional plants. Phenotypic plasticity could be an important trait in light of global climate change because individuals that can adapt to a changing environment within their own generation may be more successful than others that cannot. Moreover, the similarities between animals and Drosera could lead to some breakthroughs in understanding how the most primitive animal systems evolved. For example, the use of action potentials in Drosera capensis, which lack any specialized cells comparable to nerves, could potentially be similar to invertebrate animals that developed a primitive nervous system.


Barthlott, W., and M. Ashdown. The Curious World of Carnivorous Plants: A Comprehensive Guide to Their Biology and Cultivation. Portland, OR: Timber Press, 2007.

Chandler, G.E., and J.W. Anderson. “Studies on the Nutrition and Growth of Drosera Species with Reference to the Carnivorous Habit.” New Phytologist, 76 (1976): 129-141.

Darwin, Charles. Insectivorous Plants. London: Murray, 1875.

Ellison, A.M., and N.J. Gotelli. “Evolutionary Ecology of Carnivorous Plants.” Trends in Ecology and Evolution 16.11 (2001): 623-629.

Givnish, T.J., E.L. Burkhardt, R.E. Happel, and J.D. Weintraub. “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 124.4 (1984): 479-497. Retrieved from the World Wide Web at

Karlsson, P.S. “The effect of soil nutrient status on prey utilization in four carnivorous plants.” Oecologia 86 (199): 1-7.

Méndez, M., and P.S. Karlsson. “Costs and benefits of carnivory in plants: Insights from the photosynthetic performance of four carnivorous plants in a subarctic environment.”Oikos86.1 (1999): 105-112. Retrieved from the World Wide Web at

Millet, J., J.I. Roger, and J. Waldron. “The contribution of insect prey to the total nitrogen content of sundews determined in situ by stable isotope analysis.” New Phytologist 158 (2003): 527-534.

Millett, J., B.M. Svensson, J. Newton, and H. Rydin. “Reliance on prey-derived nitrogen by the carnivorous plant Drosera rotundifolia decreases with increasing nitrogen deposition.” New Phytologist 195 (2012): 182-188.

Poppinga, S., S. Hartmeyer, R. Seidel, and T. Masselter. “Catapulting Tentacles in a Sticky Carnivorous Plant.” PLoS One 7.9 (2012).

Thorén, L.M., J. Tuomi, T. Kämäräinen, and K. Laine. “Resource availability affects investment in carnivory in Drosera rotundifolia. New Phytologist 159 (2003): 507-511.

Williams, S.E. “The Evolution of Plant Sensory Systems.” Proceeding of the American Philosophical Society 120.3 (1976): 188-202.