Home | MITC | MonarchLab | MLMP | SEE | Monarch Store | Links | About Us | Contact Us

Do Monarch Larvae use Visual or Chemical Stimuli to find Milkweed?

Art Evans, Leland Goertz, Bethany Leach, Martin Phillips, Justin Sanderson, and Stephanie Spurgat
Texas Military Institute, San Antonio, TX

Abstract | Introduction | Methods | Results | Discussion | Acknowledgements | Literature Cited | Research Projects

Abstract

The purpose of our experiment was to examine how a caterpillar (4th or 5th instar) finds milkweed in a field with other plants and scents. We tested monarch larval use of visual, chemical, or both stimuli to find milkweed. We built a Terrarri-quad to provide four options for the larvae's sensory perception: (1) scent only, (2) visual only, (3) both scent and visual, or (4) neither visual nor scent. Each caterpillar was individually placed facing a pre-determined direction in the middle of the Terrarri-quad. Seventy-nine larvae were tested with 30 not going to any specific quadrant, 29 to the chemical stimulus only, 5 to visual only, 14 to the quadrant with both stimuli, and 1 to the quadrant with the control plant, Salvia. Our data suggest that monarch larvae primarily use chemical stimuli to find milkweed plants.

Introduction

At our school, milkweed plants are sparsely distributed, with only about 0.008 plants per square meter. Monarch larvae have small eyes that are very similar to a single adult ommatidium and that allow the larvae to see ultraviolet light and colors (Scott 1986). These eyes may help larvae find their hostplants, but olfactory strucutures may also be important in locating milkweed by allowing larvae to detect chemicals emitted from plants (Scott 1986). If this is so, which cues do they use more? We were wondering how the larvae get from their host plant to another milkweed plant. We produced four hypotheses.

Hypotheses

• H_al: Monarch caterpillars use visual stimuli to find milkweed.
• H_a2: Monarch caterpillars use chemical stimuli to find milkweed.
• H₀₁: Monarch caterpillars use both chemical and visual stimuli to find milkweed.
• H₀₂: Monarch caterpillars use neither chemical nor visual stimuli to find milkweed.

Methods

We built an open cage called a Terrarri-quad, and we kept it open so that the larvae could have a more realistic environment. In their real environment, milkweed plants emit chemicals into the air. If we had enclosed our cage, the scents from the plant would have been stronger than normal. We built this cage out of Plexiglas. The Terrarri-quad had four quadrants and was in the shape of a cross (Figure 1). All quadrants were one foot in length, height, and width.

The first quadrant was separated from the rest of the Terrarri-quad by a brown wall to prevent the larvae from seeing the milkweed plant. The wall had small holes in it to allow the plant chemicals to pass through, and a long short opening at the bottom to allow the larvae to crawl underneath the wall into the quadrant (Figure 2). To prevent larvae from seeing the plant through the bottom opening, we built up a mound of dirt that stretched from one side of the quadrant to the other. This quadrant presented larvae with chemical stimuli only.

(a)
(b)

Figure 1. The Terrari-quad (a: photo, b: diagram), our experimental cage, contained four quadrants from which the larvae could choose.

Figure 2. An opaque wall with small holes separated quadrant 1 (chemical only) from the rest of the Terrari-quad.

In quadrant two, which contained the salvia, we had no wall; we left it open. The salvia was our control plant. We had a control plant because fields have more kinds of plants that just milkweed. It is possible that larvae use neither visual nor chemical cues to find milkweed plants. They might move from plant to plant and need to actually make contact with a plant to determine whether it’s milkweed. If monarch larvae use neither visual nor chemical cues, they should go to a non-milkweed plant as often as to a milkweed plant. Quadrant 2 presented larvae with a plant that gave neither visual nor chemical cues associated with milkweed.

In quadrant three, we had a clear wall and ceiling separating milkweed from the rest of the Terrarri-quad. The clear wall allowed larvae to see, but not smell, the milkweed. This way, the larvae could use their visual stimuli only to find the milkweed.

In quadrant four, we had no wall separating a milkweed plant from the rest of the Terrari-quad. We left it open so that the larvae could use both their visual and chemical stimuli to find the milkweed stalk.

We furnished the Terrarri-quad with a layer of potting soil, and we put fresh milkweed for each day of trials in quadrants one, three, and four. We put salvia, our control plant, in quadrant two. The plants were collected from our school's butterfly garden. When bringing the salvia and milkweed stalks into our lab, we made sure they did not touch; otherwise, the scents of the plants would rub onto each other and confuse the results. One caterpillar was tested per trial. For each trial, a caterpillar was placed in the middle of the Terrarri-quad. A trial lasted ten minutes. We monitored the larva's every move. We kept the time going even if the larvae went into several different quadrans during one 10-minute period. At the end of each trial, we recorded which quadrant the caterpillar was in. If it had crawled up the walls of either quadrant one or quadrant three, we counted it as being in that specific quadrant. If the larva did not go to any specific quadrant, it was counted as "no choice." We tested 79 larvae total. We recorded what we observed on our pre-determined data sheet (Figure 3).

We received three mated female adult monarchs from the University of Minnesota and collected about 100 eggs from them. We fed the hatched larvae Asclepias currasavica, and used the same kind of milkweed in our experiment. We proceeded to raise the larvae that we had. When the larvae reached the 4th or 5th instar stage, we tested them using the methods described above. With each trial we placed the monarch caterpillar in a pre-determined direction in the middle of the Terrarri-quad. We determined the direction by placing the larvae toward a different quadrant for each trial. After testing the larvae, we dove into statistical testing. We used the Chi-square test to analyze our data.

Figure 3: Sample data collected using predetermined data sheets

Results

Of the 79 complete observations, 30 of the larvae did not make a choice because they remained or ended up in the middle. Of those that made a choice, 29 went to quadrant one, 1 went to quadrant two, 5 went to quadrant three, and 14 went to quadrant four (Figure 4). There was a significant difference between the quadrants in the number of larvae that chose each one (χ ² = 37.77, 3 d.f., p < 0.005). More larvae than expected went to Quadrant 1 (chemical only) while fewer than expected went to Quadrants 2 (neither cue) and 3 (visual only). These data suggest that monarch larvae primarily use chemical stimuli to find milkweed. We observed that an overwhelming amount of larvae (around 59%) went to the chemical stimuli quadrant. Very few went to the control quadrant, which held the control plant salvia. About 10% went to the visual quadrant, and 29% went to the quadrant with both visual and chemical stimuli.

Figure 4: Summary of larval destinations. The numbers shown correlate with the quadrants depicted in Figure 1.
Quadrant 1, chemical; Quadrant 2, control; Quadrant 3, visual; Quadrant 4; visual and chemical.

Besides this information, we had additional questions. One of those questions is "Is there a correlation between when the larvae were fed and if they made a choice?" We fed larvae twice each day, in the morning and early afternoon. Although we did not plan to feed larvae at different intervals before testing them, our schedules were such that about half of the larvae were tested in the morning and half in the afternoon. We did not feed larvae immediately before testing them, so those tested in the morning had been fed the afternoon before (about 16 hours prior) and those tested in the afternoon had been fed in the morning of the same day (about 8 hours prior). We thought that the larvae fed 16 hours earlier may be hungrier than the larvae fed 8 hours earlier and therefore more likely to reach a quadrant faster. However, we found no correlation between when they were fed and whether they made a choice (χ ² = 1.91, 1 d.f., p > 0.1) (Figure 5).

Figure 5: Observed number of larvae entering various quadrants after 8 or 16 hours of feeding.

The last question we tested was "Is there a correlation between the way the caterpillars faced and the way they went?" We found that there was no correlation between the way the caterpillars faced and the way they went (χ ² = 7.64, 9 d.f., p >0.1) (Table 1). This means that the larvae actually chose which direction they went, and we did not determine it by facing them in a particular direction at the start of each trial.

Table 1: Summary of number of larvae found in various quadrants and the direction they faced.

Discussion

Many of the larvae we tested remained in the middle portion of the Terrarri-quad during the ten- minute period. We were not sure why the larvae did not make a choice. It may have been because they were not very hungry, or maybe they were ready to molt at the time of the experiment. The larvae could have possibly been a little "stunned" by the new environment. Also, we do not know why most of the larvae that made a choice went to the 1^st quadrant, which was the chemical only quadrant. It may have been because they were attracted to the opaque wall of the 1^st quadrant. It could have been because chemical cues are the most important, but that leaves the question of why more larvae went to the quadrant with only chemical cues than the quadrant with both chemical and visual cues. Although fewer larvae went to quadrant 4 (both visual and chemical cues) than to the chemical cue only quadrant, more chose the quadrant with both cues than the quadrants with either visual only or neither cue. These data suggest that while larvae might have been attracted to the opaque wall of the chemical only quadrant, there were important differences between the quadrants with and without chemical cues. This supports the hypothesis that monarch larvae use mainly chemical stimuli to find milkweed.

In conclusion, we have learned that monarch larvae probably use chemical stimuli as their main method of finding milkweed plants. We think they use chemical stimuli because the larvae’s eyes are not developmentally advanced enough to create a complete image; therefore, they must rely on chemical cues.

Acknowledgments

The Texas Military Institute MS Monarch Team would like to thank:

• The National Science Foundation for providin a grant that allowed us to study Danaus plexippus
• Michelle Solensky for helping us with our statistical analysis and other experimental difficulties
• Karen Oberhauser for running the Monarch Workshops
• Bill Calvert for co-running the Monarch Workshops
• Leland Goertz and Art Evans for being our sponsors and building the Terrari-quad
• Mary Kennedy for helping us raise the larvae and providing the opportunity for us to learn science
• Allison Mifflin, Adrianne Borrego, Kathryn Willingham, Sarah Vellom, and Jennifer Bowes for helping us raise the larvae
• Sandy Rodie for creating a garden with milkweed in it
• Texas Military Institute for allowing us to study monarch butterflies

Literature Cited

Scott, James A. 1986. The Butterflies of North America. Stanford, CA: Stanford University Press, pp. 38-41.

Back to top | Interactions with Milkweed | Student/Teacher Research | Research Topics

MonarchLab Home | Research | Biology | Monarch Rearing | Conservation | Ask the Expert