Cory J. Wagner, Roopa Kamesh, and Randy Wilson
Miami University
Oxford, Ohio
Abstract: Cecropia trees have long been thought to have a mutualistic relationship with ants of the Azteca genus. The ants benefit from this relationship by receiving shelter and food in the form of mullarian bodies. The trees benefit from the association by receiving nitrogen from the dead ants and their byproducts and from the removal of climbing vines by the ants. Few studies have examined whether the Azteca ants actually protect the Cecropia trees against herbivory from other organisms. In this study, we performed an herbivory study between Cecropia trees in the cloud rainforest of Monteverde and the low-land rainforest of Corcovado in Costa Rica. It was thought that the trees in Corcovado would show less damage than the trees in Monteverde as a result of the protective function of the Azteca ants. The ants were known not to be present in the high altitude of Monteverde. The study found that the trees in Corcovado had showed greater herbivory than the trees in Monteverde (Chi Square test at the 95% C.L. with a P=.023) but that the damage was also less severe than that of Monteverde overall. No trees surveyed were found to have Azteca ant colonies present so the study serves as a framework and a baseline for future herbivory studies regarding the Azteca ant defense of Cecropia trees in Monteverde and Corcovado.
Introduction
There are many different ways in which organisms interact with each other as they carry out their life cycles. Some organisms develop specific and specialized associations with each other. These associations include amensalisms, commensalisms, parasitisms, and mutualisms. The latter is particularly interesting because the interaction benefits both or all the organisms involved in the mutualism (Purves et al. 1995). Some examples of mutualisms include the pollination of flowers by bees and the cleaning of the gills of larger fish by coral reef fish.
A mutualistic relationship has long been proposed between trees of the genus Cecropia and ants of the genus Azteca. Cecropia trees are common pioneer colonizing trees in the Neotropics (Kricher, 1997). They occur in disturbed areas such as clearings and along roads and trails. Cecropia trees are also common components of second growth forests (Logino, 1989). Descriptively, Cecropia trees have thin and hollow trunks with bamboo-like rings, wide palmate leaves, and a low degree of branching (Kricher, 1997). Azteca spp. ants are a particularly aggressive genus of ants that are often found living within Cecropia trees.
The prevailing theory is that the Azteca ants are obligately associated with the Cecropia tree in a mutualistic relationship (Agrawal and Dubin-Thaylor, 1999). The ants are believed to derive food and shelter from the trees and the trees in return receive possible protection from herbivory, and encroaching vines, in addition to recieivng nitrogen from the ant carcasses and by-products (Agrawal and Dubin-Thaler, 1999; Janszen, 1969). Previous studies have elucidated some aspects of this relationship. Azteca ants have been observed to chew vines and leaves from nearby trees that contacted the host Cercropia, effectively serving as allopathic agents for the tree (Janzen, 1969). Furthermore, it is believed that Cecropia trees receive up to 93% of their nitrogen from the ants in the form of frass and ant carcasses (Sagers et al., 2000).
The ants, in return, are known to receive shelter and between 19% and 47% of their food from glycogen rich Mullerian bodies located on the petiole of the Cercropia leaf (Sagers et al., 2000 and Folgarait, 1994).
However, the data is rather incomplete when it comes to whether the ants protect the tree from herbivory. Studies have shown that Cecropia trees with Azteca ants living within them display increases of between 50% and 100% in the number of ants on a given leaf when the leaf incurred physical damage, thus lending to the theory that the ants may protect the trees from being eaten (Agrawal and Dubin-Thaler, 1999). However, while it is known that the ants are recruited to sites of leaf damage, it is not known whether they provide an effective defense to protect the tree from further damage. It was our intention to determine whether the ants prevented the herbivory of Cercropia leaves. To this end, an herbivory field study was performed at two locations in Costa Rica, the Monteverde cloud rainforest and the Corcovado lowland rainforest. These two sites were chosen because the high elevation and climate of Monteverde were prohibitive for Azteca ants and the lowland forests of Corcovado are known to have Cecropia associated Azteca ants (Janzen 1973). We performed the herbivory study to test the null hypothesis that the herbivory in Monteverde and Corcovado would be at the same levels. The alternate hypothesis was that the trees in Corcovado would display less herbivory as a result of the protective function of the Azteca ants.
Materials and Methods
The herbivory study was performed at Monteverde and Corcovado over the period of time spanning May 17 to June 4 2001. Twenty Cecropia trees were observed along the El Camino trail in Monteverde and twenty trees were observed at a field station in Corcovado. For each tree, the canopy was divided into four quadrants and five leaf groups were observed for herbivory damage in each quadrant. The damage fell into four groups: no damage, light, medium, and heavy, based on visual observation. A representation of each group is found in figure 1.
Figure 1. Examples of Cecropia leaves that are a. Undamaged, b. Lightly Damaged, c. Moderately Damaged, and d. Heavily Damaged
The Cecropia trees were also examined visually for presence of Azteca ants. The attempt to verify resident ants consisted of knocking on and shaking of the trunk of the tree. All efforts were made to avoid permanent damage to the Cecropia trees, thus the absence of any cutting of the branches or trunks that has been used in other studies (Longino, 1989). The results were recorded and analyzed using the computer programs STATview and Minitab. Several Chi Square tests were performed on the trees to determine if there was any difference between trees at Monteverde, which were thought to have no ants, and trees at Corcovado. Specifically, the differences in total damage, light damage, moderate damage, and heavy damage were assayed between the two sites.
Initially, the degree of damage for each individual tree at each location was examined to look for any trends in the damage. This was calculated by taking the damage results for each tree are contained in Figure 1.
Figure 1. Graphical depiction of the herbivory damage for the twenty trees in a. Monteverde and b. Corcovado.
a.
b.
One trend that is somewhat apparent from Figure 1 is that, on average, the trees in Corcovado tend to show a higher percentage of light damage and a lower percentage of heavy damage than the trees in Monteverde. Light damage makes up a majority of the total damage for most trees in Corcovado. There are also few trees that show a high percentage of heavy damage in Corcovado. The opposite is true in Monteverde, as seven trees have heavy damage as the largest percentage of the total damage to the leaves of the tree.
Next, the amount of damaged leaves versus undamaged leaves for all the trees in the two locations was compared. This was done by taking an average of the damaged leaves and undamaged leaves of the twenty trees in each location. The results are displayed in Figure 2.
Figure 2. Graphical depiction of the average total amount of damaged and undamaged leaves for an average Cecropia tree in a. Monteverde and b. Corcovado.
a. b.
Figure 2 represents the percentage of leaves that would likely be damaged or undamaged for a tree in Monteverde and Corcovado, respectively. As can be seen, an average tree in Monteverde would have 68% of its leaves damaged and 32% undamaged where as a tree in Corcovado would have 85% of its leaves damaged and 15% undamaged. Thus, a tree in Corcovado generally has more herbivory damage to it, on average, than a tree in Monteverde.
The type of damage for an average tree at Monteverde and Corcovado was then determined and compared in the same manner as the total damage was. The results are depicted in Figure 3.
Figure 3. Graphical depiction of the average percentage of damage by type for a Cecropia tree in a. Monteverde and b. Corcovado.
a. b.
From Figure 3, it can be determined that, for the average Cecropia tree in Monteverde, 48% of the total leaf damage would be light, 26% would be moderate, and 26% would be heavy. For Corcovado, 60% of the total damage would be light, 29% moderate, and 11% heavy. Thus, we have confirmation of the initial finding that trees in Corcovado generally have higher levels of light damage and lower levels of heavy damage than tress in Monteverde.
To insure statistical relevance, a Chi Square test was performed on each of the above comparisons. The Chi Square on the total damage of an average tree in Monteverde versus the total damage of an average tree in Corcovado resulted in a statistical difference at the 95% confidence level (P=. 023). The Chi Square test on the light damage of an average tree in Monteverde versus an average tree in Corcovado resulted in a statistical difference at the 85% confidence level (P=. 110). The Chi Square test on heavy damage for an average tree in Monteverde versus an average tree in Corcovado showed a statistical difference at the 95% confidence level (P=. 0016). The Chi Square test conducted on the moderate damage between the two sites showed no statistical difference (P=. 546). Thus, it is obvious that, statistically, there is a difference between overall damage between the two sites and, when coupled with the previously described mean damages at each site, it becomes apparent that Cecropia trees in Corcovado suffered more herbivory damage than did those in Monteverde. However, the damage to the trees in Corcovado was far more in the light category and far less in the heavy category than those trees in Monteverde. Thus, while there was more damage seen in Corcovado, it was less severe overall than the damage seen in Monteverde.
It was the hope that when this project was begun that the Cecropia trees in Monteverde would serve as a control as the location was prohibitive to the Azteca ants. The Cecropia trees of Corcovado were to provide the Azteca ant colonies from which the herbivory study could be placed back into the context of the mutualistic relationship between Cecropia trees and the ants. However, none of the trees surveyed in Corcovado were observed to have the Azteca ants. Two trees did have ants present on the trunk but they did not appear to be Azteca ants as they were not aggressive towards the disturbance the experiment presented to the tree.
Discussion
What we are able to conclusively offer as a result of our herbivory study in Monteverde and Corcovado is that the patterns of herbivore consumption of Cercropia leaves in Monteverde and Corcovado are in fact different. Corcovado tends to show more herbivory than Monteverde but the damage is to a lesser degree. The trees in Monteverde display a significantly greater amount of heavy damage, on average, than the trees of Corcovado. This is an interesting result in itself and it is contrary to the results that we expected to find. It was thought that the trees of Corcovado would have less damage due to the protection of the Azteca ants. Obviously, we found no ants so that automatically renders our hypothesis inapplicable but the question of why the herbivory patterns are different still remains.
There are a few possible reasons for the difference that we will offer here. First of all, the fact that there is more herbivory of the Cecropia leaves in Corcovado may be the result of there being more creatures that eat Cecropia leaves in Corcovado than in Monteverde. The altitude and lower temperatures of Monteverde may be prohibitive to many other insects like they are for the Azteca ants. Some of these insects undoubtedly would feed on Cecropia trees and may well be present in the more hospitable lowlands of Corcovado. The same may be true for larger consumers of Cecropia leaves such as monkeys and sloths. Though present, they may be less numerous than in Corcovado.
Another possibility is that the Cecropia trees in Monteverde and Corcovado are different species and thus have different organisms that feed on them. Perhaps more organisms feed on the Cecropias of Corcovado than those of Monteverde. This possibility might also explain the differences in the degree of damage as well. Perhaps the species of Cecropia that predominates in Monteverde is appealing to large mammals, which cause more extensive damage. Conversely, the trees of Corcovado may be more appetizing to insects that cause lighter damage.
The lack of Azteca ants was particularly disturbing and damaging to this study. The reason for this is not known, as the Azteca ants were supposed to be abundant in Corcovado. There are several possibilities as to why the ants were not observed and they are to be discussed in the following.
The first possibility is that the trees surveyed were too young to have been colonized. Several of the trees surveyed were juveniles and no effort was made to only survey trees of a certain height or diameter at breast height (dbh). These trees were used because they were accessible from the trail. A second possibility is that the ants were present and the method used to draw them out was not effective. Previous studies have relied on cutting the leaves, branches, and sometimes the entire tree down to ascertain the presence of the Azteca ants (Agrawal and Dubin-Thaler, 1999; Longino, 1989). These methods were avoided, as direct damage to the tree was deemed inappropriate for this study. However, actual damage may be necessary to draw the ants out of the tree (Agrawal and Dubin-Thaler, 1999). Conversely, knocking on the tree, in retrospect, may have had the opposite of the desired effect because several species of woodpecker feed on Azteca ants. The knocking possibly could have simulated a woodpecker attack and resulted in the ants seeking shelter within the tree.
This study was strong in many ways. First of all, no trees were harmed as this was a very observational study and non-invasive. Second, the method of ranking the leaves based on the category of damage in addition to observing whether the leaves were damaged or not was a very strong idea as it allowed for the discovery different types of damage between the two forests. Finally, the number of readings taken from each tree (N=25) was also appropriate as it gave an accurate view to the overall state of the tree in terms of leaf damage.
However, the study was also limited in many ways, including the ability to access Cecropia trees as we were primarily restricted to trails, the small sample size, the inability to find Azteca ants, the inability to closely observe the leaves as they were often out of our reach and high in the canopy, the non-restriction of trees to an age range and the subjectiveness of the visual damage scale.
Future studies would help to elucidate whether the Azteca ants really protect the Cecropia trees from damage but they must be done on trees that contain the ants. The Cecropia trees should also be identified by species and restricted so that only mature trees are sampled. More trees should be sampled. Invasive techniques such as punching holes in leaves or removing branches may be necessary to identify the presence of Azteca ants, which should also be identified by species, if possible. Finally, a refinement of the damage ranking scale would be an improvement to make it more consistent and less subjective to the observer.
In conclusion, the end and useful result of this study is that a baseline of herbivory damage in both the high elevation and low elevation Cecropia trees was established that may be useful for further research on the Azteca ant protection theory. It was found that the Cecropia trees in Corcovado have a greater amount of herbivory damage but it is to a lesser degree than that of Monteverde. While we were unable to ascertain or get indications of whether the Azteca ants protect the Cecropia trees from herbivory, future research that build on this study may well be able to determine this relationship based on the results and study framework that we have provided here.
References
1. Agrawal, Anurag A. and Benjamin J. Dubin-Thaler. "Induced Responses to herbivory in the Neotropical ant-plant association between Azteca ants and
Cecropia trees: responses of ants to potential inducing clues", Journal of
Behavior, Ecologoy and Sociobiology, Vol. 45, 1999, pgs 47-54
2. Folgarait, P.J., H.L. Johnson, and D.W. Davidson. "Responses of Cecropia to Experimental removal of Mullerian Bodies", Functional Ecology, Vol. 8 Iss. 1, 1994, pgs. 22-28
3. Janzen, Daniel H. "Alleopathy by Myremecophytes: The ant Azteca as an Allelopathic Agent of Cercropia", Ecology, Vol. 50 Iss. 1, 1969, pgs. 147-
153
4. Janzen, Daniel H. "Dissolution of Mutualism between Cercropia and its Azteca ants", Biotropica, Vol. 5 Issue 1, 1973, pgs. 15-28
5. Kricher, John. A Neotropical Companion, Princeton University Press,
Princeton, New Jersey, 1997, pgs.130-132
6. Longino, John T. "Geographic Variation and Community Structure in Ant-Plant Mutualism: Azteca and Cecropia", Biotropica, Vol. 21 Iss. 2, 1989,
pgs. 126-132
7. Purves, William K., Gordan H. Orians, and H. Craig Heller. Life: The
Science of Biology. Fourth Edition, Sinauer Associates, Sunderland, Massachusetts, 1995, pgs. 1013-1014
8. Sagers, C.L. , S.M. Ginger, R.D. Evans. "Carbon and nitrogen isotopes trace nutrient exchange in an ant-plant mutualism", Oecologia, Vol. 123, 2000, pgs. 582-586
Next Article
Previous Article
Return to Topic Menu
It is 6:28:19 PM on Saturday, July 4, 2009. Last Update: Sunday, September 8, 2002