Boomeranging in structural defense

Abstract

Plant defensive behaviors that resist arthropod herbivory include trichome-mediated defenses, and variation in plant trichome morphology and abundance provides examples of the mechanistic complexities of insect-plant interactions. Trichomes were removed from Cycas revoluta cataphylls on the island of Guam to reveal Aulacaspis yasumatsui scale infestation, and predation of the newly exposed insects by pre-existing Rhyzobius lophanthae beetles commenced within one day. The quotient of predated/total scale insects was 0.5 by day 4 and stabilized at that found on adjacent glabrous leaves in about one week. The trichome phenotype covering the C. revoluta cataphyll complex offers the invasive A. yasumatsui armored scale effectual enemy-free space in this system. This pest and predator share no known evolutionary history with C. revoluta, therefore, the adaptive significance of this plant behavior in natural habitat is not yet known.

Keywords: :

• Aulacaspis yasumatsui
• Cycas revoluta
• enemy-free space
• invasive insects
• Rhyzobius lophanthae
• trichome

Introduction

The invasion of Guam by Aulacaspis yasumatsui was predicted in 2000 and verified in Oct. 2003 as a limited infestation in the landscape of the main hotel district, then dispersal of the pest from urban areas to native forest habitat was documented in Jan. 2005.¹ The scale predator Rhyzobius lophanthae was purposely introduced in Nov. 2004, and released in native forest habitats beginning Feb. 2005.² This predator rapidly established and signs of scale predation have been persistent. Despite this history, plant mortality has been epidemic and continues to date. Only a few of the Cycas revoluta trees within urban forests are still alive, and more than 90% of the native Cycas micronesica trees have been killed by the chronic scale infestations.³

After full establishment of the predator, the scale individuals that lacked predation were many times located in small crevices or other locations that appeared to be too diminutive for the predator to maneuver. In attempts to validate this observation, Cycas revoluta cataphyll surfaces were exposed by removing the very dense trichomes (Fig. 1) that cover this taxon’s cataphylls.⁴ As expected, the newly exposed plant surfaces revealed heavy infestations of cryptic scale showing no signs of predation despite the presence of Rhyzobius lophanthae beetles on the same plants.

Figure 1. The Cycas revoluta cataphyll complex is adorned with elaborate tufts of trichomes (top). Quotient of Aulacaspis yasumatsui individuals on Cycas revoluta cataphylls that were predated by Rhyzobius lophanthae beetles as a function of days after trichome removal. Mean ± SE (bottom).

Trichome morphology per se was clearly unable to deter scale infestations on the cataphyll surfaces, and we hypothesized that trichome exclusion of R. lophanthae was responsible for the complete lack of biological control of the scale insects infesting these trichome-endowed plant surfaces.⁴ Here I tested this hypothesis by exposing A. yasumatsui on cataphyll surfaces via removal of trichomes, then monitoring predation of the scales that were exposed until predation percentage reached that of the scale infestations on leaf surfaces from the same plants.

Results

The interactive effect of date X time of day was not significant (p = 0.15), and the main factor time of day was not significant (p = 0.25) for the number of beetles observed on the exposed half of the cataphyll complex. The number of beetles observed differed among the dates (f = 9.20, df = 12, p < 0.0001). No beetles were searching for or feeding on scales on the newly exposed cataphyll surfaces on the initial date, but beetles were observed beginning 0800 h on the next day. Therefore, less than 24 h was needed for this predator to move into the space exhibiting newly exposed prey. The number of beetles increased daily until peaking on day 4 at 2.44 per plant, then gradually declined until day 9 when the counts stabilized at ca 0.4 per plant.

The interactive effect of date X sex was not significant (p = 0.98), and the main factor time of sex was not significant (p = 0.70) for the quotient of predated/total scale insects. However, the quotient of predated/total scale insects differed among the dates (f = 27.46, df = 12, p < 0.0001). Predation quotient was almost 0.1 by day 1, and rapidly increased to 0.5 by day 4 (Fig. 1). The scale predation reached a stable ca. 0.65 by day 8 or 9.

Trichomes were removed on day 12 from the second half of each cataphyll complex. There were ca. 4 scales per cm², and number of R. lophanthae and predation were nil (Fig. 1) as with day 0 for the initially exposed half.

Discussion

The results of this study verify that the ability of A. yasumatsui to infest trichome-protected cataphylls on C. revoluta plants on Guam is a direct result of preferential exclusion of the R. lophanthae predators by the trichomes. Predators were observed feeding on A. yasumatsui within one day of cessation of the prophylactic behavior provided by trichomes. Moreover, the quotient of predated/total A. yasumatsui individuals reached that of adjacent glabrous leaflet surfaces, which are not protected by heavy trichome density, in about one week following trichome removal.

Structural defenses employed by plants to resist arthropod herbivory include trichome-mediated defenses, and numerous studies validate that trichomes reduce damage caused by herbivorous insects on some plant species.^5-11 However, host plant phenotype plays a crucial role in how herbivorous arthropods interact with entomophagous arthropods.¹² Some studies reveal no effect of trichome phenotype or density on performance of natural enemies of herbivorous insects.¹³ Alternatively, other studies reveal that trichomes reduce the extent of insect herbivory by aiding predators or parasitoids.¹⁴ Similar to my results, other studies reveal that trichomes increase extent of insect herbivory by decreasing performance of predators or parasitoids^15-20 and some of these included Coccinelidae predators.^21,22

In this case study from the Mariana Islands, the Cycas tree population is only recently interacting in an anthropogenic-initiated tri-trophic quagmire. The alien herbivore pressure was initiated in 2003 by an unintentional introduction,^1,23 and the alien predator protection was initiated in 2005 by a purposeful introduction.² Within the context of this artificial tri-trophic system, the ability of C. revoluta to protect plant surfaces with trichomes is disadvantageous, offering the lethal herbivore physical protection from predation.

General observations indicate trichome morphotype is a stable trait among cycad taxa and is therefore useful for taxonomy.²⁴ However, the adaptive significance of trichome production within the Cycas genus is not known. Trichomes serve purposes other than direct influence on insect herbivores and their natural enemies. For example, trichomes may increase tolerance of abiotic factors such as drought stress and UV exposure, and help regulate organ temperatures.^25,26 Furthermore, foliar trichomes have been shown to express highly efficient uptake of nitrogen,²⁷ phosphorus,²⁸ and potassium.²⁹ Production of trichomes by Cycas plants may fulfill one of these alternative roles with my observed effects on excluding a specialist pest’s predator emerging as an incidental behavior.

This paper points out the uncertainties of using generalist alien predators for control of specialist alien arthropod pests. Perhaps biological control programs that rely on generalist predators would best be served with an expectation that initial control may not be entirely effective. This expectation may enable a more aggressive initial search for the factors that reduce effectiveness, followed by a vigorous search for alternative biological control agents that do not share the defined limitations of the initial biological control organism. In this case a parasitoid that has a body size similar to scale crawlers may not be hindered by the same prophylactic exclusion imposed by the dense trichome display of C. revoluta. Following the A. yasumatsui invasion of Guam we introduced R. lophanthae because it is known as one of the most economically important natural enemies of armored scale insects³⁰ and pre-existing populations established as a predator on A. yasumatsui after the scale invaded Florida³¹ and Texas.³² We could have been more aggressive in continuing the search for a second biological control organism at the onset if we had been more astute in predicting the Cycas plant behaviors that have denied effective R. lophanthae predation over time. The results of my study indicate the introduction and release of a second Coccinelidae predator would likely be constrained by the same size differential that disallows effective R. lophanthae control on the surfaces of many Cycas organs. Biological control goals may be reached more effectively if efforts are directed toward continued search for a parasitoid with a body size that would be able to maneuver into the cryptic locations that are inaccessible to the predator.

Materials and methods

Cycas revoluta cataphyll trichomes were removed by water spray as previously described.⁴ I removed the trichome cover on one half of the cataphyll complex for each of six trees, leaving the opposite half with intact trichomes. Following this procedure I counted the number of R. lophanthae visible on the surface of the exposed cataphylls then counted the number of male scale and female scales within ten randomly placed 5x5 mm squares as previously described.⁴ I also recorded if each scale had been predated by R. lophanthae. The beetle counts were repeated at 0800 h, 1200 h, and 1800 h beginning the following day and continuing until day 12. I repeated the scale counts and determination of predation at 0800 h every day.

The beetle count data were square-root transformed (count + 0.0001) and the P values for main factors date, time of day, and the interaction between date and time of day were obtained by fitting a Generalized Linear Mixed Model (SAS Procedure GLIMMIX) with a Poisson distribution. The response variable used to quantify predation was the quotient predated scales/total scales. This variable was square-root transformed and the data were fitted with a Generalized Linear Mixed Model with Poisson distribution. The main factors were date, sex, and the interaction of date and sex.

On the final day of measurements, I removed the trichomes from the second half of each cataphyll complex to validate that increased predation on the initially exposed half was indeed due to the experimental treatment and not coincidental timing of increased predation. Following exposure of new cataphyll surfaces, I conducted the beetle counts, scale counts, and determination of predation.

Conclusions

Studying variation in plant trichome morphology and abundance in plant surface morphology provides examples of the mechanistic complexities of insect-plant interactions. The C. revoluta trichomes clearly defended the cataphyll surfaces against access by an arthropod. Unfortunately, that arthropod was an entomophagous predator. Therefore, the trichome phenotype covering the C. revoluta cataphyll complex exhibited a boomerang effect, where the invasive A. yasumatsui armored scale used the plant’s structural defense traits to find enemy-free space. To my knowledge, this is the first report on the influence of trichomes on plant-herbivore-predator interactions for any cycad species. Trichome phenotype and density may be involved in mediating invasive insect damage to some of the other 331 described cycad species,³³ which collectively comprise the most threatened group of plants worldwide.³⁴ A. yasumatsui and R. lophanthae share no known evolutionary history with C. revoluta, so the adaptive significance of this plant behavior in natural habitat is not yet known.

Disclosure of Potential Conflicts of Interest

No potential conflicts of interest were disclosed.

Acknowledgments

I thank George Fernandez for statistical analysis. Support provided by US Forest Service Project No. 10-DG-11059702-095. This material was made possible, in part, by a Cooperative Agreement from the United States Department of Agriculture's Animal and Plant Health Inspection Service (APHIS). It may not necessarily express APHIS' views