Effects of host plant architecture on colonization by galling ... - debio

Austral Ecology (2006) 31, 343–348

doi:10.1111/j.1442-9993.2006.01563.x

Effects of host plant architecture on colonization by galling insects ANA PAULA ALBANO ARAÚJO,1* JOANA D’ARC DE PAULA,2 MARCO ANTONIO ALVES CARNEIRO2 AND JOSÉ HENRIQUE SCHOEREDER3 1 Termite Laboratory, Department of Animal Biology, Federal University of Viçosa, Av. Prof. H. Rolfs, s/n. 36.570-000,Viçosa-MG, Brazil (Email: [email protected]), 2Laboratory of Animal Diversity, Department of Biological Sciences, Federal University of Ouro Preto, Campus Morro do Cruzeiro, Ouro Preto-MG, and 3Community Ecology Laboratory, Department of General Biology, Federal University of Viçosa, Viçosa-MG, Brazil

Abstract: To study the abundance and occurrence of herbivore insects on plants it is important to consider plant characteristics that may control the insects. In this study the following hypotheses were tested: (i) an increase of plant architecture increases species richness and abundance of gall-inducing insects and (ii) plant architecture increases gall survival and decreases parasitism. Two hundred and forty plants of Baccharis pseudomyriocephala Teodoro (Asteraceae) were sampled, estimating the number of shoots, branches and their biomass. Species richness and abundance of galling insects were estimated per module, and mortality of the galls was assessed. Plant architecture influenced positively species richness, abundance and survival of galls. However, mortality of galling insects by parasitoids was low (13.26%) and was not affected by plant architecture, thus suggesting that other plant characteristics (a bottom-up pressure) might influence gall-inducing insect communities more than parasitism (a top-down pressure). The opposite effect of herbivore insects on plant characteristics must also be considered, and such effects may only be assessed through manipulative experiments. Key words: abundance, Baccharis pseudomyriocephala, galling insect, richness, structural complexity.

INTRODUCTION The abundance and species richness of herbivore insects living on a plant may be determined by several factors (Lawton 1978; Strong et al. 1984). Many studies have correlated the number of insects to structural and physiological characteristics of the plants. Herbivore insects are more abundant, for instance, on plants with higher structural complexity, complexity meaning a combination of size, growth form and variety of above-ground structures of the plant (Lawton 1983; Strong et al. 1984). Plant structural complexity might influence herbivore species richness in three ways: (i) allowing the creation of microhabitats in a given plant, with different temperature and moisture (Gonçalves-Alvim & Fernandes 2001); (ii) supporting larger herbivore populations, and therefore reducing extinction probability; and (iii) providing more refuge sites and allowing the persistence of species more susceptible to predation. A well-documented pattern is the relationship of herbivore species richness on plant forms, which

*Corresponding author. Accepted for publication August 2005.

© 2006 Ecological Society of Australia

follows the sequence herbs < shrubs < trees (Strong et al. 1984). Lawton (1983) suggested that the observed pattern might be linked to the size of plants because taller plants are more apparent in space and time and are therefore more easily found by herbivores. Furthermore, they are bigger, and may have more oviposition and feeding sites for herbivores. Therefore, two factors might influence herbivory on plants with complex architecture: the architectural complexity itself and the increase of available plant area. To disentangle the relative effects of each of these factors is a first step in understanding the factors determining herbivory. Gall-inducing insects are sessile and stay on the same plant during their larval development. Therefore, to hide from predators might be more important to galls than to mobile insects. Galls induced in plants with a higher architectural complexity could be more difficult to find by parasitoids, and consequently would present higher chances of survival. This study tested the hypotheses (i) that plants with higher architectural complexity harbour more gallinducing species and more individuals per species; and (ii) that gall-forming insects on plants with higher complexity have higher survival probabilities and lower mortality due to parasitism.

344

A . P. A . A R A Ú J O E T A L .

METHODS Study site

The study was carried out in the state park of Itacolomi, Ouro Preto, MG, Brazil (20°22′30″S, 43°32′30″W). The highest altitude of the park is 1772 m above sea level and its vegetation types include semi-deciduous seasonal forests, gallery forests, Araucaria remnants, and quartzite and ferruginous fields. All plants sampled were above 1100 m.

Host plant and gall-inducing insects

The host plant, Baccharis pseudomyriocephala Teodoro (Asteraceae), is a dioecious weedy shrub, which forms dense and dominant patches in degraded and eroded areas of the park. According to Araújo et al. (2003), B. pseudomyriocephala can have 10 gall morph types, induced by Cecidomyiidae (Diptera), Psyllidae (Heteroptera) and Lepidoptera.

Sampling

Two hundred and forty plants were sampled from July to September 2000 (120 plants) and from February to April 2001 (120 plants). The plants were sampled along a transect, disposed in a single patch of B. pseudomyriocephala. The sampling was designed to choose systematically one plant from one side of the transect, and the next from the other side, leaving two unsampled plants between each pair of sampled plants. The above-ground parts of the sampled plants were entirely removed and stored in a freezer until further analysis. All apical branches of the plants, called modules, were cut. Modules are morphological subunits from meristems (Price 1991), and we considered them the newest plant tissues. These modules are believed to represent potential resources to gallinducing insects, because they constitute active meristems. The numbers of primary and secondary ramifications of each branch were also counted. The numbers of modules, primary and secondary ramifications were considered as a measure of structural complexity of the host plant. After the removal of galls, all modules were dried in an oven at 60°C for three days, until they reached constant weight. Biomass was used as a surrogate of the size of modules, assuming a positive relationship between plant growth and biomass (see Araújo et al. 2003). Gall species richness and abundance was determined in each module. The external gall morphology was used for insect species identification (Weis et al. 1988; Dreger-Jauffret & Shorthouse 1992; Floate et al.

1996). Subsequently each gall was dissected to verify the presence of alive or dead galling insects and to determine the cause of mortality. Gall mortality factors were separated into parasitism, fungal attack and unknown factors, based on the criteria adopted by Fernandes and Price (1992), considering also that galls were frozen before observations. Mortality was attributed to parasitism when parasitoid exuviae remained inside the gall and/or when the presence of parasitoid exit holes was observed. Mortality was attributed to fungal attack when the gall chamber was filled by fungi. When gall tissue was necrotic and/or when the gall chamber was obliterated by the growth of gall tissue, death was attributed to an unknown factor. The galls were considered alive when their exoskeleton was intact or, if immature, when the internal integrity of larvae tissue was observed by transparency. Empty galls, with larger exit holes caused by the exit of the gall-inducing insects, were also considered alive galls. The above methodology could not detect whether sampled insects would have completed their life cycles, and an alive insect may have been parasitized if still in the plant. Therefore, this method overestimated survival, and consequently underestimated mortality. Nevertheless, such an error would occur in all plants, and there is no reason to suppose that subsequent mortality would differ in plants with higher or lower complexity.

Statistical analyses

All analyses were carried out using generalized linear models in R (R Development Core Team 2005). Complete models were fitted, and then simplified by removing non-significant variables in turn, and verifying the effect of removal on the deviance (Crawley 2002). To test the relationship of gall species richness as well as gall abundance with architectural complexity, models that included module biomass, number of modules and the number of primary and secondary branches were fitted. Poisson distributions, corrected for over-dispersion, were used. The analyses were followed by residual analyses to test for the suitability of the models and error distribution. The relationship between gall-inducing insect survival and plant architecture were tested through regression analysis, in which the response variable was the proportion of surviving insects and total number of galls, and explanatory variables were the same as in above models. After checking the residuals, the model was simplified and a non-linear model was fitted: proportion of surviving insects = a − b × exp (–c ×number of modules). The same model was used to test for the response of the proportion of galls that died due to parasitism and the total number of dead galls per plant. Binomial distribution corrected for over© 2006 Ecological Society of Australia

P L A N T A R C H I T E C T U R E A N D G A L L C O L O N I Z AT I O N

result regarding the effect of architecture (see Schoereder et al. 2004).

dispersion was used. The analyses were followed by residual analyses to test for the suitability of the models and error distribution. Module biomass was used as one of the explanatory variables in all models because host plant size might affect the colonization by gall-inducing insects. As plant architecture and biomass are possibly correlated, the model is non-orthogonal. To use a more conservative model, most deviance was attributed to module biomass, entering this variable first in the models. Therefore, we are confident of any possible significant

RESULTS All collected plants sampled contained galls. Species richness (F1,238 = 32.5645; P < 0.001) and abundance (F1,238 = 165.464; P < 0.001) of gall-inducing insects were positively correlated with the biomass of modules (Figs 1a,2a). After the positive effect of biomass has (b)

8

Gall-inducing species richness

Gall-inducing species richness

(a)

345

7 6 5 4 3 2

8 7 6 5 4 3 2 1

1 0

20

40

60

80

100

0

Weight of modules (g)

20 40 60 80 Number of branches

100

(a)

(b)

100

100 Abundance of galls

Abundance of galls

Fig. 1. Relationship between gall-inducing species richness and (a) the weight of plant modules and (b) the number of branches per plant in Baccharis pseudomyriocephala.

80 60 40 20 0

80 60 40 20 0

0

20

40 60 80 Weight of modules (g)

100

0

20

40 60 80 Number of modules

(c)

Abundance of galls

100 80 60 40 20 0 0

20 40 60 80 Number of branches

100

Fig. 2. (a) Relationship between gall-inducing abundance and the weight of modules; (b) relationship between abundance of galls and the number of modules and (c) relationship between gall abundance and the number of branches of plants.


346


been accounted for, species richness (F1,235 = 11.429; P < 0.001) and abundance (F1,235 = 50.147; P < 0.001) of gall-inducing insects were significantly larger in plants with higher structural complexity. Species richness was positively correlated with the number of branches (F1,237 = 5.7174; P = 0.01758) (Fig. 1b). Gall abundance was significantly related both to the number of modules (F1,237 = 16.165; P < 0.001) and to the number of branches (F1,236 = 19.792; P < 0.001) of the host plant (Fig. 2b,c). From a total of 3925 galls analysed, 71.6% contained galling insects that were alive when sampled. Attacks by parasitoids accounted for 13.3% of all dead galling insects, fungi accounted for 7.2%, whereas the cause of death of most dead galls (79.5%) was unknown. The number of modules positively influenced the survival of gall-inducing insects (F1,239 = 5844.7; P < 0.001) (Fig. 3). Plant architectural complexity was not significantly related to parasitism (F4,214 = 1.96; P = 0.10), neither was module biomass significantly related to survival (F1,238 = 0.74; P = 0.39) or parasitism (F1,217 = 0.11; P = 0.74).

DISCUSSION In our study gall-inducing insects occurred on all plants that were sampled and their survival was around 70%. The structural complexity of the host plant was one factor that was related to both species richness and abundance of galls. A well-documented pattern in other studies is that species richness is higher in sites with higher heterogeneity (Rosenzweig 1995). Plants with higher archi-

Survival of gall-inducing insects

1.0

0.8

0.6

0.4

0.2

0.0 0

20

40

60

80

Number of modules

Fig. 3. Relationship between the proportion of surviving galls and the number of modules of the host plant in Baccharis pseudomyriocephala.

tectural complexity may represent a heterogeneous habitat to insects (Strong et al. 1984) because they increase (i) resource spatial distribution and (ii) the abundance of apical meristems. These two factors may explain the increase of species richness and abundance of gall-inducing insects in plants with higher architectural complexity. Spatial resource partition increases diversity of microhabitats and allows the coexistence of different species (Ettema & Wardle 2002). According to Marquis et al. (2002), an increase of architectural complexity of plants has a positive effect on the number of insects, because proximity among leaves eases the locomotion and resource finding by herbivores. Furthermore, plants with higher structural complexity possess more apical meristems, which are considered as high-quality sites for oviposition because they are active tissues. Therefore, it is expected that more complex plants maximize the reproductive success of gall-inducing insects, facilitating the selection of oviposition sites. In our study, gall species richness and abundance were also influenced by plant biomass. An increase of gall abundance with increasing resource have also been observed in the congeneric plant species Baccharis dracunculifolia (Gonçalves-Alvim et al. 1999), as well as in the Leguminosae Bauhinia brevipes (Cornelissen & Fernandes 2001). In our data, the increase of biomass may have resulted in a higher amount of resources, supporting an increase of populations (higher gall abundance) and allowing the coexistence of different species. On the other hand, plant biomass increase may also reflect more area (larger plants). Colevatti and Sperber (1997) have shown that B. dracunculifolia plant size correlated with higher gall abundance, and attributed such result to a sampling artefact. The difficulty of separating effects of size and structural complexity has limited the study of species–area relationships (Ricklefs & Lovette 1999). However, our results showed that gall-inducing insect species richness correlated positively with the increase of plant structural complexity, independent of plant size. Survival of gall-inducing insects also showed a positive relationship with increasing plant structural complexity. Higher complexity may have facilitated the selection of adequate oviposition sites, increasing survival probabilities. Mortality from parasitism was not related to plant architecture, and parasitism was not the main cause of gall mortality. Low gall mortality from parasitism has been observed before (Fernandes & Price 1992; Fernandes & Negreiros 2001; EspíritoSanto & Fernandes 2002), and has been attributed to hypersensitivity reactions. Plants that grow larger usually show lower investment in defence, and therefore plants with higher biomass would support larger densities of gall-inducing insects. Our results, as well as other, suggest that bottom-up control (the pressure © 2006 Ecological Society of Australia

P L A N T A R C H I T E C T U R E A N D G A L L C O L O N I Z AT I O N

from the primary trophic level) might be more important to the regulation of the community of gall-inducing insects than the top-down control (the pressure from parasitoids and predators) (Fernandes & Negreiros 2001). On the other hand, our results may suggest a response of plants to herbivory instead of a response of herbivores to plant structure and biomass. Some species of gall-inducing insects prefer meristems as sites to oviposit their eggs (Weis et al. 1988; Shorthouse & Rohfritsch 1992). Destruction of these meristems may reduce the dominance of such meristems over secondary meristems, which would in turn increase the number of branches and the structural complexity of the attacked plant. Therefore, herbivory would promote a compensatory growth of the plant resulting in increased complexity (Callaway et al. 2003). The effect of herbivores on plant growth and productivity has been previously reported (Owen 1980; Abrahamson & Weis 1987). Some studies showed that previously galled branches were longer and bore more leaves and shoots than non-galled branches (Fernandes & Ribeiro 1990; Prado & Vieira 1999). The effect of gall-inducing insects on plant structural complexity and biomass, or vice versa, can only be tested by manipulative experiments. Other host plant characteristics, such as the nutritional quality, may also influence the settling and colonization success of gall-inducing insects. Gallinducing insect communities might also respond to host plant phenology (Yukawa 2000), which may induce seasonal changes in nutritional quality (EspíritoSanto et al. 1999). Our results corroborate the hypothesis that the architectural complexity of host plants may increase species richness and abundance of gall-inducing insects. It is important to notice that plant architecture determined species richness independently of plant biomass. In addition, gall survival was correlated with the number of branches of the plants. However, the increase of structural complexity seems to be related more to plant resistance or nutritional quality than to reduction of parasitism. Such results must be viewed with the necessary caution of every correlative study, because an effect of gall-inducing insects on plant architecture cannot be entirely dismissed.

ACKNOWLEDGEMENTS The authors are indebted to Arne Janssen (Visting Teacher, scholarship Professor Visitante Estrangeiro from CAPES (Coordenação de Aperfeiçoamento de Pessoal de Nível Superior)/Brazil) for comments and English revisions, and to two anonymous reviewers. This study and Ana Paula A. Araújo were supported by FAPEMIG (Fundação de Amparo à Pesquisa do © 2006 Ecological Society of Australia

347

Estado de Minas Gerais – CRA 2893/98). José H. Schoereder is supported by a CNPq (Conselho Nacional de Desenvolvimento Científico e Tecnológico) grant. Our thanks to IEF (Instituto Estadual de Florestas) and especially to the state park of Itacolomi for their support.

REFERENCES Abrahamson W. G. & Weis A. E. (1987) Nutritional ecology of arthropod gall makers. In: Nutritional Ecology of Insects, Mites, Spiders, and Related Invertebrates (eds F. J. Slansky & J. G. Rodriguez) pp. 235–58. Wiley-Interscience, New York. Araújo A. P. A., Carneiro M. A. A. & Fernandes G. W. (2003) Efeitos do sexo, do vigor e do tamanho da planta hospedeira sobre a distribuição de insetos indutores de galhas em Baccharis pesudomyriocephala Tedodoro (Asteraceae). Revta Bras. Entomol. 47, 483–90. Callaway R. M., Pennings S. C. & Richards C. L. (2003) Phenotypic plasticity and interactions among plants. Ecology 84, 1115–28. Colevatti R. G. & Sperber C. F. (1997) The gall maker Neopelma baccharidis Burck. (Homoptera:Psyllidae) on Baccharis dracunculifolia DC. (Asteraceae): individual, local and regional patterns. Ann. Soc. Entomol. Brasil 26, 45–53. Cornelissen T. G. & Fernandes G. W. (2001) Patterns of attack by herbivores on the tropical shrub Bauhinia brevipes (Leguminosae): Vigour or chance? Eur. J. Entomol. 98, 37–40. Crawley M. (2002) Statistical Computing: An Introduction to Data Analysis Using S-plus. John Wiley & Sons, Baffins Lane. Dreger-Jauffret F. & Shorthouse J. D. (1992) Diversity of gallinducing insects and their galls. In: Biology of Insect-induced Galls (eds J. D. Shorthouse & O. Rohfritsch) pp. 8–33. Oxford University Press, London. Espírito-Santo M. M. & Fernandes G. W. (2002) Host plant effects on the development and survivorship of the galling insect Neopelma baccharidis (Homoptera: Psyllidae). Austral Ecol. 27, 249–57. Espírito-Santo M. M., Fernandes G. W., Allain L. R. & Reis T. R. F. (1999) Tannins in Baccharis dracunculifolia (Asteraceae): effects of seasonality, water availability and plant sex. Acta Bot. Bras. 13, 167–74. Ettema C. H. & Wardle D. A. (2002) Spatial soil ecology. Trends Ecol. Evol. 17, 177–83. Fernandes G. W. & Negreiros D. (2001) The occurrence and effectiveness of hypersensitive reaction against galling herbivores across host taxa. Ecol. Entomol. 26, 46–55. Fernandes G. W. & Price P. W. (1992) The adaptative significance of insect gall distribution: survivorship of species in xeric and mesic habitats. Oecologia 90, 14–20. Fernandes G. W. & Ribeiro S. P. (1990) Plant response to herbivory: two examples from neotropics. Ecotropicos 3, 77– 86. Floate K. D., Fernandes G. W. & Nilsson J. A. (1996) Distinguishing intrapopulacional categories of plants by their insect faunas: galls and rabbitbrush. Oecologia 105, 221–9. Gonçalves-Alvim S. J., Faria M. L. & Fernandes G. W. (1999) Relationships between four neotropical species of galling insects and shoot vigor. Ann. Soc. Entomol. Brasil 28, 147– 55. Gonçalves-Alvim S. J. & Fernandes G. W. (2001) Biodiversity of galling insects: historical, community and habitat effects in four neotropical savannas. Biodivers. Conserv. 10, 70–98.

348


Lawton J. H. (1978) Host plant influences on insect diversity: the effects of space and time. In: Diversity of Insect Faunas (eds L. A. Mound & N. Waloff) pp. 105–25. Backwell, Oxford. Lawton J. H. (1983) Plant architecture and the diversity of phytophagous insects. Annu. Rev. Entomol. 28, 23–9. Marquis R. J., Lill J. T. & Piccinni A. (2002) Effect of plant architecture on colonization and damage by leaftying caterpillars of Quercus alba. Oikos 99, 531–7. Owen D. F. (1980) How plants may benefit from the animals that eat them. Oikos 35, 230–5. Prado P. I. K. L. & Vieira E. M. (1999) The interplay between plant traits and herbivore attack: a study of a stem galling midge in the tropics. Ecol. Entomol. 24, 80–8. Price P. W. (1991) Plant vigor hypothesis and herbivore attack. Oikos 62, 244–51. R Development Core Team (2005) R: A Language and Environment for Statistical Computing. R Foundation for Stastical Computing, Vienna.

Ricklefs R. E. & Lovette I. J. (1999) The roles of island area per se and habitat diversity in the species–area relationships of four Lesser Antillean faunal groups. J. Anim. Ecol. 68, 1142–60. Rosenzweig M. (1995) Species Diversity in Space and Time. Cambridge University Press, Cambridge. Schoereder J. H., Galbiati C., Ribas C. R. et al. (2004) Should we use proportional sampling for species–area studies? J. Biogeogr. 31, 1219–26. Shorthouse J. D. & Rohfritsch O., eds (1992) Biology of Insectinduced Galls. Oxford University Press, Oxford. Strong D. R., Lawton J. H. & Southwood T. R. E. (1984) Insect on Plants: Community Patterns and Mechanisms. Backwell Science, London. Weis A. E., Walton R. & Creco C. L. (1988) Reactive plant tissue sites and the population biology of gall makers. Annu. Rev. Entomol. 33, 467–86. Yukawa J. (2000) Synchronization of gallers with host plant phenology. Popul. Ecol. 42, 105–13.
