"r"'

."

An. Soc. Entomol. Brasi128(2)Junho. 1999 21

ECOLOGY. BEHAVIOR AND BIONOMICS

Cynipid Gall Growth Dynamics and Enemy Attack: Effects of GallSize, Toughness and Thickness

GERALDO W. FERNANDES. MARIO M. EspIRITO-SANTO AND MAURicIo L. FARIA

Ecologia Evolutiva de Herbívoros Tropicais/DBG, ICB/Universidade Federal de MinasGerais, Caixa postal 486, 30161-970, Belo Horizonte, MO, Brasil.

An. Soc. Entomo1. Brasil 28(2): 211-218 (1999)

Dinâmica de Crescimento de um Cinipídio e Ataque por Inimigos Naturais:Efeitos do Tamanho, Dureza e Espessura da Galha

RESUMO- Insetos g~lhadores desenvolveram várias estratégias para evitar oureduzir o ataque por inimigos naturais, tais como maior espessura da parede noinício de formação, e maior diâmetro e dureza da parede em estágios posterioresdo desenvolvimento. Neste estudo. analisamos a eficiência de cada um dessesparâmetros na redução da mortalidade das galhas de Atrusca caprone Weld(Hymenoptera: Cynipidae) em Quercus turbinella Greene (Fagaceae), duranteos três meses de seu desenvolvimento (junho-agosto). O diâmetro e a durezadas galhas aumentaram do primeiro para o terceiro mês de estudo (ANOVA,diâmetro: F = 88.73, p < 0.0001; ANOVA, dureza: F = 26.13, p < 0.0001).

Entretanto, a espessura das galhas apresentou um aumento de junho parajulho,diminuindo de julho para agosto (ANOVA, F = 35.84, p < 0.0001). A análise de

regressão linear múltipla mostrou que apenas a dureza das galhas influenciou asusceptibilidade das larvas ao ataque por parasit6ides (r = 0.52, F = 13.84, p <

0.01, n = 15). A sobrevivência das galhas foi elevada no primeiro mês,diminuindo com o tempo, provavelmente devido à menor dureza da parede.Estes resultados sugerem que o primeiro mês seja crítico para o estabelecimentodas galhas, sendo que galhas não parasitadas neste estágio são menos suscetíveis

a um ataque posterior.,

Insecta, galhadj!F, Atrusca caprone, parasitóides,Quercus turbinella, janela de vulnerabilidade.

PAT_AVRAS CHAVE:

ABSTRACT. Galling insects have developed many strategies to preclude orreduce the attack by natural enemies, such as an increased wall thickness earlyin the season, and switching to larger and tougher walls later in the maturationstage. In this study, we observed the efficiency of each one of these parametersin reducing mortality of the leaf galling wasp Atruscacaprone Weld (Hymenop-tera: Cynipidae), on Quercus turbinella Greene (Fagaceae), during the threemonths of gall development (June-August). Gall diameter and wall toughnessincreased from the first to the third month of the study (ANOV A, Diameter: F =

88.73, p < 0.0001; ANOVA, Toughness: F= 26.13, p < 0.0001). However, gallwall thickness increased from June to July, decreasing in August (ANOVA, F =

212 Fernandes etal.

35.84, p < 0.0001). Gall survivorship was only 2% in June, increasing to 10 %in July and to 29.3% in August. Multiple regression analyses showed that onlygall wall toughness influenced gall susceptibility to parasitoid attack (r = 0.52,F = 13.84, p < 0.01). Gall survivorship was very lowin the first month due

probably to low wall toughness, which led to a higher success of oviposition byparasitoids during this phase. These results suggest that the first month is criti-cal to gall establishment, and the galls which are not parasitized at this stage areless likely to be attacked later.

KEY WORDS: Insecta, gall development, parasitoids, Atrusca caprone, win.dow of vulnerability, Quercus turbinella.

toughness in time. With increasing age the galltissue becomes harder and the gall larger,therefore precluding parasitism (Weis et al.1985). In this way, a "window ofvulnerabil-ity': (CornellI983) appears to exist early inthe season, whenthe developinggall is smallerand softer. However, these early galls maypreclude or decrease mortality by natural en-emies if they have thicker walls which act asa barrier to the parasitoids' ovipositors. Wepostulate that galls may have different 'on-togenetic' barriers to avoid parasitoid attack,if their walls are thicker in the early seasonand harder and larger in the late season.

We tested this hypothesis observing thephenological development and thesurvivorship rates of a leaf galler, Atruscacaprone Weld (Hymen'optera: Cynipidae ), onQuercus turbinella Greene (Fagaceae). Thisstudy aimed to answer the following ques-tions: a) How does gall development affectsgall toughness, gall size, and gall thickness of

~.. caprone ? b) Which ofthese variables aremore effective against the natural enemies of

A. caprone ?

Material and Methods

Quercus turbinella is a shrub widely dis-tributed in the chaparral vegetation of Ari-zona- USA (McDougall 1973). Galls werefound throughout the distribution of the hostplant, and were more abundant between 1,200and 1,500 meters altitude (adjacent to high-ways 1-17, and 89 A South) in northern Ari-

The induction of galls by insects has beenshown to be of adaptive nature. Gall forrna-tion may protect its inducers from harshmicroen.vironmental factors, such as dryness,and temperature (Price et al. 1987). In addi-tion, galls can provide enemy-free space tothe gall-inducer, by offering a protective ref-uge against predators and parasites (Askew1961, Hodkinson 1984, Price et al. 1987,Price & Pschorn-Walker, 1988, Fernandes &Price 1992; but see Fernandes et al. 1987,Hawkins 1988).

Several factors may affect the rates ofparasitism on gall-forrning insects. Gall sizemay interfere with parasitoid oviposition suc-cess in an inverse relationship, since oviposi-tors of limited length are unable to reach thelarvae inside the gall chamber (e.g., Weis &Abrahamson 1985, Rossi et al. 1992, Rossi& Stiling 1995). The toughness of gal1 tissuemay also preclude parasitoid attack, ifthe gal1wal1 is not easily penetrable (Cornel1 1983,Craig et al. 1989). Final1y, the thickness ofthe gal1 wal1s is also important to reduce para-sitism rates, since the success of ovipositionby a parasitoid is related to the length of theovipositor and to the depth of the larvae inthe gal1 (Askew 1961, Price 1972,Weis et al.1985, see also BrandI & VidaI1987).

Parasitoids may synchronize their activ-ity pattern with gal1 development, by attack-ing early in the gal1 growing season (Askew1975, Wiebes-Rijks 1982, Jones 1983,Schonrogge et al. 1996). This should be re-lated to their ability to overcome gal1 size and

:.,

Junho. 1990 An. Soc. Entomol. Brasil 28(2)211

zona.The cynipid galls occured on the abaxial

leaf surface, were spherical, glabrous, and

one-chambered with only one gall-makinglarva per chamber (Femandes et al. 1990).Galls varied from pale yellow to red. Youngga1ls are solid, but the centrally located larvalchamber becomes separated from the gallwalls during development. However, the lar-vaI chamber remains attached to the gall wallsby fibers of conductive tissue during gall de-

velopment (see Femandes et al. 1990).Galls were first observed in early June

1987 , at the beginning of their development,and by the end of September alI of the gaIIswere mature. GaIIs belonged to the same co-hort. These gaIIs were induced early in theseason, when plant leaves are flushing. Werandomly removed 100 gaIIs in June, 50 gaIIsin July and 75 gaIIs in August, from severalplants in the same patch. Galls were dissectedand analyzed for survivorship, diameter, waII

thickness, and wall toughness. GaII mortalitywas caused by several different species of

hymenopteran parasitoids. GaII diameter wasmeasured with a caliper (mm), while gaII waIIthickness was measured under a dissectingscope (mm). GaII waII toughness was ob-tained by using a Volander Texture Analyzer ,which measured the force needed to push aneedle probe 2 mm into the gaII at a speed ofI mmls. The device was calibrated in "grams"of force; measurements were converted toNewtons of force, using the conversion fac-tor of I "gram-force" = 9.80665 mn (Craig et

al. 1990). AII these parameters were com-pared between the three months using a one- ...

way ANOV A foIIowed by the Tukey test (Zar

1984).We used linear regression ana1yses to

verify the relationship between gaII diameterand survivorship rates. Since there was norelationship between these variables (r2 =0.005, F = 1.22, p > 0.05, n = 225), we di-

vided gaII survivorship, thickness, and tough-ness into eight classes of diameter at eachmonth. This procedure was conducted in anattempt to transform gaII survivorship from abinomial to a continuous variable. Data on ..v. ~vv". .v v" "'U""'IIL LV I-"""IULJÇ IJ(U(1~II.U'U

..

gaII survivorship were then arcsine-square-root transfonned, and data on gaII thicknessand toughness were Iog transfonned to meet

nonnality (Zar 1984). We used stepwise mul-tiple regression analysis to observe the influ-ence of gall age, galI wall thickness, andtoughness on the survivorship of A. caprone

galls (Zar 1984).

Results

GaIls of A. caprone varied in diameter,waII thickness, and toughness during devel-

opment (Fig. 1). On average, gall diameterand gaII waII toughness were significantlyhigher at the end of maturation period (Au-gust) (ANOVA and Tukey test, Table 1; Fig.

1). Thus, gaII development influenced posi-tively these variables. However, gall wallthickness increased from June to July, de-creased from July to August, and attained asignificantly Iower average at the end of the

gall developmental period (ANOVA andTukey test, Table 1, Fig. 1 ).

GalI survivorship, on the other hand, in-creased with gaII development (Table 1; Fig.2). In June, survivorship rates were very low,with only 2% ofthe galls remaining alive, in-creasing to 10% in July and 29.3 % in Au-gust. The stepwise multiple regression analy-ses showed that only gaII wall toughness pre-sented a significant positive relationship withgall survivorship (r2 = 0.52, F = 13.84, p <0.01, n = 15, Fig. 3). Therefore, aII the other

variables were excluded from the model, as, .they did not show significant relationships

with gall survivorship (p > 0.05, alI). Thus,the higher gall survivorship in the later stageof development may be related to an increasein the toughness of gall waIIs.

Discussion

The growth dynamics of A.caprone expe-rienced a change from thickerand softer wallsat the initial stage of development to thinnerbut harder waIIs later in the development.However, the higher gall waII thickness did..."t " t~ ..d dÇÇ;~;d~. .~ ~.d~I.."- :._;..

214 Fernandes e! al.

~E

.§...aIãíE

.~"O-U)U)aIt:

~u

:i:-

ro~

ro"

Figure 1. Average diameter, wall toughness, and wall thickness ofthe galls of A. caproneon Q. turbinella during three months of study. Statistically significant differences are repre-sented by different letters above the bars (ANOVA and Tukey test, p < 0.05).

not strong enough to overcome gall walltoughness (see also Askew 1965). In addi-tion. Craig et ai. ( 1990) stated that larval in-

attack, since mortality of the earlier galls wasalmost 98 %. Our results suggest that gallwall toughness is more effective as a defense

Table I. One-way ANOVA comparison of variations on gall diameter, gall wall toughness,thickness and survivorship rates of galls of A. caprone on Q. turbinella between the three

months of study.

-

p

-

DF

-

FVariable

88.7326.1335.8416.42

< 0.0001< 0.0001< 0.0001< 0.0001

2222

Gall DiameterGall wall toughnessGall wall thicknessGall survivorship

against natural enemies in A. caprone galls,assuming that rates of parasitoid attack aresimilar over gall developmental period. In

fact, even longer ovipositors may generallyhe found hroken off in manv ~alls. if thev are

.

accessibility (i.e., larval depth inside the gall)is less important than gall wall toughness in

determining parasitism rates.The existence of this window of vulner-

abilitv mav decrease the range of resource

An. Soc. Entomol. Brasil 28(2) 21~.!unh(J. !999

b35

30-

~-

O)-m..

ao

.cU!..o

.~>..~

(I)

25

20

a15

10

a

n

AugJul

Month

Jun

Figure 2. Average mortality rates of the galls of A. caprone on Q. turbinella during threemonths of study. Statistically significant differences are represented by different letters above

the bars (ANOVA and Tukey test, p < 0.05).

avai1abi1ity, decreasing the efficiency of natu-

ral enemies (Craig et al. 1990). However,Schonrogge et al. ( 1996) suggested that

parasitoids are morpho1ogica11y, pheno1ogi-ca11y and/or physio1ogica11y we11 adapted to ,

their host ga11s, being ab1e to circumvent any,protective function offered by the ga11s. Thus,gall wal1 toughness may diminish theparasitoid attack, but is far from prec1uding itcompletely (ga11 mortality, even in the thirdmonth, remained higher than 70 % ).

The higher ga11 mortality at the initia1 stageof deve1opment is probably due to a highersuccess of oviposition by natural enemiesduring this phase. This pattern was already

observed in other systems (Wiebes-Rijks1982, Craig et al. 1990, Shorthouse et al.1990). Since ga11 wa11 thickness is not effi-

cient in precluding parasitoid attack, and galltoughness is stilllow on the first month aftergall induction, this should be the most criticalphase ofgall establishm:ent. The galls remain-ing alive in the second and third months of

.development have tougher waQs. possiblyexcluding certain species of parasitoids withovipositors of limited strength, and this fac-tor seems to determine the vulnerability of thegalls. In this way, if the larvae of galling in-sects escape detection when galls are soft andwithin reach of parasitoids, they are protectedonce imbedded inside hard maturing galls.

We concluded that gall wall thickness ofA. caprone can not avoid the "window of vul-

nerability" in the initial phase of development.Thus, the mortality rate~ ofthe cynipid larvaeare very high in this stage. probably due to a

216 Femandes et al.

60

50

40

30

20

l

Do

.c

~O

.~>...:J<n

ro~

10

o1.0 1.20.2 0.4 0.6 0.8

Gall wall toughness (N)

Figure 3. Linear regression analyses showing the relationship between survivorship ratesand wall toughness of A. caprone galls (r= 0.52, y = 0.006x + 0.163).

Literature Citedhigher oviposition success by parasitoids.Gall wall toughness seems to limit larva sus-ceptibility to parasitism and, as this param-eter increased with gall age, galls in the Iateseason are less likely to be affected by thepressures of natural enemies.

Askew, R. R. 1961. On the bio1ogy of theinhabitants of oak ga11s of Cynipidae(Hymenoptera) in Britain. Trans. Soc.Brit. Entomo1. 14: 237-258.

Askew, R. R. 1965. The biology of the

British species of the genus Torymus(Hymenoptera: Torymidae) associatedwith the galls of the Cynipidae(Hymenoptera) on oak with specialreference to alternation of forms. Trans.Soc. Brit. Entomol. 16: 217-232.

Acknowledgementsi.

We are very grateful to D Yanega for hishelpful comments on the manuscript and twoanonymous reviewers that carefully reviewedand commented on drafts of this work.Logistical support by the US Forest Service,and Bilby Research Center are greatly appre-ciated. We also thank the US Forest Servicefor allowing us to work on their land. This

project was supported by CNPq-200.747/84-3-Z0, CNPq-521772/95-8, FAPEMIG-078/91, 1950/95 and a Sigma-Xi Grants-in-Aid

to GWF.

Askew, R. R. 1975. The organization ofchalcid-dominated parasitoid commu-

nities centred upon endophytic hosts. p.130-153. In P. W. Price (ed.). Evo-lutionary strategies of parasitic insectsand mites. Plenum. New York. USA. "'f'

An. Soc. Entomol. Brasil 28(2) 217Junho. 1999

59-77. In T. N. Ananthakrishnan (ed.),Biology of gall insects. Oxford & IBHPubl. Co.. India.

Brandi, R. & S. Vidal. 1987. Ovipositorlength in parasitoid and tentifonn leafmines: adaptations in eulophids(Hymenoptera: Chalcidoidea). Biol. J .Linno Soco 32: 351-355. Jones,D. 1983. Theinfluenceofhostdensity

and gall shape on the survivorship ofDiastropus kincaidii Gill (Hymenoptera:Cynipidae). Can. I. Zool. 61: 2138-2142.

Cornell, H. V. 1983. The secondarychemistry and complex morphology ofgalls formed by the Cynipidae(Hymenoptera): why and how? Amer.Midl. Nat. 110: 225-234.

MacDougall, W. B. 1973. Seed plants ofNorthern Arizona. The Museum of

Northern Arizona, Flagstaff, Arizona,USA, 594 pp.Craig, T. P.,J. K. Itami & P. W. Price. 1989.

A strong relationship between ovipositionpreference and larval perfonnance in ashoot-galling sawfly. Ecology 70: 1691-1699.

Price, P. W. 1972. Parasitoids uti1izing thesame host: adaptive nature of differencesin size and form. Eco1ogy 53: 190-195.

Craig, T. P., J. K. Itami & P. W. Price. 1990.The window of vu1nerabi1ity of a shoot-galling sawfly to attack by a parasitoid.Eco1ogy 71: 1471-1482.

Price, P. W., G. W. Fernandes & G. L.Waring. 1987. Adaptive nature of insectga11s. Environ. Entomo1. 16: 15-24.

Price, P. W. &H. Pschorn-Walker. 1988.

Are ga11ing insects better protectedagainst parasitoids than exposed feeders?A test using tenthredinid sawflies. Eco1.

Entomo1. 13: 195-205.

Fernandes, G. W., R. P. Martins & E.Tameirão-Neto.1987. Food webrelationships invo1ving Anadiplosis sp.ga11s (Diptera: Cecidomyiidae) onMachaerium aculeatum (Leguminosae).Rev. Bras. Bot. 10: 117-123. Rossi, A. M., P. D. Stiling, D. R. Strong &

D. M. Johnson. 1992. Does galldiameter affect the parasitism rate ofAsphondylia borrichiae (Diptera:Cecidomyiidae? Ecol. Entomol. 17: 159-1CA

Fernandes, G. W., R. W. Preszler & J. N.Grim. 1990. The occurrence of crystalsin a cynipid leaf gall on Quercusturbinella. Beitr. Biol. Pflanz. 65: 377-383.

Rossi, A. M. & P. D. Stiling. 1995.Intraspecific variation in growth rate, sizeand parasitism of galls induced byAsphondylia borrichiae (Diptera:Cecidomyiidae) on three host species.Ecol. Pop. Biol. 88: 39-44.

Fernandes, G. W. & P. W. Price. 1992. The j,.adaptive significance of insect ga11distribution: survivorship of species in

xeric and mesic habitats. Oeco1ogia90:14-20.

Hawkins, B. A. 1988. Do galls protectendophytic herbivores from parasitoids?A comparison of galling and non-gallingDiptera. Ecol. Entomol. 13: 473-477.

Schonrogge, K., G. N. Stone & M. J.Crawley. 1996. Abundance patterns andspecies richness of the parasitoids and

inquilines of the alien gall-former

Andricus quercuscalicis (Hymenoptera:Cynipidae). Oikos 77: 507-518.

Hodkinson, I. D. 1984. The biology of the

gall-forrning Psylloidea (Homoptera), p.

.

218 Fernandes etal.

Shorthouse, J. O., I. F. Mackay, & T. J.Zmijowskyj. 1990. Role of parasitoidsassociated with gall induced by Remadasnubilipennis (Hymenoptera: Pteromali-dae) on lowbush blueberry. Environ.Entomol. 19: 911-915.

Eurytoma gigantea. Ecol. Ent9mol. 10:341-348.

Wiebes-Rijks, A. A. 1982. Ear1y parasitationof oak-app1e ga11s (Cynipis quercusfoliiL., Hymenoptera). Nether. J. Zool. 32:112-116.

Weis, A. E. & W. G. Abrahamson. 1985.Potential selective pressures by pa-rasitoids on a plant-herbivore interaction.Ecology 66: 1261-1269.

Zar, j. H. 1984. Biostatistical Analysis,2nd

ed., Prentice-Hall, Englewood Cliffs, N.

I., 620 pp.

Weis, A. E., W. G. Abrahamson & K. D.McCrea. 1985. Host gall size and ovi-oosition success bv the oarasitoid Received 16/11/98. Accepted 10/111/99.

...

.