ORIGINAL PAPER

Herbivory on Handroanthus ochraceus (Bignoniaceae)along a successional gradient in a tropical dry forest

Jhonathan O. Silva • Mario M. Espırito-Santo •

Geraldo A. Melo

Received: 23 August 2010 / Accepted: 16 September 2011 / Published online: 15 October 2011

� Springer Science+Business Media B.V. 2011

Abstract This study determined the temporal patterns of

herbivory on Handroanthus ochraceus (Cham.) Mattos

(Bignoniaceae) along a successional gradient in a seasonally

dry tropical forest (SDTF) in southeastern Brazil. We

assessed the diversity of free-feeding herbivore insects (sap-

suckers and leaf-chewers), leaf herbivory rates, leaf nitrogen

content, phenolic compounds, and spider abundance through

the rainy season in three different successional stages (early,

intermediate, and late). Sampling was conducted in

December, at the beginning of the rainy season (with fully

expanded young leaves), February (mid-aged leaves), and

April, at the end of rainy season (old leaves). Fifteen

reproductive trees of H. ochraceus were sampled per suc-

cessional stage in each month of sampling. Herbivore

diversity was highest in the early stage of succession, but

herbivory rates were highest in the intermediate and late

stages. This result was probably related to differences in

herbivore community composition and leaf quality across

successional stages. The highest herbivore abundance was

found in April in the early successional stage. In addition, we

found low levels of herbivory in the intermediate and late

successional stages in the second half of the rainy season.

For each successional stage, leaf nitrogen content decreased

through the rainy season, whereas the concentration of

phenolic compounds increased. For the intermediate and

late successional stages, the temporal changes that took

place as the rainy season progressed corroborated the fol-

lowing hypotheses postulated for SDTFs: (1) both the

abundance of chewing insects and herbivory rates

decreased, (2) the abundance of natural enemies (i.e., spi-

ders) increased, and (3) leaf quality decreased. These results

suggest that the described herbivory patterns are robust for

advanced successional stages (intermediate and late) of the

SDTFs, but may not apply to early successional stages of

these forests.

Keywords Chemical defenses � Deciduous forest �Herbivores � Insects � Natural regeneration

Introduction

The intensity of herbivore damage to their host plants is

influenced by many bottom-up and top-down factors

(Rhoades 1979; Mattson 1980; Price 1997; Stiling and

Moon 2005). These factors vary with time and across space

according to changes in habitat conditions, such as those

observed during ecological succession. The successional

process in tropical forests is characterized by an increase in

ecosystem structural complexity, with a reduction in the

amount of light reaching lower strata and enhanced plant

competition for soil nutrients (Davidson 1993; Poorter

et al. 2004; Pezzini et al. 2008; Madeira et al. 2009). The

availability of these resources is severely limiting to plant

metabolism, affecting the growth rates and palatability of

plants to herbivores (Mattson 1980; Bryant et al. 1983;

Coley et al. 1985; Herms and Mattson 1992).

The resource availability hypothesis (Coley et al. 1985)

predicts that plants in early successional stages, with light

Handling Editor: Gimme Walter.

J. O. Silva (&) � M. M. Espırito-Santo � G. A. Melo

Departamento de Biologia Geral, Centro de Ciencias Biologicas

e da Saude, Universidade Estadual de Montes Claros, CP 126,

Montes Claros, MG 39401-089, Brazil

e-mail: jhonathanos@gmail.com

J. O. Silva

Departamento de Ecologia, Instituto de Ciencias Biologicas,

Universidade de Brasılia, Brasılia, DF, Brazil

123

Arthropod-Plant Interactions (2012) 6:45–57

DOI 10.1007/s11829-011-9160-5

readily available but nutrient availability low, have ele-

vated rates of carbon fixation and low investment in

chemical defenses. The hypothesis suggests that plants in

early successional stages grow faster and more easily

replace tissues lost to herbivores than their advanced suc-

cessional counterparts. This phenotypic outcome may

result from inherent differences among species (Poorter

et al. 2004) and/or plastic responses of the plants to dif-

ferent environmental conditions (Bryant et al. 1983; Coley

et al. 1985; Herms and Mattson 1992). According to pre-

dictions of the resource availability hypothesis, trees of the

same species occurring in different successional habitats

are likely to differ in their chemical defense levels and

herbivory rates, with early successional individuals suf-

fering higher damage from herbivores because of their

lower phenolic and higher nitrogen leaf content.

In addition to the effects of host plant quality change,

insect herbivore performance is also affected by modifi-

cations in forest structural complexity along the succession

(Brown and Ewel 1987; Kalacska et al. 2004; Lewinsohn

et al. 2005; Madeira et al. 2009). Advanced forest suc-

cessional stages harbor higher plant richness and more

vertical strata, providing higher niche diversity and better

microclimatic conditions (Basset et al. 2003) for both

herbivores and their natural enemies, such as parasitoids

and predators (Ernest 1989; Lewinsohn et al. 2005). Her-

bivores thus possibly suffer stronger top-down influences

in more complex, late successional habitats. Anyway, the

strength of herbivore–natural enemy interaction also

depends on intrinsic species traits (Sanders et al. 2008) and

probably varies between different insect herbivore guilds.

Though many studies have addressed herbivory patterns

along successional gradients (Ernest 1989; Davidson 1993;

Siemann et al. 1999; Gruner and Polhemus 2003), practi-

cally nothing is known about this process in seasonally dry

tropical forests (SDTFs) (Quesada et al. 2009). In this

vegetation type, most plant species lose all their leaves

during the dry season and this short lifespan may be

associated with low investment in chemical defenses and

high herbivory rates (Dirzo and Boege 2008). Some

authors have argued that the more constant selection

pressure imposed by herbivores in the perennial vegetation

of wet forests have resulted in plants in these environments

evolving stronger chemical resistance to leaf herbivory

(Stanton 1975; Dirzo and Domınguez 1995; Coley and

Barone 1996; Dirzo and Boege 2008). Other distinctive

characteristics that may affect herbivory patterns across

successional stages in SDTFs are the strong synchroniza-

tion in resource production (i.e., young tissue) and the

limited resource availability through the year (Mendonca

2001), since plants are leafless during the entire dry season

(5–6 months). Differences in the timing of leaf production

between successional stages have already been reported

(Pezzini et al. 2008) and may have profound effects on

herbivores in seasonal environments.

A consequence of the strong seasonality in leaf pro-

duction in SDTFs is that herbivory is concentrated in the

rainy season, though insect herbivore damage rates may

vary significantly during this period (Janzen 1981; Filip

et al. 1995). Several studies reported a peak of folivore

activity during the first half of the rainy season, followed

by a decrease in herbivory rates toward the end of this

period (Janzen 1981; Filip et al. 1995; Boege 2004). This

pattern is attributed to synchronization between herbivore

attack and host plant production of new, highly nutritious

tissues (Janzen and Waterman 1984; Filip et al. 1995;

Boege 2005) and to a possible temporal escape from nat-

ural enemies (Dirzo and Domınguez 1995). Some studies

demonstrated a reduction in plant water and nitrogen

content and/or an increased phenolic concentration with

leaf age as the rainy season progressed, thus reducing tissue

palatability and herbivore consumption rates (Boege 2004;

Boege 2005). Also, natural enemy abundance may lag

herbivore density, with lower predation rates early in the

rainy season, followed by an intensification of predation

rates during the second half of this season.

In this study, we investigated leaf chemistry, spider

abundance, herbivore diversity, and the level of herbivory

on Handroanthus ochraceus (Cham.) Mattos (Bignonia-

ceae) over time and across a successional gradient in a

SDTF in southeastern Brazil. The following questions were

addressed: (1) How do herbivore diversity and leaf damage

vary over time and across early, intermediate and late

successional stages? (2) How do spider abundance and leaf

chemistry change over time and across early, intermediate

and late successional stages?

Materials and methods

Study area and system

This study was conducted in the Mata Seca State Park, a

conservation unit of restricted use created in 2000 with

10,281.44 ha located in Manga, Minas Gerais, between

14�4803600–14�5605900S and 43�5501200–44�0401200W. Its

climate is considered tropical semiarid (Koppen’s classifi-

cation) and is characterized by dry winters. The average

temperature in the region is 24�C, with an average annual

precipitation of 871 mm (Antunes 1994). The original

vegetation of the park is seasonally dry tropical forests,

growing on plain and nutrient-rich soils (IEF 2000). The dry

season extends from May to October (Fig. 1) when most

trees lose 90–100% of their leaves (Pezzini et al. 2008).

About 1,525 ha of the park are covered with abandoned

pasture fields in different stages of regeneration, and the

46 J. O. Silva et al.

123

remaining area is a mosaic of secondary and primary dry

forests (IEF 2000; Madeira et al. 2009). Our samples were

taken in forest fragments categorized as early, intermedi-

ate, and late successional stages. The early successional

fragment was used as pasture for at least 20 years and

abandoned in 2000. It is characterized by herbaceous–

shrubby vegetation with sparse trees forming a discontin-

uous canopy up to four meters in height. The intermediate

successional fragment was used as pasture for an unknown

period and abandoned at the late 1980s. It has two vertical

strata: the first is composed of 10–12 m tall deciduous trees

and the second comprises a dense understory with many

young trees and abundant lianas. The late successional

fragment has no records of clear-cutting for the last

50 years (Madeira et al. 2009) and is also characterized by

two strata. The first stratum is composed of deciduous trees

that form a closed canopy 18–20 m high. The second

stratum comprises a sparse understory of low liana density.

Handroanthus ochraceus (Bignoniaceae), commonly

known in Brazil as ‘‘ipe-amarelo,’’ is a perennial species

reaching up to 20 m and found from Argentina to Mexico

(Gentry 1992). It was previously known as Tabebuia

ochracea (Cham.) (Grose and Olmstead 2007). This tree

occurs in all successional stages in the park, where it is

ranked the third species in Importance Value Index (IVI—

combines density, basal area and frequency; see Madeira

et al. 2009). In SDTF ecosystems, H. ochraceus individuals

usually flush their leaves in November (the beginning

of the rainy season), reaching full expansion in

December (Pezzini et al. 2008). Leaves fall between May

and September (dry season), and flowering and fruiting

usually occur at the end of the dry season (Pezzini et al.

2008). In the Brazilian Cerrado, some studies have been

conducted with various Handroanthus species, especially

on H. ochraceus (Ribeiro and Brown 1999; Ribeiro and

Brown 2006). However, little is known about this species

in deciduous forests, a vegetation type where it is among

the most abundant tree species (Sullivan 2000; Madeira

et al. 2009; Moreira et al. 2009). Since it occurs along the

entire successional gradient, H. ochraceus allows the

evaluation of the effects of forest succession on insect

herbivore richness and abundance and intensity of herbiv-

ory using a single host plant, eliminating possible idio-

syncrasies of multiple host species comparisons.

Arthropod and leaf sampling

Sampling was initiated at the beginning of the rainy season,

in December 2007, when leaves were newly formed, fol-

lowed by collections in February and April, representing

the middle and the end of the rainy season, respectively. In

each successional stage, sampling took place in three plots

(50 m 9 20 m) at least 200 m apart from each other.

Based on the plant–herbivore interactions literature, this

distance is considered long enough to avoid pseudo-repli-

cation between plots (see Fernandes et al. 2004; Santos

et al. 2011). The floristic composition of each plot has been

characterized (see Madeira et al. 2009). For each succes-

sional stage, 15 H. ochraceus trees were randomly marked

(five per plot), all at least 10 m from each other. Tree

height was 3.91 ± 0.150 m in the early stage and

8.37 ± 0.645 m in the late stage and all were reproductive

trees. Also, a cohort of leaves from different branches was

arbitrarily marked at crown height of each individual in

December. Nine leaves from this cohort were removed per

sampling period, totaling 27 leaves per tree. Leaves were

carefully removed at the base of their petioles to avoid the

rupture of the leaf blade and consequent phenolic oxida-

tion. They were immediately photographed in the field

against a white board with one centimeter marks as scale

reference to quantify leaf damage. Photography of leaves in

situ and across time was not possible, as any manipulation

caused them to detach. Leaves were stored individually in

dark paper bags for transport to the laboratory.

The percentage of leaf damage was calculated per plant

using the software ImageJ (Rasband 2006). Each image

was processed through a binary conversion (black/white),

with the scale set to centimeters (distance calibration). In

this way, it was possible to determine the number of pixels

present in one centimeter (cm) and extrapolate to the

picture area (cm2). Then, we calculated the total leaf area

and the area removed by herbivores. The percentage of leaf

damage was obtained by [(removed area/total area)9100]

and averaged per plant.

The free-feeding insects were collected by beating three

branches of each sampled tree (each about 40 cm long) with

an entomological umbrella (see Campos et al. 2006; Neves

et al. 2010a). The sampled shoots were similar in size and

0

20

40

60

80

100

120

140

160

180

200

0

5

10

15

20

25

30

35"2007-2008"

Historical

Temperature

Tot

al r

ainf

all (

mm

)

Mea

n te

mpe

ratu

re (

°C)

Aug Sept Oct Nov Dec Jan Feb Mar Apr Jun JulMay

Fig. 1 Monthly total rainfall during 2007–2008 (filled circle),

historical average precipitation (open circle), and monthly average

temperature (filled triangle) during 1939–2006 in the study site. Data

from Mocambinho Meteorological Station, 15 km from Mata Seca

State Park

Leaf herbivory in a tropical dry forest 47

123

were situated in different places of the crown, to homogenize

the effects of the different microclimatic conditions

across the canopy. The insects were grouped into two guilds

(sap-sucking and chewing) (Moran and Southwood 1982),

separated into morphospecies and identified at the family

level. We considered as part of the sap-sucking guild all

adult and juvenile Hemiptera (Auchenorryncha, Sternor-

ryncha, and Heteroptera). The chewing guild was composed

by adult and juvenile Coleoptera, Orthoptera, Phasmida, and

juveniles of Lepidoptera (Borror et al. 2002). Among the

predatory arthropods, only spider abundance was quantified,

due to the limited efficacy of the entomological umbrella to

collect other predators. All the insects and spiders collected

from the three shoots were grouped per plant for statistical

analyses. Throughout the text, we used the term ‘‘insect

diversity’’ to refer to insect richness and abundance.

Leaf chemistry

To evaluate tree foliage quality, we quantified two types of

leaf secondary metabolites: total phenolics and tannins, as

these are important herbivory deterrents in SDTFs (Boege

2004; Boege 2005; Dirzo and Boege 2008). To assess plant

nutritional quality, we quantified leaf nitrogen content,

since this nutrient is considered the most limiting for her-

bivorous insects (Mattson 1980; Stiling and Moon 2005).

Also, there are several quantification methods for these

compounds that are quick, relatively inexpensive, and

widely used in ecological studies (see Cornelissen and

Fernandes 2001; Boege 2005; Silva et al. 2009).

For each sampling period, nine leaves per plant were dried

at 70�C for 96 h. Since a single leaf usually did not provide

enough dry matter for all chemical analyses, we haphazardly

allocated three leaves to each of three subsamples. The total

phenolic content was determined through the Folin-Dennis

assay (Swain and Hillis 1959), using quercetin as standard.

The same extracts were used to quantify condensed and

hydrolyzable tannins with the radial diffusion method (Ha-

german 1987) and tannic acid as standard. Leaf nitrogen

content was determined with the same dry leaves, digested

using the micro-kjeldahl method, followed by the distillation

and titration of the resulting solution (Allen et al. 1974).

Total phenolic and nitrogen contents from the three subs-

amples were averaged per plant for statistical analyses. No

tannins were recorded in H. ochraceus leaves (with the

detection level of the radial diffusion method); so, this var-

iable was not included in further analyses.

Statistical analyses

We calculated two types of herbivory indicators: (1) the

mean percentage of leaf damage per plant at each interval

and (2) the herbivory increment, given by the difference in

the percentage of leaf damage between two subsequent

periods (December–February and February–April) for each

individual tree, determined through the subtraction of the

average percentage of leaf damage in a given period by the

previous one. To calculate the herbivory increment during

the first period (from leaf production to December), we

assumed that the percentage of leaf damage at leaf pro-

duction was zero. Thus, for the first period, both herbivory

indicators had the same values. This analysis allowed us to

assess at which point of the rainy season periods leaf

consumption was highest.

To verify the effects of the successional stages and time on

the mean percentage of leaf damage, herbivory increment, leaf

total phenolic and nitrogen contents, richness and abundance

of herbivore insects, and spider abundance, we adjusted a

linear mixed-effect model (LME) for each of these response

variables (Crawley 2002). These analyses were employed

because the data were obtained repeatedly from the same

plants at different times and the temporal autocorrelation

generated by consecutive counting violates the assumption of

sampling independence. Assuming independence when it is

not true would inflate the error degrees of freedom and could

lead to spurious significance (Type I error) (Crawley 2002).

To overcome this problem, the data were grouped by plant and

the error variances were calculated for each different group.

The response is thus not the individual measure, but the

sequence of measures in an individual (Crawley 2002). The

resulting groups were treated as a random effect, whereas the

successional stage, sampling date, and the interaction between

successional stage-sampling date were used as explanatory

variables (fixed effects).

To test for the effects of predator abundance and leaf

chemistry on herbivore diversity and herbivory increment,

we also used LME. In this case, spider abundance, phenolic

and leaf nitrogen contents, and the interactions between

these factors with the successional stage were used as

explanatory variables. Herbivore abundance and herbivory

increment were log-transformed to meet normality and

used as response variables. All LME models were com-

pared with null models, and minimal adequate models were

adjusted with the removal of the nonsignificant terms.

Finally, the junction of nonsignificant categorical groups

(amalgamation) was performed by contrast analyses. All

model construction and analyses were conducted using the

software R.2.10 (The R Development Core Team 2009).

Results

Herbivore diversity and predator abundance

We sampled 825 free-feeding herbivore insects distributed

across 104 morphospecies. The chewing guild was

48 J. O. Silva et al.

123

composed of 453 individuals (60 morphospecies), whereas

372 individuals (44 morphospecies) were sap-sucking

insects. Most herbivores were collected in the early suc-

cessional stage: 404 individuals (47 morphospecies), fol-

lowed by the intermediate and late stages, with 224 and

197 individuals belonging to 58 and 45 morphospecies,

respectively (Table 1). Among the chewing insects, the

most abundant species was Oedionychus sp.1 (Coleoptera:

Chrysomelidae), which dominated the samples, especially

in the early successional stage, where its abundance

increased substantially through the rainy season. The pro-

portion of orthopteroid herbivores (Orthoptera and Phas-

mida) declined through the rainy season in the intermediate

and late successional stages, whereas their abundance

increased in the early successional stage (Table 1). Among

the sap-sucking insects, immature stage individuals were

most abundant, particularly in the early and intermediate

forests but with contrasting temporal patterns between

these successional stages. Among the adult sap-sucking

insects, the most common morphospecies were Rhabdot-

alebra sp. 1 (Homoptera: Cicadellidae) and Tingis stecoma

(Hemiptera: Tingidae), particularly in the early succes-

sional stage where they increased in abundance through the

rainy season (Table 1).

As a whole, the average abundance and richness of

chewing insects per tree were significantly higher in the

early successional stages than in the later stages

(‘‘Appendix’’). In the intermediate and late stages, both the

abundance and richness of this guild decreased in the first

half of the rainy season (December–February), but no sig-

nificant changes were observed in the second half (Febru-

ary–April). The same pattern was observed for the early

stage during the first half of the rainy season, but the

abundance and richness of chewing insects increased sig-

nificantly at the end of this period (‘‘Appendix’’; Fig. 2a, b).

The spatial and temporal patterns of abundance and

richness of sap-sucking bugs were quite different from

those of the chewing insects. The abundance of sap-suck-

ing insects was lower in early succession vegetation, but

increased strongly in April. A steady, significant increase

in the abundance of this guild was evident through the

rainy season in both the early and late stages of succession,

but no temporal variation was observed in the intermediate

stage (‘‘Appendix’’; Fig. 3a). As a whole, sap-sucking bug

richness did not vary between successional stages

(‘‘Appendix’’), and no clear temporal pattern was observed

through the rainy season (Fig. 3b). The richness of sap-

sucking bugs increased steadily from December to April in

the early stage of succession, decreased slightly in the

intermediate stage, and remained constant in the late stage

(‘‘Appendix’’; Fig. 3ab).

A total of 163 spiders were collected on H. ochraceus.

The abundance of these predators did not vary significantly

between successional stages, but showed a marked increase

as the rainy season progressed (‘‘Appendix’’; Fig. 4). In

early succession, the highest increase occurred from

Table 1 Free-feeding herbivorous insects associated with Handroanthus ochraceus through the rainy season in three successional stages (early,

intermediate, and late) in a tropical dry forest, Minas Gerais, Brazil

Taxonomic group Successional stage

Early Intermediate Late

December February April December February April December February April

Chewing insects

Chrysomelidae 11 (8) 08 (7) 06 (5) 11 (7) 08 (5) 04 (4) 08 (6) 05 (4) 03 (2)

Oedionychus sp. 1 08 20 93 03 07 05 02 01 07

Curculionidae 08 (5) 01 (1) 07 (4) 13 (8) 06 (2) 05 (7) 05 (4) 06 (3) 03 (2)

Orthopteroids 14 (2) 15 (3) 29 (8) 13 (4) 11 (3) 04 (3) 39 (8) 12 (5) 06 (3)

Other 11 (19) 05 (5) 05 (3) 04 (4) 02 (2) 03 (3) 02 (2) 03 (2) 11 (9)

Total chewing individuals 52 (35) 49 (17) 140(21) 44 (24) 34 (12) 21 (18) 56 (21) 27 (15) 30 (17)

Sap-sucking insects

Cicadellidae 00 03 (2) 04 (3) 00 00 05 (2) 01 (1) 00 00

Rhabdotalebra sp.1 04 09 23 00 00 07 00 00 02

Tingidae 00 00 04 (3) 00 01 (1) 01 (1) 00 00 13 (4)

Tingis stecoma 01 10 18 01 00 01 04 09 05

Immature stages 09 (5) 08 (6) 34 (12) 22 (14) 24 (8) 04 (3) 06 (5) 07 (2) 12 (8)

Other 03 (2) 05 (4) 28 (18) 25 (16) 16 (11) 18 (13) 12 (7) 05 (5) 08 (07)

Total sap-sucking individuals 17 (9) 35 (14) 111(38) 48 (31) 41 (20) 36 (21) 23 (14) 21 (8) 40 (21)

Total 69 84 251 92 75 57 79 48 70

The number of morphospecies is given in brackets

Leaf herbivory in a tropical dry forest 49

123

February to April (second half), whereas in the interme-

diate and late stages, a huge increment in spider abundance

took place during the first half of the rainy season (Fig. 4).

Leaf damage

The mean percentage of leaf damage per plant was higher

in intermediate and late successional stages than in the

early stage (‘‘Appendix’’; Fig. 5a). The herbivory incre-

ment varied through the season and between successional

stages (‘‘Appendix’’; Fig. 5b). In the early successional

stage, a higher herbivory increment was detected from

February to April, whereas in the intermediate and late

0

2

4

6

8

10

12

0

0.5

1

1.5

2

2.5

3

3.5

4

4.5

(a)

(b)

a a

b

a a

b

bb

c

d

bc

bb

c

a

b

c

Early

Intermediate

Late

Che

win

g ab

unda

nce

Che

win

g ric

hnes

s

December February April

Fig. 2 Mean a abundance and b richness of chewing insects

(±standard error) per tree on Handroanthus ochraceus through the

rainy season in three successional stages: early, intermediate, and late

(n = 15 per stage). Different letters above the bars represent

statistically different means

0

2

4

6

8

10

12

0

0.5

1

1.5

2

2.5

3

3.5

4

4.5

5

(a)

(b)

a

b

c cb

c

d

b

e

a

b

c c cc c

c

d

Sap

-suc

king

abu

ndan

ceS

ap-s

ucki

ng r

ichn

ess

EarlyIntermediateLate

December February April

Fig. 3 Mean a abundance and b richness of sap-sucking insects

(±standard error) per tree on Handroanthus ochraceus through the

rainy season in three successional stages: early, intermediate, and late

(n = 15 per stage). Different letters above the bars represent

statistically different means

a

a a

a

bb

b

b b

0

0.5

1

1.5

2

2.5

3 Early

Intermediate

Late

Spi

der

abun

danc

e

December February April

Fig. 4 Mean abundance of spiders (±standard error) per tree on

Handroanthus ochraceus through the rainy season in three succes-

sional stages: early, intermediate, and late (n = 15 per stage).

Different letters above the bars represent statistically different means

0

2

4

6

8

10

12

0

2

4

6

8

10

12

14(a)

(b)

a aa

b

c

c c

d

e

Her

bivo

ry in

crem

ent (

%)

Leaf

dam

age

(%)

December February April

December -February February -April

a

b

c c c

a

Early

Intermediate

Late

Fig. 5 a Mean percentage of leaf damage and b herbivory increment

(±standard error) per tree on Handroanthus ochraceus through the

rainy season as assessed in each of three successional stages: early,

intermediate, and late (n = 15 per stage). Different letters above the

bars represent statistically different means

50 J. O. Silva et al.

123

successional stages, a higher herbivory increment was

recorded between December and February (Fig. 5b).

Leaf chemistry

The concentration of phenolic compounds in H. ochraceus

did not differ between successional stages, but showed a

significant increase through the rainy season (‘‘Appendix’’;

Fig. 6a). In general, leaves in all successional stages had

low phenolic concentrations in December but these were

significantly higher in February and April (Fig. 6a). The

leaf nitrogen content differed between successional stages,

as follows: early \ intermediate \ late (‘‘Appendix’’;

Fig. 6b). Moreover, the nitrogen content decreased signif-

icantly during the rainy season for all stages (‘‘Appendix’’;

Fig. 6b).

Plant defense and predator effects

The abundance and richness of chewing insects showed a

significant relationship with leaf nitrogen content, but these

variables were not correlated with concentration of phe-

nolic compounds and predator abundance (‘‘Appendix’’).

The sap-sucking insect abundance showed no significant

relationship with the tested variables, whereas the richness

of these organisms was negatively correlated with spider

abundance (‘‘Appendix’’). Herbivory increment was posi-

tively correlated with leaf nitrogen content and concen-

tration of phenolic compounds, but this variable was not

correlated with abundance of chewing insects and spiders

(‘‘Appendix’’).

Discussion

Insect community composition

The composition of herbivore families found on H. ochraceus

corroborates the diversity patterns described in other studies

conducted in tropical forests, with sap-sucking cicadellids

and folivorous chrysomelids being the most abundant canopy

herbivorous insects (Basset et al. 2001; Campos et al. 2006;

Varanda and Pais 2006; Neves et al. 2010a). The most

abundant herbivore species recorded here, such as Oedio-

nychus sp.1, Rhabdotalebra sp. 1 and T. stecoma, were also

dominant on H. ochraceus individuals in the Brazilian

Cerrado (Ribeiro and Pimenta 1991; Ribeiro et al. 1994)

and on T. aurea in the Cerrado and Pantanal (Ribeiro

1998). The Rhabdotalebra individuals are common on

fully expanded, mature leaves of H. ochraceus, apparently

because of their lower trichome density that facilitates

herbivore locomotion and feeding (Ribeiro et al. 1994).

Thus, H. ochraceus in the SDTF we studied seems to

support a widely distributed fauna of oligophagous free-

feeding herbivores that are commonly associated with

congeneric plants in other Brazilian vegetation types.

Rainy season patterns

In general, the temporal patterns of insect diversity and leaf

damage in H. ochraceus partially corroborated the

hypotheses and mechanisms proposed for SDTFs by Janzen

(1981); there was a decrease in the abundance and richness

of chewing insects and in leaf damage through the rainy

season in the intermediate and late stages, coupled with an

increase in predator abundance and a decrease in leaf

quality. However, trends were contradictory for chewing

diversity and leaf damage in the early successional stage

(Fig. 7a, b).

The hypothetical higher insect diversity early in the

rainy season in SDTFs is based on two different processes:

(i) escape from predation, since predator response to insect

proliferation lags in time and (ii) synchronization with the

plant production of highly nutritious, young tissues.

Indeed, the abundance of spiders conformed to this pre-

diction for all successional stages, peaking in the middle of

the rainy season, in February, and may have influenced the

reduction in chewing insect diversity in the intermediate

and late stages of succession. Though we have partial

0

5

10

15

20

25(a)

0

0.5

1

1.5

2

2.5

3(b)

EarlyIntermediate

Late

Phe

nolic

com

poun

ds (

µg/m

L)N

itrog

en c

onte

nt (

%)

December February April

a a a

b b

b

b bb

ab

c

da a

d dd

Fig. 6 a Mean concentration (±standard error) of phenolic com-

pounds (lg/ml) and b nitrogen content (% dry weight) per tree in

Handroanthus ochraceus leaves through the rainy season in three

successional stages: early, intermediate, and late (n = 15 per stage).

Different letters above the bars represent statistically different means

Leaf herbivory in a tropical dry forest 51

123

evidence to corroborate the importance of the top-down

control mechanism, a more precise understanding of

predator effects on temporal patterns of herbivore diversity

will only be achieved through exclusion experiments. If

spiders really affect herbivore populations through time, an

experimental approach would also help clarify whether

such effects vary across the early and advanced succes-

sional stages.

Alternatively, the decline in the abundance of chewing

insects in the intermediate and late successional stages

could be a consequence of the decrease in leaf nitrogen

content observed through the rainy season. In plants with

seasonal growth cycles, peak nitrogen concentrations occur

whenever the cells of a plant tissue or organ are rapidly

expanding in number and size (Mattson 1980; Bryant et al.

1983). As soon as tissue growth begins to wane, nitrogen

levels drop sharply (Herms and Mattson 1992), with a

gradual decline in its concentrations during the course of a

growing season until tissue senescence (Mattson 1980).

Since chewing insects are ‘‘early-season feeders’’ highly

adjusted and limited by leaf nitrogen levels, life cycles that

synchronized with the highest availability of young leaves

were probably selected positively in SDTFs. Thus, our data

also provide partial support to the bottom-up control

mechanism driving temporal variations in herbivore

diversity in SDTFs.

Why chewing insect diversity and herbivory increment

showed opposite trends in the early successional stage is

unclear, since the rainy season increase in spider abun-

dance and decrease in leaf nitrogen were detected for all

stages (Fig. 7). These variations could have been caused by

changes in the herbivore community composition in the

early successional stage in April, with an increase in the

proportion of large orthopteroids usually able to feed on

low-quality tissues (Novotny et al. 2003). Thus, early

successional stages might be more prone to insect out-

breaks (loss of stability) (Ludwig et al. 1997; Swanson

et al. 2011), possibly due to a lower diversity of predators

other than spiders. In fact, Neves et al. (2010b) and Leite

et al. (2008) found lower richness and abundance of

predatory ants and birds, respectively, in the early stage

plots in the Mata Seca State Park. Therefore, advanced

successional stages are probably more buffered against

swings in insect herbivore’s population size due to erratic

temporal distribution of precipitation in SDTFs.

The temporal patterns of abundance and richness of sap-

sucking insects were totally different from chewing insects,

and the only clear trend was the progressive increase in sap-

sucking abundance observed at the early and late succes-

sional stages. This result also contradicts those obtained by

Novotny and Basset (1998), who found a higher abundance

of sap-sucking insects early in the rainy season in several

Ficus (Moraceae) species in a lowland rain forest of New

Guinea. Apparently, the insect abundance in this guild was

not affected by temporal variations in spider abundance and

plant quality, which was reinforced by its lack of significant

relationship with all the explanatory variables tested.

However, it is necessary to confirm these temporal and

successional trends across consecutive years, since climatic

factors are key drivers of insect diversity (Tauber et al.

1986; Wolda 1988; Didham and Springate 2003), and the

intra- and interannual distribution of rain is very unpre-

dictable in semiarid regions such as the one studied here

(Antunes 1994; Pezzini et al. 2008).

Successional patterns

Successional comparisons showed that the early stage

differed from intermediate and late stages for most vari-

ables tested on H. ochraceus. Spider abundance and the

concentration of phenolic compounds did not conform to

this pattern, showing no successional trends (Table 2).

Insect herbivore diversity was higher in the early stage,

whereas herbivory levels were greater in advanced stages.

Nitrogen content was the only variable that increased

gradually along the successional gradient, differing

between the three stages (Table 2).

Despite an increase of up to 92 times in the structural

complexity (Holdridge complexity index) from early to late

Fig. 7 Rainy season patterns of leaf damage, plant defenses, and

abundance of chewing insects and spiders associated with Handro-anthus ochraceus in a early and b intermediate, and late successional

stages in a tropical dry forest. Linear and exponential curves were

plotted based on parameters estimated from the analyses of linear

mixed-effect models

52 J. O. Silva et al.

123

stage plots in the Mata Seca State Park (Madeira et al.

2009), insect herbivore diversity on H. ochraceus was

generally higher in the earliest successional stage studied.

Thus, the increase expected in general forest insect diver-

sity because of higher niche availability or better micro-

climatic conditions in advanced, more complex succes-

sional stages may not be reflected in the fauna associated

with single plant species in those forests. It has also been

proposed that herbivores suffer higher rates of predation in

more complex habitats, because they are expected to har-

bor a higher diversity of natural enemies (Siemann et al.

1999; Sanders et al. 2008; Silva et al. 2010). However, no

statistical differences in spider abundance were observed

between successional stages (except for February, see

Fig. 4). It is possible that other natural enemies not con-

sidered here (e.g., birds and parasitoids) may contribute to

the lower herbivore abundance in intermediate and late

stages, and this deserves further investigation. Before

robust generalizations are made, however, the highest

herbivore abundances recorded on H. ochraceus in the

early successional stages need support from interannual

comparisons using multiple host plant species.

The greatest insect diversity, recorded in the early suc-

cessional stage, was not associated with relatively higher

leaf herbivory levels, and the mean percentage of leaf

damage was consistently greater in intermediate and late

stages. Indeed, no relationship was found between herbivory

increment and the abundance of chewing insects. Several

studies have recorded such an absence of spatial linkage

between the herbivore diversity and herbivory rates (Ernest

1989; Campos et al. 2006; Varanda and Pais 2006). No

mechanistic explanation has been provided, but is probably

related to herbivore diversity being influenced by several

diffuse factors such as climatic variables (temperature, air

humidity, wind speed, and insolation) and complex multi-

trophic interactions (interspecific competition, predation,

parasitism, mutualisms), whereas their consumption rates

are proximately controlled by plant quality (Strong et al.

1984; Didham and Springate 2003; Stiling and Moon 2005).

According to the resource availability hypothesis, it was

expected that individuals of H. ochraceus in the early stage

would have lower phenolic and higher nitrogen leaf con-

tents, thus suffering higher damage from herbivores than in

advanced stages (Coley et al. 1985). However, the opposite

successional patterns were observed (Table 2). In the

present study, herbivory increment was positively corre-

lated with foliar nitrogen content at the plant level, and

plants in the intermediate and late stages have both higher

mean percentage of leaf damage and nitrogen content than

in the early stage. Thus, the unexpected herbivory increase

in H. ochraceus individuals along the successional gradient

is likely a consequence of their increased foliar nitrogen.

Other studies have found a strong positive relationship

between insect herbivory and leaf nitrogen content in dif-

ferent ecosystems, including SDTFs (Mattson 1980; Filip

et al. 1995; Cornelissen and Fernandes 2001; Boege 2004;

Boege 2005; Stiling and Moon 2005; Barret and Stiling

2007). The inverse successional pattern of leaf nitrogen

content observed for H. ochraceus is probably a conse-

quence of the higher soil nutrient content (N, P, K-base

saturation) observed in advanced compared to the early

successional stages at the Mata Seca State Park (unpub-

lished data).

Leaf phenolic concentration did not differ in plants

across the different successional stages, and this result

contradicts the resource availability hypothesis and also

other studies in that they detected an increase in plant

defenses as forest succession progresses (Cates and Orians

1975; Bryant et al. 1983; Davidson 1993; Poorter et al.

2004). It is likely that this classic successional pattern,

typical of tropical rain and temperate forests, does not

apply to STDFs. However, such differences are consistent

with soil nutrient richness and also leaf longevity in the

studied forest. Pezzini (2008) reported that leaf fall is lower

for individuals of H. ochraceus in the early successional

stage, increasing gradually in advanced stages, a pattern

observed for the whole tree community. Thus, the longer-

lived leaves of H. ochraceus, produced under limited

nutrient availability, probably accumulate high levels of

carbon-based compounds. Although phenolic compounds

are well known for their reduction in foliage digestibility,

their effects depend on the specific response of each her-

bivore species and on the type of compound produced

(Close and McArthur 2002). If fact, there was a positive

relationship between herbivory increment and the

concentration of phenolic compounds in individuals of

H. ochraceus, suggesting that these chemicals may have a

limited defensive effect against herbivores in this host

species. Alternatively, phenolic compounds may be

involved in protection against photodamage and/or were

produced as an induced defense (Close and McArthur

2002; Boege 2004). Experimental manipulations are

Table 2 Successional trends for several variables assessed on

Handroanthus ochraceus in a tropical dry forest at the north of Minas

Gerais, Brazil

Variable Successional gradient

1. Richness and abundance of chewing

insects

Early [ intermediate = late

2. Abundance of sap-sucking insects Early [ intermediate = late

3. Richness of sap-sucking insects Early = intermediate = late

4. Leaf damage (%) Early \ intermediate = late

5. Spider abundance Early = intermediate = late

6. Phenolic compounds Early = intermediate = late

7. Nitrogen content Early \ intermediate \ late

Leaf herbivory in a tropical dry forest 53

123

therefore required to verify the production of phenolics as

induced defenses and their role in herbivory deterrence on

H. ochraceus in SDTFs.

Conclusion

This is one of the first studies to investigate temporal

patterns of herbivory along a successional gradient in

SDTFs, including insect herbivore diversity, natural ene-

mies, leaf damage, and plant chemistry. Our results indicate

that rainy season variations can be markedly different from

early to intermediate and late successional forests. Also, the

successional trends detected in the present study did not

corroborate those predicted by the resource availability

hypothesis. The reasons for such a pattern deserve further

investigation, but most hypotheses and theories in herbi-

vore–host plant interactions in tropical wet and temperate

forests were based on studies conducted in intermediate and

old-growth forests (Coley and Barone 1996). Thus, most of

the temporal patterns described in the literature may not

apply to early successional ecosystems, and successional

models in general may be revisited for the poorly studied

SDTFs (see Quesada et al. 2009). Understanding herbivory

variations across time and space in complex environments

such as tropical forests is a difficult task, and long-term

studies with experimental approaches are necessary to

confirm whether such variations are idiosyncratic or deter-

mined by bottom-up and top-down mechanisms.

Acknowledgments We are very grateful to M Fagundes, SP Ribeiro

and FS Neves for their valuable comments on the early versions of

this manuscript. We also thank RA Andrade, KN Oliveira, SF

Magalhaes and A Mendes for field assistance, and SP Ribeiro for

insect identification. Logistical support was provided by the Instituto

Estadual de Florestas (IEF), and the financial support was provided by

Conselho Nacional de Pesquisa—CNPq (474508-07), Fundacao de

Amparo a Pesquisa de Minas Gerais-FAPEMIG (CRA-2288/07 and

CRA APQ-3042-5.03/07), and the Inter-American Institute for Global

Change Research (IAI-CRN II-021). We gratefully acknowledge

FAPEMIG for a MSc scholarship to JO Silva and a research schol-

arship to MM Espırito-Santo. This study was in partial fulfillment of

requirements for the Master degree at Universidade Estadual de

Montes Claros.

Appendix

See Table 3.

Table 3 Analyses of variance of the linear mixed-effect models

constructed to test the effects of successional stage and temporal

variations in the abundance and richness of chewing and sap-sucking

insects, spider abundance, leaf phenolic and nitrogen contents, mean

percentage of leaf damage, and herbivory increment on Handroanthusochraceus (n = 15 per successional stage)

Response variable Explanatory variable nunDF denDF F P

Abundance of chewing insects Intercept 1 88 305.461 \0.0001

Successional stage 2 42 43.521 <0.0001*

Time 2 87 22.670 0.0001*

Spider abundance 1 88 6.718 0.0756

Nitrogen content 1 88 10.816 0.0107*

Phenolic compounds 1 88 2.558 0.3157

Successional stage 9 time 4 85 11.101 0.0296*

Successional stage 9 nitrogen 2 87 9.473 0.0388*

Richness of chewing insects Intercept 1 88 236.745 \0.0001

Successional stage 2 42 19.538 0.0295*

Time 2 87 23.670 0.0028*

Spider abundance 1 88 12.236 0.1657

Nitrogen content 1 88 45.564 0.0005*

Phenolic compounds 2 88 0.501 0.7895

Successional stage 9 nitrogen 4 85 28.194 0.001*

Abundance of sap-sucking insects Intercept 1 88 228.308 \0.0001

Successional stage 2 42 11.785 0.0005*

Time 2 87 15.133 0.0003*

Spider abundance 1 88 3.684 0.1784

Nitrogen content 1 88 2.945 0.2742

Phenolic compounds 1 88 1.378 0.3789

Successional stage 9 time 4 85 9.716 0.0001*

54 J. O. Silva et al.

123

References

Allen SE, Grimshaw HM, Parkinson JA, Quarmby C (1974) Chemical

analysis of ecological materials. Blackwell Scientific Publica-

tions, Oxford

Antunes FZ (1994) Caracterizacao climatica–caatinga do estado de

minas gerais. Info Agro 17:15–19

Barret MA, Stiling P (2007) Relationships among key deer, insect

herbivores and plant quality. Ecol Res 22:268–273

Basset Y, Aberlenc HP, Barrios H, Curletti G, Berenger JM, Vesco JP,

Causse P, Haug A, Hennion AS, Lesobre L, Marques F, O’Meara

R (2001) Stratification and diel activity of arthropods in a lowland

rainforest in Gabon biological. J Linn Soc 72:585–607

Basset Y, Novotny V, Miller SE, Kitching RL (2003) Arthropods of

tropical forests: spatio-temporal dynamics and resource use in

the canopy. Cambridge University Press, Cambridge

Boege K (2004) Induced responses in three tropical dry forest

plant species–direct and indirect effects on herbivory. Oikos

107:541–548

Boege K (2005) Herbivore attack in Casearia nitida influenced by

plant ontogenetic variation in foliage quality and plant architec-

ture. Oecologia 143:117–125

Borror DJ, Triplehorn CA, Johnson NF (2002) An introduction to the

study of insects. Saunders College Publishing, New York

Brown BJ, Ewel JJ (1987) Herbivory in complex and simple tropical

successional ecosystems. Ecology 68:108–116

Table 3 continued

Response variable Explanatory variable nunDF denDF F P

Richness of sap-sucking insects Intercept 1 88 183.837 \0.0001

Successional stage 2 42 0.956 0.5681

Time 2 87 8.670 0.003*

Spider abundance 1 88 10.490 0.0025*

Richness of sap-sucking insects Nitrogen content 1 88 0.907 0.5933

Phenolic compounds 1 88 0.006 0.9839

Successional stage 9 time 4 85 12.342 <0.0001*

Spider abundance Intercept 1 88 304.785 \0.0001

Successional stage 2 42 0.088 0.9558

Time 2 87 34.062 <0.0001*

Successional stage 9 time 4 85 12.205 0.0237*

Phenolic compounds Intercept 1 88 305.757 \0.0001

Successional stage 2 42 0.092 0.8305

Time 2 87 33.585 <0.0001*

Nitrogen content Intercept 1 88 968.741 \0.0001

Successional stage 2 42 22.065 <0.0001*

Time 2 87 67.926 <0.0001*

Successional stage 9 time 4 85 4.511 0.0738*

Leaf damage (%) Intercept 1 88 431.887 \0.0001

Successional stage 2 42 8.898 <0.0001*

Time 2 87 156.539 <0.0001*

Successional stage 9 time 4 85 11.06 <0.0001*

Herbivory increment (%) Intercept 1 88 383.596 \0.0001

Successional stage 2 87 25.677 <0.0001*

Time 2 87 31.169 <0.0001*

Spider abundance 1 88 3.223 0.1873

Nitrogen content 1 88 5.184 0.0334*

Phenolic compounds 1 88 9.809 0.0086*

Chewer abundance 1 88 0.172 0.7358

Successional stage 9 time 4 85 7.692 <0.0001*

Successional stage 9 nitrogen 2 87 6.391 0.02596*

Successional stage 9 phenolic 2 87 13.664 0.0038*

Also, the effects of plant quality (phenolic and nitrogen contents) and spider abundance on herbivore diversity and herbivory increment were

assessed. In these models, all variables were log-transformed to reach a normal distribution. Only the significant interactions between explanatory

variables are shown

* Explanatory variable retained in the minimum adequate model (P \ 0.05)

Leaf herbivory in a tropical dry forest 55

123

Bryant JP, Chapin FS III, Klein DR (1983) Carbon/nutrient balance of

boreal plants in relation to vertebrate herbivory. Oikos 40:

357–368

Campos RL, Vasconcelos HL, Ribeiro SP, Neves FS, Soares JP

(2006) Relationship between tree size and insect assemblages

associated with Anadenanthera macrocarpa. Ecography 29:

442–450

Cates RG, Orians GH (1975) Successional status and the palatability

of plants to generalized herbivores. Ecology 56:410–418

Close DC, McArthur C (2002) Rethinking the role of many plant

phenolics–protection from photodamage not herbivores? Oikos

199:166–172

Coley PD, Barone JA (1996) Herbivory and plant defenses in tropical

forests. Annu Rev Ecol Syst 27:305–335

Coley PD, Bryant JP, Chapin FS III (1985) Resource availability and

plant anti-herbivore defense. Science 230:895–899

Cornelissen TG, Fernandes GW (2001) Induced defences in the

neotropical tree Bauhinia brevipes (Vog.) to herbivory: effects of

damage-induced changes on leaf quality and insect attack. Tree

15:236–241

Crawley M (2002) Statistical computing: an introduction to data

analysis using S-Plus. Wiley, London

Davidson DW (1993) The effects of herbivory and granivory on

terrestrial plant succession. Oikos 68:23–35

Didham RK, Springate ND (2003) Determinants of temporal variation

in community structure. In: Basset Y, Novotny V, Miller S,

Kitching R (eds) Arthropods of tropical forests: spatio-temporal

dynamics and resource use in the canopy. Cambridge University

Press, Cambridge, pp 28–39

Dirzo R, Boege K (2008) Patterns of herbivory and defense in tropical

dry and rain forests. In: Carson W, Schnitzer SA (eds) Tropical

forest community ecology. Blackwell Science, West Sussex,

pp 63–78

Dirzo R, Domınguez CA (1995) Plant-herbivore interactions in

Mesoamerican tropical dry forest. In: Bullock SH, Mooney A,

Medina E (eds) Seasonally dry tropical forest. Cambridge

University Press, Cambridge, pp 304–309

Ernest KA (1989) Insect herbivory on a tropical understory tree:

effects of leaf age and habitat. Biotropica 21:194–199

Fernandes GW, Castro FMC, Faria ML, Marques ESA, Greco MKB

(2004) Effects of hygrothermal stress, plant richness, and

architecture on mining insect diversity. Biotropica 36:240–247

Filip V, Dirzo RJ, Maass M, Sarukhan J (1995) Within- and among-

year variation in the levels of herbivory on the foliage of

trees from a Mexican tropical deciduous forest. Biotropica 27:

78–86

Gentry AH (1992) Bignoniaceae–part II (Tribe Tecomeae). Flora

neotropica, monograph 25(II). New York Botanical Garden,

New York

Grose SO, Olmstead RG (2007) Taxonomic revisions in the

polyphyletic genus Tabebuia s.l. (Bignoniaceae). Syst Bot 32:

660–670

Gruner DS, Polhemus DA (2003) Arthropod assemblages across a

long chronosequence in the Hawaiian islands. In: Basset Y,

Novotny V, Miller S, Kitching R (eds) Arthropods of tropical

forests: Spatio-temporal dynamics and resource use in the

canopy. Cambridge University Press, Cambridge, pp 135–145

Hagerman AE (1987) Radial diffusion method for determining tannin

in plant extracts. J Chem Ecol 13:437–449

Herms DA, Mattson WJ (1992) The dilemma of plants: to grow or

defend. Q Rev Biol 67:283–335

IEF—Instituto Estadual de Florestas (2000) Parecer tecnico para a

criacao do Parque Estadual da Mata Seca. Relatorio tecnico,

Belo Horizonte

Janzen DH (1981) Patterns of herbivory in a tropical deciduous forest.

Biotropica 13:271–282

Janzen DH, Waterman PG (1984) A seasonal census of phenolics,

fibre and alkaloids in foliage of forest trees in Costa Rica: some

factors influencing their distribution and relation to host

selection by Sphingidae and Saturniidae. Biol J Linn Soc 21:

439–454

Kalacska M, Sanchez-Azofeifa GA, Calvo-Alvarado JC, Quesada M,

Rivard B, Janzen DH (2004) Species composition, similarity and

diversity in three successional stages of seasonally dry tropical

forest. For Ecol Manage 200:227–247

Leite LO, Borges MAZ, Lima CA, Goncalves RMM, Siqueira PR

(2008) Variacao espaco-temporal do uso de recursos pela

avifauna do Parque Estadual da Mata Seca. MG Biota 1:54–60

Lewinsohn TM, Novotny V, Basset Y (2005) Insects on plants:

diversity of herbivore assemblages revisited. Annu Rev Ecol

Syst 36:597–620

Ludwig D, Walter B, Holling CS (1997) Sustainability, stability, and

resilience. Conserv Ecol 1:1–27

Madeira BG, Espırito-Santo MM, DAngelo-Neto S, Nunes YRF,

Sanchez-Azofeifa GA, Fernandes GW, Quesada M (2009)

Changes in tree and liana communities along a successional

gradient in a tropical dry forest in south-eastern Brazil. Plant

Ecol 291:291–304

Mattson JMJ (1980) Herbivory in relation to plant nitrogen content.

Annu Rev Ecol Syst 11:119–161

Mendonca MS (2001) Galling insect diversity: the resource synchro-

nization hypothesis. Oikos 95:171–176

Moran CV, Southwood TRE (1982) The guild composition of

arthropod communities in trees. J Anim Ecol 51:289–306

Moreira PA, Fernandes GW, Collevatti RG (2009) Fragmentation and

spatial genetic structure in Tabebuia ochracea (Bignoniaceae), a

seasonally dry Neotropical tree. For Ecol Manage 258:2690–2695

Neves FS, Araujo LS, Fagundes M, Espırito-Santo MM, Fernandes

GW, Sanchez-Azofeifa GA, Quesada M (2010a) Canopy

herbivory and insect herbivore diversity in a dry forest-savanna

transition in Brazil. Biotropica 42:112–118

Neves FS, Braga RF, Espırito-Santo MM, Delabie JHC, Fernandes

GW, Sanchez-Azofeifa GA (2010b) Diversity of arboreal ants in

a Brazilian tropical dry forest: effects of seasonality and

successional stage. Sociobiology 56:177–194

Novotny V, Basset Y (1998) Seasonality of sap-sucking insects

(Auchenorrhyncha, Hemiptera) feeding on Ficus (Moraceae) in a

lowland rain forest in New Guinea. Oecologia 115:514–522

Novotny V, Basset Y, Kitching R (2003) Herbivore assemblages and

their food resources. In: Basset Y, Novotny V, Miller S, Kitching

R (eds) Arthropods of tropical forests: spatio-temporal dynamics

and resource use in the canopy. Cambridge University Press,

Cambridge, pp 40–53

Pezzini FF (2008) Fenologia e caracterısticas reprodutivas em

comunidades arboreas de tres estagios sucessionais em Floresta

Estacional Decidual do norte de Minas Gerais. Master disserta-

tion, Universidade Federal de Minas Gerais, Belo Horizonte

Pezzini FF, Brandao D, Ranieri BD, Espırito-Santo MM, Jacobi CM,

Fernandes GW (2008) Polinizacao, dispersao de sementes e

fenologia de especies arboreas no Parque Estadual da Mata Seca.

MG Biota 1:37–45

Poorter L, Plassche MV, Willems S, Boot RGA (2004) Leaf traits and

herbivory rates of tropical tree species differing in successional

status. Plant Biol 6:746–754

Price P (1997) Insect ecology. Wiley, New York

Quesada L, Sanchez-Azofeifa GA, Alvarez-Anorve M, Stoner KE,

Avila-Cabadilla L, Calvo-Alvarado J, Castillo J, Espırito-Santo

MM, Fagundes M, Fernandes GW, Gamonb J, Lopezaraiza-Mikel

M, Lawrence D, Morellato LPC, Powers JS, Neves FS, Rosas-

Guerrero V, Sayago R, Sanchez-Montoya G (2009) Succession

and management of tropical dry forests in the Americas: review

and new perspectives. For Ecol Manage 258:1014–1024

56 J. O. Silva et al.

123

R Development Core Team (2009) R: a language and environment for

statistical computing. R foundation for statistical computing,

http://www.r-project.org. Accessed 21 July 2009

Rasband WS (2006) ImageJ, US. National Institutes of Health,

Bethesda, Maryland, http://rsb.info.nih.gov/ij. Accessed 15 July

2009

Rhoades DF (1979) Evolution of plant chemical defense against

herbivores. In: Rosenthal GA, Janzen DH (eds) Herbivores: their

interaction with secondary plant metabolites. Academic Press,

London, pp 3–54

Ribeiro SP (1998) The role of herbivory in Tabebuia spp. life history

and evolution. Ph.D Dissertation, Imperial College at Silwood

Park, London

Ribeiro SP, Brown VK (1999) Insect herbivory within tree crowns of

Tabebuia aurea and T. ochracea (Bignoniaceae): contrasting the

Brazilian Cerrado with the wetland ‘Pantanal Matogrossense’.

Selbyana 120:159–170

Ribeiro SP, Brown V (2006) Prevalence of monodominant vigorous

tree populations in the tropics: herbivory pressure on Tabebuiaspecies in very different habitats. J Ecol 94:932–941

Ribeiro SP, Pimenta HR (1991) Padroes de abundancia e de

distribuicao temporal de herbıvoros de vida livre em Tabebuiaochracea (Bignoniaceae). Ann Soc Entomol Brasil 20:428–448

Ribeiro SP, Pimenta HP, Fernandes GW (1994) Herbivory by

chewing and sap-feeding insects on Tabebuia ochracea. Biotro-

pica 26:302–307

Sanders D, Nickel H, Grutzner T, Platner C (2008) Habitat structure

mediates top–down effects of spiders and ants on herbivores.

Basic Appl Ecol 9:152–160

Santos JC, Almeida-Cortez JS, Fernandes GW (2011) Diversity of

gall-inducing insects in the high altitude wetland forests in

Pernambuco, Northeastern Brazil. Braz J Biol 71:47–56

Siemann E, Haarstad J, Tilman D (1999) Dynamics of plant and

arthropod diversity during old field succession. Ecograph 22:

406–414

Silva JO, Jesus FM, Fagundes M, Fernandes GW (2009) Esclerofilia,

taninos e insetos herbıvoros associados a Copaifera langsdorffiiDesf. (Fabaceae: Caesalpinioideae) em area de transicao Cerra-

do-Caatinga no Brasil. Ecol Austral 19:197–206

Silva JO, Oliveira KN, Santos KJ, Espırito-Santo MM, Neves FS,

Faria ML (2010) Efeito da estrutura da paisagem e do genotipo

de Eucalyptus na abundancia e controle biologico de Glycaspisbrimblecombei Moore (Hemiptera: Psyllidae). Neotrop Entomol

39:91–96

Stanton N (1975) Herbivore pressure on two types of tropical forests.

Biotropica 7:8–11

Stiling P, Moon DC (2005) Quality or quantity: the direct and indirect

effects of host plants on herbivores and their natural enemies.

Oecologia 142:413–420

Strong DR, Lawton JH, Southwood TRE (1984) Insects on plants:

community patterns and mechanisms. Blackwell Scientific

Publication, London

Sullivan JJ (2000) How the sapling specialist shoot-borer, Cromarchastroudagnesia (Lepidoptera, Pyralidae, Chrysauginae), alters the

population dynamics of the Costa Rican tropical dry forest tree

Tabebuia ochracea (Bignoniaceae). Ph.D Dissertation, Univer-

sity of Pennsylvania, Philadelphia

Swain T, Hillis WE (1959) The phenolic constituents of Prunusdomestica I. The quantitative analysis of phenolic constituents.

J Sci Food Agric 10:63–68

Swanson ME, Franklin JF, Beschta RL, Crisafulli CM, DellaSala DA,

Hutto RL, Lindenmayer DB, Swanson FJ (2011) The forgotten

stage of forest succession: early-successional ecosystems on

forest sites. Front Ecol Environ 9:117–125

Tauber MJ, Tauber CA, Masaki S (1986) Seasonal adaptations of

insects. Oxford University Press, Oxford

Varanda EM, Pais MP (2006) Insect folivory in Didymopanaxvinosum (Apiaceae) in a vegetation mosaic of Brazilian Cerrado.

Braz J Biol 66:671–680

Wolda H (1988) Insect seasonality, why? Ann Rev Ecol Syst 19:1–18

Leaf herbivory in a tropical dry forest 57

123