J Appl Entomol. 2019;143:21–33. wileyonlinelibrary.com/journal/jen
|
21
© 2018 Blackwell Verlag GmbH
1 | INTRODUCTION
Understanding how agroecosystem components can change the at-
tack intensity of a pest species on its host plant is a key step in the
development of integrated pest management (IPM) programmes
(Pereira, Picanço, Bacci, Crespo, & Guedes, 2007; Semeão et al.,
2012). This may be even more important when considering a crop
system such as tomatoes, where pest insect control interventions
are intensively carried out with synthetic insecticide applications.
However, in some cases, the desired outcome was not achieved
due to control failures and pest resistance to insecticides (Campos,
Silva, Silva, Silva, & Siqueira, 2015; Campos et al., 2014; Gontijo
et al., 2013; Roditakis et al., 2015; Silva, Assis, Ribeiro, & Siqueira,
2016; Silva et al., 2011, 2015; Siqueira, Guedes, & Picanço, 2000).
Therefore, the management practices adopted aim at conserving
and increasing the role natural control agents play as agroecosys-
tem components, being able to help to reduce losses due to pest
attack. Such practices are environmentally safer and more sustain-
able since they have low or no negative economic impact on crop
production.
Received:31March2018
|
Revised:20August2018
|
Accepted:21August2018
DOI: 10.1111/jen.12567
ORIGINAL CONTRIBUTION
Seasonal variation in natural mortality factors of Tuta absoluta
(Lepidoptera: Gelechiidae) in open- field tomato cultivation
Leandro Bacci
1
| Ézio Marques da Silva
2
| Júlio Cláudio Martins
3
|
Marianne A. Soares
4
| Mateus Ribeiro de Campos
5
| Marcelo Coutinho Picanço
6
1
Departamento de Engenharia
Agronômica, Universidade Federal de
Sergipe, São Critovão, Brazil
2
Instituto de Ciências Agrárias, Universidade
Federal de Viçosa, Viçosa, Brazil
3
Instituto Federal de Educação, Ciência
e Tecnologia Baiano, Campus Teixeira de
Freitas, Teixeira de Freitas, Bahia, Brazil
4
Departamento de
Entomologia, Universidade Federal de
Lavras, Minas Gerais, Brazil
5
INRA (French National Institute
for Agricultural Research), CNRS,
UMR 1355-7254, Institut Sophia
Agrobiotech, Université Côte d’Azur, Sophia
Antipolis, France
6
Departamento de
Entomologia, Universidade Federal de
Viçosa, Viçosa, Brazil
Correspondence
Júlio Cláudio Martins, Instituto Federal de
Educação, Ciência e Tecnologia Baiano,
Campus Teixeira de Freitas, Teixeira de
Freitas, Bahia, Brazil.
Email: julioufv@gmail.com
Abstract
The seasonal variation in natural mortality of phytophagous insects is determined by
the relative importance of biotic and abiotic factors in agroecosystems. Knowledge
regarding these factors throughout the year represents a key concern for IPM pro-
grammes. Seasonal population fluctuations of tomato pinworm, Tuta absoluta, led to
an investigation of its natural mortality factors during the rainy season when the
population level is low and during the dry season when population peaks occur. The
aim of this study was to verify the seasonal variation in T. absoluta mortality factors
in tomato crops. Immature stages of T. absoluta were obtained from laboratory-
rearing in the laboratory. These were taken to the field and monitored over two
years. The mortality causes for each stage of insect development from egg to adult
were assessed daily. Multiple biotic and abiotic mortality factors affected the imma-
ture T. absoluta stages such as rainfall, physiological disturbances, diseases, parasi-
toids and predators. The key T. absoluta mortality factor during summer–spring was
predation. In addition, larvae predation correlated positively with temperature, wind
velocity, photoperiod and rainfall. Nevertheless, during winter–fall, the key mortality
factor was parasitism. Therefore, the critical stage for mortality was 3rd- and 4th-
instar larvae, being more vulnerable to natural control factors. Finally, the results
showed the importance of vertical and horizontal action on natural mortality
factors.
KEYWORDS
climatic elements, insect seasonality, parasitoid, physiological disturbance, predator, Tomato
borer
22
|
BACCI et Al.
Maintenance of pest population balance can be achieved
through the action of several natural mortality factors (Pereira
et al., 2007; Semeão et al., 2012). This kind of information can be
obtained using ecological life tables from pest studies to deter-
mine the critical stage and key mortality factors of a pest (Gonring,
Picanço, Guedes, & Silva, 2003; Pereira et al., 2007; Rosado et al.,
2014; Semeão et al., 2012; Silva et al., 2017). Interaction rele-
vance of leafminer in demographic parameters is affected be-
tween miners and host plants, such as miners’ insects and natural
enemies (vertical sources) or in intra- and inter- specific compe-
tition, including meteorological factors, such as wind, storms,
frosts, heavy rain, moisture and extreme temperatures (horizontal
sources; Auerbach, Connor, & Mopper, 1995). Seasonal variation
in agroecosystems, such as rainfall and dramatic maximum/mini-
mum temperature changes, can be the most important factors to
reduce insect abundance (Cornell & Hawkins, 1995). Among the
factors involved in natural mortality, natural enemies and climatic
variables stand out because they cause mortality during all life
cycle stages of insect pests and generally have a great impact on
the population dynamics of these organisms (Naranjo & Ellsworth,
2005). These two groups of mortality factors may present season-
ality throughout the crop cycle and during the different seasons
(Pereira et al., 2007; Semeão et al., 2012). Therefore, it is import-
ant to know the duration and magnitude of these factors at differ-
ent times, because this information is fundamental for population
dynamics studies and sustainable development of pest manage-
ment programmes.
The tomato pinworm, Tuta absoluta (Meyrick; Lepidoptera:
Gelechiidae), is the main tomato pest in South America (Gontijo
et al., 2013; Silva et al., 2011). Currently, this pest has been recog-
nized as a worldwide threat to tomato production, being one of the
major concerns as a quarantine pest for North American and Asian
countries, where the presence of this pest has not yet been con-
firmed (Biondi, Guedes, Wan, & Desneux, 2018; Campos, Biondi,
Adiga, Guedes, & Desneux, 2017; Desneux, Luna, Guillemaud, &
Urbaneja, 2011; Desneux et al., 2010). This is a result of the high
capacity of T. absoluta to cause economic damage to tomato crops
due to its larvae attacking various parts of the plant such as the
leaves, flowers, stems and especially fruits (Biondi et al., 2018;
Galdino et al., 2015; Tropea Garzia, Siscaro, Biondi, & Zappalà,
2012). The losses caused by T. absoluta to tomato crops, depend-
ing on the time of year, may reach 100% (Desneux et al., 2011;
Guedes & Picanço, 2012). This variability in attack is probably
due to the seasonality of its population density. Considering that
the harmfulness of T. absoluta depends on its abundance, under-
standing the mechanisms that determine its population regulation
is fundamental in the elaboration of integrated pest management
programmes.
Natural enemies and climatic conditions can overcome defi-
ciencies in the method normally used to control T. absoluta, that
is, synthetic insecticide applications, which are a critical IPM
component that can greatly impact population dynamics of pests.
Therefore, the objectives of this study were (a) to investigate the
impact of natural mortality factors on T. absoluta; (b) the seasonal-
ity of these mortality factors throughout the year in tomato crops;
(c) and to determine the relationship between these components
of T. absoluta mortality.
2 | MATERIALS AND METHODS
The study carried out in an experimental tomato crop area in
Viçosa, Minas Gerais, Brazil (20°48′45″S, 42°56′15″W, altitude
600 m). The experimental setup was 12 rows with 30 tomato
plants cv Santa Clara seedling in a spacing of 0.5 m between plants
and 1 m between rows in a total area of 180 m
2
. The experimen-
tal area used standard agronomic practices for tomato cultivation;
however, no other method to control insects and diseases was
used.
Seasonal and natural variabilities in T. absoluta mortality fac-
tors were evaluated using ecological life tables for eight different
periods over two years. The assessments were performed for four
seasons. The periods in each year were chosen to represent dif-
ferent seasons (fall, winter, spring and summer) for the Southern
Hemisphere.
2.1 | Rearing of Tuta absoluta
The leaves of tomato plants with T. absoluta were collected from
commercial crops and taken to the laboratory for study (Silva
et al., 2011). A T. absoluta colony was established in the laboratory
using a system of four wood cages (40 × 40 × 40 cm), covered with
nylon mesh. The laboratory- rearing system consisted of cages for
oviposition, 1st- and 2nd- instar larvae, 3rd- and 4th- instar lar-
vae, pupae and adults. The T. absoluta colony in the laboratory
was placed under conditions of 25 ± 2°C, relative humidity (RH)
75 ± 5% and photoperiod of 12:12 (L: D) hr. Larvae were fed with
cv. Santa Clara tomato leaves from the greenhouse without insec-
ticide application.
2.2 | Mating establishment
Tuta absoluta stages from laboratory- rearing were taken to field and
their mortality monitored throughout the seasons. Ten experimental
plots were used for each instar, with each plot having a tomato plant
at the reproductive stage chosen at random from the crop. Egg and
1st- instar mortality was evaluated on the apical canopy. Second- ,
third- and fourth- instar T. absoluta mortality was evaluated on the
median canopy. These canopy positions were selected for evaluation
of mortality, because they are the preferred plant sites for female
oviposition (Torres, Faria, Evangelista, & Pratissoli, 2001). Therefore,
ten leaves from the apical canopy (plots with eggs and 1st- instar lar-
vae) and ten from the median canopy with 2nd- , 3rd- and 4th- instar
larvae were carefully inspected to the presence of T. absoluta eggs,
larvae and pupae using 10× magnifier lens. If any egg, larva or pupa
were found, they were removed with the aid of a soft paint brush
|
23
BACCI et Al.
as well as any dust or eggs and larvae of other species that were
present on the leaves.
During initial mating establishment, adult couples were kept in
the laboratory. After mating, they were taken to field to obtain eggs.
Thirty T. absoluta adults were placed in each plot, closed using nylon
mesh, and were left for 24 hours for female oviposition. Eggs were
counted using a 10× magnifying manual lens, and their positions on
the tomato leaf were recorded using schematic drawing. In each plot,
were left 150 eggs, totalling 1,500 eggs per season.
Two hundred and fifty larvae from each instar from the T. ab-
soluta colony were used for larval stage establishment. From each
instar, 25 larvae/plot were used. They were transferred using a fine-
tipped brush. The different instars were identified through the de-
gree of cuticle sclerotization, body size and larval cephalic capsule
(Imenes, Fernandes, Campos, & Takematsu, 1990). At each larval
stage, the insects were collected from the field at the beginning of
the change to the next instar. An exception was made for the 4th-
instar larvae, which were removed from the field when they pre-
sented a pinkish appearance, which is indicative of the change from
larval stage to pupae (Imenes et al., 1990). This removal of insects
occurred only at the end of the larval 4th instar, before pupal stage,
and then, the larvae remained exposure during all the 4th- instar lar-
val stage. This procedure was adopted, because the 4th- instar larvae
generally migrate to the soil before they start to pupate.
A total the 300 individuals were selected from the field and
grouped into 30 individual/plot to evaluate pupae mortality. The
pupae were placed into plastic trays (5 × 25 × 30 cm) with a 1 cm
layer of sand, which were placed underneath the tomato leaves in
the pupation cages stored in the laboratory for 24 hr. After this pe-
riod, the sand was sifted, and the pupae removed for use in the field.
This was done due to T. absoluta pupae often being found in the soil
or in the dry leaflets or stems of tomato plants (Imenes et al., 1990).
The pupae from each plot were conditioned in plastic pots (10 cm
in diameter and 7 cm of height), perforated on the side and at the
base to allow water drainage from the occasional rains that occurred
during the experimental period. The pupae were placed inside plas-
tic pots with a thin layer of sand at the bottom of approximately
3 mm height. Then, the pots were placed at soil level next to the
base of the tomato plant. In order to standardize the time of pupae
removal from field, a pilot test was performed in the laboratory to
measure the time taken for adult emergence.
2.3 | Assessment of mortality factors
Tuta absoluta mortality factors were monitored in the field from the
beginning of the egg stage to the end of the pupal stage (Pereira
et al., 2007; Semeão et al., 2012). Egg stage data were recorded, that
is, position, dates and disappearance cause. The rainfall action on
mortality of eggs was determined by counting the number of eggs
before and after each rainfall event. It was considered a predator
attack when eggs were not found in their previously registered posi-
tions and there had been no rain between two consecutive assess-
ments. This assumption is plausible since the T. absoluta eggs remain
strongly adhered to the leaves making it difficult for the wind to
dislodge them. Damaged eggs, which only presented chorion, were
considered dead due to predation. The presence of predators on
each leaf was recorded, and representative specimens were col-
lected and placed in 70% ethanol for subsequent identification. The
eggs which did not hatch were placed in the glass tube (10 cm length
and 2 cm of diameter). The tubes were covered with perforated PVC
plastic and stored in the laboratory under conditions of temperature
at 25 ± 0.5°C, RH 75 ± 5% and photoperiod 12:12 (L: D) hr for sub-
sequent assessment of parasitoid emergence. Incubated eggs that
did not present parasitoid emergence were considered dead due to
inviability.
First- instar larvae mortality was evaluated in two stages: before
and after they penetrated the leaves. The time for leaf penetration
was 20–45 min (Imenes et al., 1990). Before leaf penetration, 1st-
instar larvae that disappeared following rainfall were considered
dead due to this factor, whereas in the absence of rainfall, they were
considered dead due to predation. After leaf penetration (formation
of the mines), larvae that disappeared or mines that were torn were
considered dead due to predation. At the end of the 1st- instar stage,
all the leaves evaluated at this stage were collected and analysed
on a stereoscopic microscope (40× magnification). First- instar larvae
that did not show signs of incomplete moulting (adhered to their ex-
uvie) and disease symptoms were considered dead due to unknown
factors (Pereira et al., 2007).
Second- , third- , and fourth- instar T. absoluta larvae that were
found during mortality assessment dehydrated inside the mines,
with ectoparasitoid, ecdysis disturbance or with disease symptoms
were considered dead due to predatory bugs. However, T. absoluta
larvae with disease symptoms had a mortality factor classified ac-
cording to the symptoms caused by fungi or bacteria (Gouli, Gouli,
& Marcelino, 2011). Larval mortality due to predation by predatory
wasps was determined by the direct observation of this action on
the T. absoluta mines which when attacked showed torn or removed
surfaces. The mortality caused by other predators was determined
by the observation of the direct action of their organisms on the lar-
vae in the field. The same was the case with ectoparasitism, through
evidence of larval paralysis and the observation of ectoparasitoids
on T. absoluta larvae. Additionally, all the natural enemies found on
the leaves were collected and their representative specimens were
kept in 70% ethanol for complete identification. Like the egg stage,
the effect of rain on larvae was determined by checking leaves im-
mediately before and after rainfall. The mortality from rainfall was
identified in the field by verifying dead larvae in mines flooded
with water after the occurrence of this factor or by observation of
drowning dead larvae trapped between leaf epidermis as the water
evaporated.
The mortality was only attributed to predation when predators
were visually observed feeding on egg, larval and pupal stages.
Moreover, a parallel experiment was also carried out on the same
fields to monitor the activity of the predators and identify the most
frequent preying on T. absoluta stages according with the method-
ology used by Asiimwe et al. (2007). During this time, 24- hr checks
24
|
BACCI et Al.
done at 3- hr intervals over a 1- month period at all the evaluated sea-
sons were carried out.
To determine the level of parasitism, the 1st- , 2nd- , 3rd- and 4th-
instar larvae alive at the end of each stage were transferred to new
tomato leaves and conditioned in 500- ml plastic pots. The plastic
pots with a 2- cm
2
hole in the top were covered with nylon mesh to
allow for ventilation. The larvae were kept in the laboratory until the
adult stage to verify parasitoid emergence two times a week on the
same assessment date. New tomato leaflets were added to feed lar-
vae. Emerged parasitoids were counted and kept in 70% ethanol for
subsequent identification. Emerged T. absoluta adults were counted
and removed from the plastic pots. Tomato leaflets were kept in the
plastic pots until the T. absoluta adults or parasitoids were no longer
observed.
During the pupal stage, natural mortality was determined by di-
rectly counting the number of pupae remaining in each plastic pot at
the end of the period in the field. The number of pupae dead from
predation was determined by the difference between the total initial
pupae and pupae remaining in the plastic pots. The mortality was
only attributed to predation when predators were visually observed
feeding on pupal stage. Remaining pupae were placed in a glass tube
similar to the procedure adopted with un- hatched eggs. The pupae
were kept in the glass tube for 30 days until T. absoluta adult or par-
asitoid emergence. After this, the pupae still in the glass tubes were
classified as dead due to physiological disturbances (malformation)
or rain. Pupae mortality due to rainfall was estimated in the labora-
tory where the experiment simulating the addition of rainfall to the
plastic pupae pots under field conditions was realized. The daily rain-
fall volumes that occurred during the period that pupae remained in
the field were recorded using a gauge installed in the crop. Rainfall
action was attributed through the difference of pupae mortality
between plastic pots that had added water and those that did not.
Pupae that did not emerge T. absoluta adults were classified as dead
due to physiological disturbances.
Tuta absoluta parasitoids were identified by Dra. Angélica Maria
Penteado Martins Dias of the Universidade Federal de São Carlos,
by Dr. Paulo Sérgio Fiuza Ferreira of the Universidade Federal de
Viçosa and the Coleoptera by Dr. Ayr de Moura Bello. Others natural
enemy specimens were collected, assembled and identified at the
Entomology Museum of the Universidade Federal de Viçosa.
2.4 | Statistical analysis
For the parameters analysed, standard methods for building
and ecological analysis of life tables were used (Southwood &
Henderson, 2000). Significant differences in total mortality caused
by natural mortality factors throughout the four seasons and the
years were analysed using two- way ANOVA (PROC ANOVA) in SAS
v.9.2 Software (SAS Institute, 2009). If a significance was detected,
post hoc analyses were conducted using confidence interval at 95%
(PROC UNIVARIATE SAS Institute, 2009). Since mortality levels can
affect population growth, net reproductive rate (R
0
) in each season
was estimated and compared to a reference net reproductive rate
(R
0
= 1) using a t test (TTEST Procedure) in the SAS v.9.2. The net
reproductive rate data were calculated predicting how the T. abso-
luta population would behave in the light of the seasonality of the
mortality factors. The net reproductive rate was estimate by division
of egg number expected in the next generation (number of surviving
adults in the current mating × sexual rate × fecundity) by the initial
number of eggs in the current mating. The sex ratio (i.e., number of
female/number of female + number of male) was considered to be
0.5973 with fecundity of 183 eggs/female (Mihsfeldt & Parra, 1999).
This information was used because it allows for standardization with
respect to sex and fecundity rates for all populations in the different
seasons of the year. In this way, it was possible to demonstrate the
impact of mortality on population development, without being af-
fected by the reproductive biology of the insect.
The K value was determined to identify the difference between
the mortalities by stage and mortality factors within each stage. The
K value was calculated using the formula [K = log(100qx)], where
100qx is the apparent mortality rate (Southwood & Henderson,
2000). The significant difference between mortalities by stage and
mortality factors within each stage was analysed using one- way
FIGURE1 Seasonal analysis of Tuta absoluta life history of two
parameters. (a) Differences in total mortality between seasons
were compared using ANOVA and confidence interval at 95%.
(b) Difference in net reproductive rate (R
0
) between seasons. For
each season, the R
0
was compared to an equilibrium value (R
0
= 1)
using the t test. In Figure (a) are shown the mean values and 95%
confidence intervals. In Figure (b) are shown the mean values and
standard errors. Asterisks indicate that R
0
for those seasons is
different from the equilibrium value
|
25
BACCI et Al.
ANOVA (PROC ANOVA) and, if a significance was detected, post hoc
analyses were conducted using confidence interval at 95% (PROC
UNIVARIATE SAS Institute, 2009).
For the T. absoluta critical stage and key mortality factor, a sim-
ple linear regression analysis of the partial K values of each stage or
each mortality factor as a function of the total K values (K = Σk) for
p < 0.05 was performed for each period. The critical stage or key
mortality factor was considered the one with the regression curve
showing a higher slope coefficient at p < 0.05 (Pereira et al., 2007;
Semeão et al., 2012). The difference between slopes was verified by
the confidence interval at 95% probability.
The seasonality of mortality factors potentially associated with
T. absoluta environmental factors such as wind velocity, rainfall, air
temperature and photoperiod were analysed using principal compo-
nent analysis (PCA) and other natural control agents using the pro-
gram Canoco 4.5 system (ter Braak & Smilauer, 2002). The biplot
ordering gradient was generated using CanoDraw 3.0. The response
gradient was represented by the vectors with the origin at the cen-
tre point of the two axes of the sorting diagram. The vector length
is proportional to the significance of the variable. Vectors with the
same direction and sorting represent a variable with positive cor-
relation, while vectors with the same direction opposite represent a
negative correlation. Variables were not correlated when the angle
between the vectors is 90°.
3 | RESULTS
There were significant differences in the total mortality caused by the
natural factors, considering different seasons of years (F
3;72
= 30.20,
p < 0.001), between the years (F
1;72
= 16.74, P < 0.001) and in the in-
teraction between seasons and year (F
3;72
= 12.19, p < 0.001).
High T. absoluta mortality rates were observed in the summer
and spring and low mortality was observed in the winter and fall
during two years as was observed by confidence interval at 95%
(Figure 1a). The T. absoluta net reproductive rate (R
0
) was signifi-
cantly higher than 1 in the winter in the first year (t test = 4.4, df = 36,
p = 0.001). The R
0
of T. absoluta was significantly <1 in the summer
FIGURE2 Relative proportion of
mortality caused by different mortality
factors and represented by K- value in
each season. The x- axis represents the
different stages and development stage of
Tuta absoluta