CC BY
42Russian Journal of Nematology, 2014, 22 (2), 89-100
Spatial patterns of Tuta absoluta and heterorhabditid nematodes
Mahfouz M.M. Abd-Elgawad
Phytopathology Department, National Research Center, EI-Tahrir St., Dokki 12622, Giza, Egypt
e-mail: mahfouzian2000@yahoo.com
Accepted for publication 31 May 2014
Summary. The tomato leafminer, Tuta absoluta, is a serious pest that can spread rapidly and damage tomato production substantially. Understanding the spatial distribution of the pest population and its natural enemies under specific conditions will aid in the formation and application of integrated control strategies against the pest. The dispersion parameter of T. absoluta larval population derived from Taylor's Power Law was studied at plant, leaf and leaflet scales in three tomato fields in Egypt. Values of b, the distribution parameter of the Power Law, increased from plant to leaflet, indicating a more aggregated distribution as the size scale diminished. Data transformations were calculated from parameters of the Power Law fitted to the data obtained across the three fields and compared with log and square root transformations. Using baiting technique with Gallería mellonella, entomopathogenic nematodes (EPN) were not detected from the three fields but were found in 29 out of 30 soil samples obtained from a nearby mango orchard. Mortality of T. absoluta larvae caused by Heterorhabdítís species isolated from the orchard ranged from 90.0 to 96.7% suggesting that its inundative release for pest control may be feasible. Horizontal spatial patterns of the nematodes were aggregated within the orchard, with individual soil samples being much more likely to contain sufficient nematodes to infect multiple G. mellonella larvae than was predicted based on the total number of positive soil samples. The distribution of EPN-infected G. mellonella larvae within a series of extractions was random in 27 out of the 29 samples; the larvae were aggregated in the two remaining positive samples. The possibility of finding infected G. mellonella larvae in the positive samples was not more than that expected by chance. Key words: distribution, entomopathogens, Gallería mellonella, insects, mango, tomato.
The tomato leafminer, Tuta absoluta (Meyrick) (Lepidoptera: Gelechiidae), is a very injurious pest affecting tomato plants from seedlings to mature stage, both in glasshouse and open field cultivations (Zappala et al2012). Now, it is considered as a serious threat to tomato production worldwide (e.g., Desneux et al., 2010, 2011). The insect has a high reproductive potential and a short life cycle (Pereyra & Sanchez, 2006). The insect larvae bore into and develop inside the plant, continuously searching for new feeding locations. The larvae usually mine in the leaves producing large galleries, penetrate into the stems causing tunnels leading to fragility of the stem and collapse of the plant, and burrow into the fruits causing a substantial losses of tomato production (Ibrahim & Hasan, 2012). The larval stage (caterpillars) does not enter diapause as long as food is available. There are four larval instars. In between moulting, caterpillars can be found temporarily outside the leaf mines or fruits. The larval period is most damaging period, which
usually lasts 12-15 days. Tuta absoluta infests other solanaceous crops (eggplant, sweet pepper, potato and tobacco) as well as spontaneous plants, such as the black nightshade, Solanum nigrum L.; occasional damage has been reported on green bean (Desneux et al., 2010). Moreover, Abolmaaty et al. (2010) predicted peaks in T. absoluta annual generation numbers under current and expected climate change in Egypt based on future degree days.
Control strategies for T. absoluta consist mainly of early detection through sex pheromone traps, the use of insecticides and cultural methods (Batalla-Carrera et al., 2010). Biocontrol of the pest by predators, parasitoids, entomopathogens, resistant plants, and botanical insecticides has met with variable success (Miranda et al., 1998; Marchiori et al., 2004; Desneux et al., 2010; Jacobson & Martin, 2011). Batalla-Carrera et al. (2010) stressed that effective biocontrol agents against T. absoluta are urgently needed to replace broad spectrum
insecticides to avoid hazards to humans and development of resistance in the pest (Siqueira et al, 2000).
Entomopathogenic nematodes (EPN) are lethal to a broad range of economically important insect pests (Gaugler & Kaya, 1990; Grewal et al., 2005). The only function of a nematode-infective juvenile (IJ) is to seek out and infect insect hosts. After entering an insect, IJ release their associated mutualistic bacterium. The nematodes and bacteria feed on the liquefying host, and reproduce for several generations inside the cadaver. When food resources in the host become scarce, the adult EPN produce new IJ adapted to withstand the outside environment and so they leave the host cadaver in search of new hosts. The nematodes Heterorhabditis bacteriophora (Poinar), Steinernema carpocapsae (Weiser) and S. feltiae (Filipjev) showed high efficacy against T. absoluta larvae, but limited success against pupae (Batalla-Carrera et al., 2010; Jacobson & Martin, 2011).
The aggregated distribution of nematodes in soil usually confounds parametric statistical analyses necessary for their study and developing management strategy (Goodell & Ferris, 1980; Abd-Elgawad, 1992; Salama & Abd-Elgawad, 2010). Approaches based on estimate of nematode dispersion indices have been used to solve these problems (Barker & Campbell, 1981). Taylor's power law (Taylor, 1961) provides such an index. It states that: the variance (S2) of a population is proportional to a fractional power (b) of the arithmetic mean (X): S2 = aX or log S2 = log a + b log X (i), where a and b are population parameters, a depends chiefly on the sample size and b is an index of nematode dispersion (McSorley et al., 1985). The law is useful in determining transformations (Taylor, 1970) and developing nematode sampling plans (Ferris, 1984).
In this paper, the spatial distribution patterns of T. absoluta larvae, as the most EPN-susceptible and tomato-damaging insect stage, at three separate tomato fields were studied. The patterns may serve as a basis for management-decision making (Bechinski & Pedigo, 1981; Faleiro et al, 2002) and as important information on ecological characteristics of insect and EPN species (Taylor, 1984; Salama & Abd-Elgawad, 2010). On the other hand, most surveys of natural populations have merely assessed the occurrence of EPN in single soil samples at different Egyptian sites and provide no information on the nematode distribution within sites (e.g., Abd-Elgawad & Nguyen, 2007; Abd-Elgawad et al, 2009; Abu-Shady et al, 2011). Hence, EPN populations in the three tomato fields
were examined to know the possibility of their use in augmentation or conservation biological control in Egypt. Also, EPN in a nearby area were surveyed to assess their spatial patterns as a solid basis of information for effective use in biocontrol programmes (Ferris, 1985; Cabanillas & Raulston, 1994; Stuart & Gaugler, 1994; Salama & Abd-Elgawad, 2010).
MATERIALS AND METHODS
The study area, the Sugar Beet village, is located near the Alexandria-Cairo desert road running from the southern edge of the Maryut company farm south towards El Nasr canal, 55 to 66 km from Alexandria, Egypt. It represents one of the relatively newly reclaimed areas in the west of the Nile delta. Its climate is semi arid with an annual rainfall varying from 136 mm at the north to 35 mm at the south. No rain falls between June and September and is sporadic during November to February, the main rainy season, when the rainfall is localised and unpredictable (Erian & Yacoub, 2000). The mean monthly maximum temperature rises to 32.2°C at west Nubariya. The mean monthly minimum temperature drops to 7.5°C.
Insect and EPN sampling and laboratory bioassay. Following cultivation of Egyptian clover (Trifolium alexandrinum L.), the sandy soil (sand = 94.7%, silt = 3.2%, clay = 2.1%, pH = 7.88, organic matter = 1.14%) at Sugar beet village was ploughed and prepared for planting seeds of tomato, Solanum lycopersicum L. cvs Alisa, Hadeer and Hynez, in three separate fields between 3 and 20 March twice in 2011 and 2012. Plants grew on both sides of I m-wide raised beds, and the distance between the plants at each side of a bed was 40-50 cm where seeds were planted, covered and then irrigation took place between beds so that water could spread across to seed level; all agricultural practices were carried out as recommended (Mohamed & Dosuki, 2009). Since Tuta absoluta sex pheromone trap is highly attractive to male moths, the pheromone based control system was used according to Ibrahim & Hasan (2012) only in fields with cvs Alisa and Hynez. Pheromone traps were placed on the tops of 25 cm wood sticks and randomly scattered, just above the plant. Also, a light trap consisted of bright bulb suspended above blue plastic pan (25 cm diam., 8.5 cm height) filled with 0.3% detergent was randomly scattered in the field with cv. Alisa (six traps/Feddan). Tomato cv. Hadeer served as non-treated control. Thus, the experiment was based on a priori reasons: reported efficacy of pheromone and light traps (cv. Alisa), light traps only (cv. Hynez)
and untreated control (cv. Hadeer) for predicting the observed ranks of treatment populations. About 300 m2 (in 2011) and 400 m2 (in 2012) of each tomato field was delimited for sampling tomato leafminer larvae according to Sanchez et al. (2009) at three different scales: plant, leaf, and leaflet. Only the upper third of the plant was sampled as representative of the distribution of the pest because T. absoluta concentrates mainly in the apical parts of the tomato plants (Torres et al, 2001). At each field, three plants at three successive beds were randomly sampled from each 100 m2 of the delimited area; 300 and 400 m2 of each field were sampled in years 2011 and 2012, respectively. The apical parts of 9 (in 2011) and 12 (in 2012) tomato plants, each plant having seven consecutive expanded leaves, were harvested at random at each field/cultivar. To examine the spatial distribution of T. absoluta larvae at the three scales, each apical part of the plant was sorted in all component leaves and each leaf in its component leaflets; seven leaflets were counted per leaf. The leaves were examined under a binocular microscope to count T. absoluta larvae at the three spatial scales (Sanchez et al., 2009). Insect sampling took place on July 5 (in 2011) and 11 (in 2012) during the tomato harvest season. For EPN sampling, soil samples were simultaneously collected from the rhizosphere of the same insect-sampled plants with a shovel to a depth of ca 25 cm beneath the plant shoots. Three subsamples, from three of the insect-sampled plants in 100 m2 plot, with a total volume of approx. 1500 cm3 were collected and mixed thoroughly to form a composite soil sample; three (in 2011) and four (in 2012) composited samples were collected from each of the three tomato fields.
On July 5, 2012, soil samples were also collected from 11-year-old mango, Mangifera indica L., trees planted at 2.5 m between rows and 5 m between trees of a row in a geographical North/South row direction. The mango orchard, located near to the Sugar Beet village, had sandy soil (sand = 94.8%, silt = 3.5%, clay = 1.7%, pH = 7.83, organic matter = 1.29%). It apparently lacked a tree management programme and seemed as mango semi-jungle. Plastic tube openings for dropping under tree canopies were spaced 70 cm with two parallel longitudinal tubes per a tree row; a tube at each side of the tree stem. The orchard was irrigated at weekly intervals and sampling time corresponded to 3 days after irrigation time. Samples were taken once from 30 trees, ten successive trees in each of three consecutive rows, with a shovel to a depth of ca 25 cm. Under each tree canopy, three subsamples, within a distance of 40-60 cm apart, with a total
volume of approx. 1500 cm3 were randomly collected and mixed to form a composite sample representing the tree canopy. Each sample was mixed thoroughly and a 500-cm3 aliquot from the soil sample was assayed for EPN using the Galleria-bait method (Bedding & Akhurst, 1975) in multiple cycles according to Stuart & Gaugler (1994) with modifications; i.e., samples were placed in assay chambers (plastic cups, 10 cm diam., 7 cm height). Each plastic cup was examined twice weekly for the first 2 weeks and once weekly thereafter. Infected juveniles (IJ) emerging from collected insect cadavers in the traps (White, 1927) were used to reinfect additional Galleria to confirm pathogenicity and to fulfill Koch's postulates (Pelczar & Reid, 1972).
EPN identification and preparation.
Taxonomic identification was based not only on characteristics of cadavers, IJ and males (Poinar, 1990; Stock & Hunt, 2005), but also on morphological and molecular identification of species previously isolated from the same area (Abd-Elgawad et al., 2013). Nonetheless, the isolates obtained in the present investigation are receiving further study, and the current analysis remains relevant whether considered at the generic or the species level.
Larval susceptibility in Petri dish assay. Three EPN species/isolates extracted from the mango soil, i.e., H. bacteriophora, Heterorhabditis sp. Mgo2, and Heterorhabditis sp. Mgo3, were separately processed through G. mellonella last-instar larvae several times to obtain a pure culture of the nematodes. IJ in White traps were harvested then counted (Woodring & Kaya, 1988) and suspended in distilled water within flasks at 12-15°C for 2 days before testing against T. absoluta late (third to fourth) larvae instars collected from the tomato fields according to Batalla-Carrera et al. (2010) with some modifications; i.e., insects were singly exposed to only one nematode dose (50 IJ cm-2 or 981 IJ per 5-cm-diameter Petri dish). Insect mortality was checked at 72 h and presence of nematodes inside dead insects was observed to confirm nematode infection. There were 10 replicates/treatment and the experiment was repeated twice.
Statistical analyses. Counts of T. absoluta larvae were subjected to ANOVA using a completely randomised design and their arithmetic means were compared using Duncan's New Multiple Range Test. Means of dead insects by the three nematode species/isolates were also subjected to similar analysis. Mean and variance of T. absoluta larvae counts were calculated from each seven leaflets
Table 1. Mean (±SE) and ranges of densities of T. absoluta larvae at three spatial scales in three Egyptian fields each
planted to a tomato cultivar*.
Tomato Treatment Plant Leaf Leaflet
cultivar Mean (SE) Min. Max. Mean (SE) Min. Max. Mean (SE) Min. Max.
Alisa Phermone and Water traps 18.14b (4.42) 0 67 2.59b (0.36) 0 28 0.37b(0.04) 0 10
Hadeer Non 35.86a (4.10) 0 75 5.12a (0.45) 0 39 0.73a (0.05) 0 14
Hynez Phermone 22.67b (4.11) 0 65 3.24b (0.39) 0 31 0.47b(0.04) 0 14
* Means in each column sharing a common letter are not significantly (P < 0.05) different according to Duncan's New Multiple Range Test.
1 - m , -s V n * > + «■ + tit, ' * + + + + ♦i+* * ■ * -í ' * + ■ft
4_ 1 1 1 1 * J ° -2.5 -2 -1.5 + -1 t * ! " - _--■ + + + 12 - -2 ■ Leg of ins E-ct n 0.? 1 15 2 4 y= 1.6542X +0.7614 + FT =0.7248 iE an
Fig. 1. Relationship between the logarithms of the mean (X) and variance (Y) of Tuta absoluta larvae counts on leaflets of tomato plants across three cultivars (n = 229).
of the same plant leaf for the leaflet scale; from each seven leaves of the same plant for the leaf scale; and from 3 plants at 100 m2 area for the plant scale in each field and pooled for the two years together. These means and variances of larval counts were natural-log transformed. Mean and variance of EPN-infected Gallería numbers from the bait method on different observation dates for the positive soil samples were also transformed. Log variance was regressed against log mean to determine coefficients of Taylor's Power Law for each distribution scale of single tomato cultivars, and also for all cultivars pooled together, as well as for distribution of EPN-infected Gallería from positive soil samples (Taylor, 1961). For example, log variance was plotted against log mean of each three plants sampled from each 100 m2 plot in a tomato field for
each plant scale (7 plots/field/2 years for plant scale et seq). Data of these infected Gallería were analysed using the chi-squared, variance-to-mean ratio for agreement with a Poisson series (small samples, n < 30) according to Elliott (1971). The index of dispersion (I) or variance-to-mean ratio was calculated: I = S2/X', where S2 = sample variance and X' = sample arithmetic mean. The significance of departures from unity was assessed by reference to tables of chi-square values (Pearson & Hartley, 1966).
The chi-squared value was calculated where a scale was not fit to the Power Law. A transformation procedure to normalize data in order to perform parametric statistics was determined (Taylor, 1970) as follows: Y = Xci"05b) (ii) and compared.
RESULTS
Insect sampling and spatial distributions. Data of T. absoluta larvae for each scale of a tomato cultivar did not differ (P < 0.05) between the two years; and therefore were pooled together. A total of 3,087 leaflets from 441 leaves of 63 tomato plants and 21 soil samples from the three tomato fields surveyed in 2011 and 2012 as well as 30 soil samples from mango sites in 2012 were examined. Overall mean densities (± standard error) for T. absoluta at the three spatial scales were 25.56 (±2.58) leafminer larvae/plant; 3.65 (±0.24)
larvae/leaf; and 0.52 (±0.03) larvae/leaflet. Densities of T. absoluta larvae in fields planted with cvs Alisa or Hynez were lower (P < 0.05) than on cv. Hadeer at the three scales when calculations were performed on the untransformed insect counts (Table 1). Out of 21 plants, 147 leaves and 1029 leaflets examined, 20, 96, and 284 for cv. Hadeer, 15, 58, and 134 for cv. Alisa, and 20, 75, and 232 for cv. Hynez, respectively, were infected by T. absoluta.
Taylor's Power Law was fitted to seven of the nine spatial scales related to the three tomato cultivars (Table 2). Generally, slope values of regression lines lie within the range of values which would indicate
6 •u o n] 3 ■ ■C 5 -e 2' J ° * --- + 4 * 4 +
-3 -2.5 -2 -1.5 < - t -0.5 I * -2 ■ -O ai 1 1Ï 2 2i 3 Ï-1J154Ï + a.T631 FÎ-a.7439 Log of insect mean
Fig. 2. Relationship between the logarithms of the mean (X) and variance (Y) of Tuta absoluta larvae counts on leaves of tomato across three fields (n = 55).
Table 2. Taylor's Power Law regression equations for data of Tuta absoluta larvae from three tomato cultivars/fields.
Tomato cultivar Distribution scale n R2 Power law parameters Effective range
a b x" S2
leaflet 58 0.846** 1.08 1.73 0.14-4.0 0.14-20.33
Alisa Leaf 15 0.834** 1.11 1.24 0.29-9.57 0.57-82.60
Plant 7 0.796** 2.01 1.18 3.33-43.99 25.0-586.4
leaflet 96 0.736** 0.78 1.92 0.14-5.57 0.14-27.48
Hadeer Leaf 20 0.294* 0.70 1.22 0.71-10.71 1.57-221.63
Plant 7 0.022 ns 1.55 1.09 27.33-46.67 56.51-1410.52
leaflet 75 0.664** 0.42 1.39 0.14-4.43 0.14-37.94
Hynez Leaf 20 0.921** 0.70 1.50 0.14-9.29 0.14-142.91
Plant 7 0.459 ns 1.50 0.96 2.0-57.0 4.0-390.33
n = number of non-zero points, R2 = coefficient of determination for fit to equation: log S2 = log a + b log x, where x- = mean, S2= variance. Effective range = range of x'and S2 data. * and ** = significant at P < 0.05 and P < 0.01, respectively; ns = no significant regression.
Fig. 3. Relationship between the logarithms of the mean (X) and variance (Y) of Tuta absoluta larvae counts on plants of tomato across three fields (n = 21).
Table 3. Comparison of F values from Analysis of Variance Tables from T. absoluta larvae counts and from transformed counts on three scales of tomato cvs Alisa, Hadeer, and Hynez.
Sample scale F value
Df+ Natural log x Square root Taylor law* Untransformed
Leaflet 2&3084 27.61 22.75 26.01 17.16
Leaf 2&144 3.27 3.17 3.15 2.73
Plant 2&18 3.00 3.35 3.35 3.04
Samples were collected from three tomato fields/cultivars, representing three different treatments. +Degrees of freedom for treatments (the three tomato cultivars) and error of a completely randomized design.
Transformation calculated from y = x(1 - 0 5b) where x is the population count and b is the aggregation index from Taylor's Power Law (Table 2; Taylor, 1970).
from equation (ii) that log (slope = 2.0) or square root (slope = 1.0) data transformations are most appropriate to equalise experimental treatment variances (Taylor, 1970). The chi-squared values were well above the upper 5% significance level described by Elliot (1971) in the two plant scales that did not fit to the power law. Therefore, agreement with a Poisson series was rejected at the 95% probability level for cv. Hadeer (x2 = 197.18, P < 0.05) and cv. Hynez (x2 = 313.04, P < 0.05) indicating that T. absoluta larvae had clumped distribution on their plant scale. When insect counts for each scale of the three cultivars were pooled together, this law was fitted (P < 0.01) to the three grouped scales, namely leaflets, leaves and plants of the two-year-combined data (Figs 1-3). Consequently, slope values for each of the three grouped scales were used in equation (ii), resulting in the following normalising transformation: Y =
X017, X034 and X0'51 for leaflets, leaves and plants, respectively, across the three tomato fields. The used transformations resulted in different F values of ANOVA tables (Table 3) and consequently in various statistical probability (P) levels for separation between tomato cultivars concerning the mean counts of insect larvae in each of the three scales. This probability (P) for log, square root, Taylor law, and untransformed counts was P < 0.001 for leaflets, P = 0.046, 0.050, 0.051 and 0.075 for leaves, and P = 0.075, 0.058, 0.058, and 0.073 for plants, respectively.
For leaflets, results of mean separation techniques were similar regardless of which transformation was used (Table 4). However, in such a scale, use of any of the three transformations (log, square root, Taylor law) permitted separation of treatment population levels, between cvs Alisa and Hynez (Table 4), in which differences were
statistically nonsignificant when untransformed data were analysed (Table 1). For leaves, only Taylor law could separate treatment population levels, between cvs Alisa and Hadeer (Table 4). Also, for plants, only Taylor law could separate treatment population levels, between cvs Hynez and Hadeer (Table 4).
Table 4. Differences in mean separation of tomato leafminer larval population counts according to Duncan's multiple range test on transformed data.
Spatial scale Tomato cultivar Treatment mean*
Log x Square root Taylor law**
Alisa 0.066c 1.114c 1.032c
Leaflet Hadeer 0.137a 1.231a 1.065a
Hynez 0.098b 1.157b 1.045b
Alisa 3.58a 2.28a 1.359b
Leaf Hadeer 3.11ab 1.90ab 1.681a
Hynez 2.77b 1.68b 1.454b
Alisa 2.48b 3.87b 3.940b
Plant Hadeer 3.57a 5.97a 6.246a
Hynez 2.79ab 4.44ab 4.720b
Data are from experiment in Table 1. The three tomato cultivars were treated differently for control of Tuta absoluta.
* Means in each column sharing a common letter are not significantly (P < 0.05) different according to Duncan's New Multiple Range Test.
** Transformation calculated from y = x(1 - 05b) where x is the population count and b is the aggregation index from Taylor's Power Law (Table 2; Taylor, 1970).
EPN sampling and spatial distributions.
Insect-parasitic nematode isolates of the genus Heterorhabditis Poinar were recovered from 29 (56.9%) of 51 soil samples (21 from tomato + 30 from mango) tested. EPN were not detected from any tomato soil samples in 2011 and 2012, while 29 (96.7%) of 30 samples from mango soil in 2012 were positive. Morphological and molecular identification revealed the presence of H. bacteriophora (25 samples), and two unidentified species designated as Heterorhabditis sp. Mgo2 (in 2 samples), and Heterorhabditis sp. Mgo3 (in 2 samples). These three isolates induced 90.0, 93.3 and 96.7% mortality, respectively, to late larvae instars of the tomato leafminer. All dead larvae were dissected to confirm nematode parasitism.
Individual soil samples from the mango orchard often contained sufficient EPN to infect multiple Galleria. Overall, 152 Galleria of 29 soil samples from the mango orchard were infected with heterorhabditid nematodes with three samples producing one infection, one producing two infections, seven producing three infections, five producing four infections, seven producing six infections, one producing eight infections, two producing nine infections, one producing 12 infections, and two producing 13 infections. Such a large number of heterorhabditid isolates obtained from the mango orchard permits analysis of their distribution in three different levels. Firstly, the distribution of infected G. mellonella larvae within each positive sample was mostly random, that is, chi squared (x2) value for each of 27 positive samples lie between the appropriate 5% significance levels
Fig. 4. Relationship between the logarithms of the mean (X) and variance (Y) of Galleria mellonella larvae infected by heterorhabditid nematodes (n = 29).
for agreement with a Poisson series (x2 = 2-12, P > 0.05). These chi-squared values were well above the upper 5% significance level described by Elliot (1971) in the remaining two positive samples with equal chi-squared value. Therefore, agreement with a Poisson series was rejected at the 95% probability level (x2 = 14, P < 0.05). So, in these two samples, infected G. mellonella larvae were aggregated in the observation schedule of the multiple Gallería bait method. So, the spatial variability of infective EPN-natural populations would mostly change randomly over time. Secondly, when log variance was regressed against log mean for the positive samples (n = 29), Taylor's Power Law fitted (P < 0.01) to the counts of infected G. mellonella larvae obtained from multiple cycles of the Galleria bait method (Fig. 4). The value of parameter b of the Power Law as an index of dispersion was close to one suggesting randomness. This indicated that the possibility of finding larvae infected with G. mellonella during any cycle of the multiple Gallería bait method is not more than expected by chance. Thirdly, the value for the variance-to-mean ratio in the 29 positive samples was significantly greater than unity (chi-square test). The chi-squared value was well above the upper 5% significance and therefore, agreement with a Poisson series was rejected at the 95% probability level (x2 = 59.776, P < 0.05). This high value of the variance-to-mean ratio indicated that the horizontal spatial pattern of heterorhabditid nematodes had aggregated distribution in the mango orchard.
DISCUSSION
Pheromone treatment could significantly affect the abundance of T. absoluta larvae on the three scales for the tested tomato cultivars (Table 1) documenting the efficacy of the pheromone (Caparros Megido et al, 2013). Also, Batalla-Carrera et al. (2010) demonstrated that EPN are capable of infecting larvae, pupae and adults of T. absoluta and killing the larvae inside the galleries in tomato plant leaves. They found that all three EPN tested, i.e., Steinernema feltiae, S. carpocapsae and H. bacteríophora had high efficacy against larvae irrespective of the dose. The high mortality obtained on T. absoluta larvae in the present study is similar to the mortality observed by them (86.6-100%) and by Rodriguez et al. (2005) in Petri dish trials using the entomopathogenic fungus Beauveria bassiana (96% mortality) but superior to results found by Goncalves-Gervasio & Vendamim (2007) using neem seed extract (52.4-95.4% mortality).
All the values of the Power Law parameters a and b (Table 2) lie within the range reported by
Taylor (1961) for other biological contexts. Sánchez et al. (2009) found that the spatial distribution of T. absoluta larvae on the same scales of tomato deviated significantly from three of the most common frequency distributions in three out of 12 studied cases, i.e., larval counts on plant scale in one tomato site and leaf as well as leaflet scale in another site did not adjust either to Poisson or to negative or positive binomial distributions. On the other hand, differences in nematode patch size can affect the relationship of the sample variance to the mean (Duncan et al., 1989). For EPN, this is particularly evident when variable insect (host) size or environmental factors affect the nematode carrying capacity of the soil. Orchards with low carrying capacities in which many samples are positive, e.g., many small-size insects are infected, may have mean EPN population levels similar to orchards with high carrying capacities, but in which fewer samples are positive due to a few infected large-size insects. The EPN variance to mean ratio of the former orchards would be smaller than those of the latter. Therefore, application of the Power Law in a geographic survey using sampling regime similar to the present survey, i.e., sampling individual trees, could distinguish between two subsets of data according to the perceived size of patches of the citrus nematode Tylenchulus semipenetrans (Duncan et al., 1989). This information enables us to select the appropriate EPN application strategy to manage the target insect species, such as EPN augmentation or conservation (Campos-Herrera et al., 2009; Salama & Abd-Elgawad, 2010). Therefore, an insect baiting technique was also used herein to give information on the ability of the nematodes to infect insects. A fine scale (e.g., 5 * 5 cm) distribution of natural populations may also be studied to visualise where events of EPN emergence from insect cadavers have occurred, and give some indication of how dauer juveniles may have dispersed (Spiridonov et al., 2007).
In the present study, F values of ANOVA from transformed counts were generally higher than those of the original counts (Table 3). This resulted in further separation (P < 0.05) of insect means for each of these scales could be obtained through Taylor's Power Law, natural logarithm, and square root transformations of T. absoluta counts (Table 4). Yet, among the three transformation methods used, only the Power Law has the advantage of maintaining the untransformed counts-based separation between the means carried out for both plant and leaf scales (Table 4). This Law covers a wider range compared to the above-mentioned distributions and the transformations derived from
the Law parameter b (Table 2; Figs 1-4) are often easier to apply, as index of dispersion, than those of the negative binomial (e.g., Elliott, 1971; McSorley et al, 1985; Salama & Abd-Elgawad, 2010). Transformations based on Taylor's Power Law were found by Abd-Elgawad & Hasabo (1995) to be more effective than log or square root transformations in stabilising variances for nematode population data. When the parameter b was used as index of citrus nematode dispersion (Duncan et al., 1989), it differed (P < 0.05) from values which would indicate from the Power Law that log (slope = 2.0) or square root (slope = 1.0) data transformations are most appropriate to equalise experimental treatment variances (Taylor, 1970). Yet, the primary thrust of the present study was to compare the spatial patterns of T. absoluta and its biocontrol agent, EPN, as a solid basis for drawing up an improved T. absoluta management strategy.
The distribution pattern observed often depends upon the scale over which it is measured (Southwood, 1978). Patchy distributions are often an apparent consequence of the distribution of resources upon which a species depends or of interactions among conspecifics or heterospecifics (Stuart & Gaugler, 1994). Admittedly, the very high percentage of positive samples in the mango orchard is an unprecedented case in Egypt (e.g., Shamseldean & Abd-Elgawad, 1994; Abd-Elgawad & Nguyen, 2007; Abu-Shady et al, 2011) and elsewhere (e.g., Cabanillas & Raulston, 1994; Stuart & Gaugler, 1994; Zadji et al., 2013). This was possibly because numerous small Ceratitis capitata larvae, susceptible to EPN (Abd-Elgawad et al., 2012), were seen during sampling in and around Mediterranean fruit fly infected-mangoes that have fallen from the trees. Individual positive samples from this orchard often contained sufficient nematodes to infect multiple Galleria as reported elsewhere (Stuart & Gaugler, 1994). Also, the number of nematodes infecting a Galleria larva under these circumstances is probably highly variable and is probably higher for first infections than later infections, as reported by Stuart & Gaugler (1994), which may have, at least partially, accounted for the randomness in the nematode-infected G. mellonella larvae during the series of extractions by the baiting method within and among most of the positive samples. On the other hand, horizontal spatial patterns of the nematodes were aggregated in the orchard, which agreed with EPN distribution in other regions (Cabanillas & Raulston, 1994; Stuart & Gaugler, 1994; Salama & Abd-Elgawad, 2010). Nematode aggregation can be caused by patchy host distribution coupled with
high reproductive capabilities and limited mobility (McGraw, 2009). Yet, this latter found that nematode dispersion values at the scale of several square meters fluctuated from aggregated to random in very short time frames (1-2 wk). Such a cycling of dispersion demonstrate that EPN probably exist as complex metapopulations within an area of several square meters, fluctuating from uniform or random to aggregated in response to migration, immigration, death, infection and reproduction (Stuart et al., 2006; Spiridonov et al., 2007).
The spatial distribution of T. absoluta larvae was clumped at plant, leaf and leaflet scales at four sites in Buenos Aires province, Argentina (Sanchez et al., 2009). In parallel, eight out of the nine slope values of the insect in the present study (Table 2) had slope > 1.0 indicating different degrees of larval aggregations too. Only plant scale of cv. Hynez had slope (b) ~ 1.0 (Table 2) suggesting random distribution. Moreover, values of b generally increased from plant to leaflet at each field, indicating a more aggregate distribution as scale diminished, which corroborated a previous report (Sanchez et al., 2009). This same trend was quite obvious at each spatial scale when analyses were also carried out for all fields/cultivars combined (Figs 1-3). Also, when Spiridonov et al. (2007) used both Lloyd's index of patchiness and spatial analysis by distance indices (SADIE), they found that EPN species had aggregated distributions. Yet, based on assigning these soil-inhabiting EPN into four groups (I-IV) of increasing physiological age, Lloyds index identified group I as being the most aggregated, whereas SADIE identified group II as the most aggregated; also probably because group I nematodes were aggregated at a scale finer than that used in the sampling regime (Spiridonov et al., 2007).
It is likely that the undisturbed habitat of the perennial mango has contributed to such an occurrence of EPN reported herein. On the other hand, because of crop, pest and natural enemy removal at the end of the tomato growing season as well as soil ploughing and preparation for the new crop, and because of usual chemical applications on tomato, no EPN were detected from the three fields. Hence, such factors may hinder the reliability of EPN as conservation biological control agents. Cabanillas & Raulston (1994) attributed low levels of suppression of corn earworm prepupa and pupa by naturally occurring S. riobravis to the aggregated distribution of the nematode in soil.
REFERENCES
Abd-Elgawad, M.M.M. 1992. Spatial distribution of the phytonematode community in Egyptian citrus groves. Revue de Nématologie 15: 367-373.
Abd-Elgawad, M.M.M. & Hasabo, S.A. 1995. Spatial distribution of the phytonematode community in Egyptian berseem clover fields. Fundamental and Applied Nematology 18: 329-334.
Abd-Elgawad, M.M.M. & Nguyen, K.B. 2007. Isolation, identification and environmental tolerance of new heterorhabditid populations from Egypt. International Journal of Nematology 17:116-123.
Abd-Elgawad, M.M.M., Kabeil, S.S.A. & El-Taweel, M.I.A. 2009. Agricultural features of Borg El-Arab region, Egypt with special reference to nematodes and other ecological factors. International Journal of Nematology 19: 33-39.
Abd-Elgawad, M.M.M., Abdel-Razek, A.S. & Abd El-Wahab, A.E. 2012. Efficacy of kaolin, entomopathogenic nematodes and fungi on the Mediterranean fruit fly infecting citrus. Egyptian Journal of Agronematology 11: 178-192.
Abd-Elgawad, M.M.M., Jian, H., Qiao, Y. & Hammam, M.M.A. 2013. Entomopathogenic nematodes and their potential against insects and phytonematodes in Egypt. In: Proceedings of the 10th International Nematological Symposium (M.V. Pridannikov, Zh.V. Udalova & N.N. Butorina Eds.). pp. 98-101. Moscow (Bolshie Vyazemy), Russia, Russian Research Institute of Phytopathology.
Abolmaaty, S.M., Hassanein, M.K., Khalil, A.A. & Abou-Hadid, A.F. 2010. Impact of climatic changes in Egypt on degree day's units and generation number for tomato leaf miner moth Tuta absoluta (Meyrick) (Lepidoptera: Gelechiidae). Nature and Science 8: 122-129.
Abu-Shady, N.M., Shamseldean, M.M., Abd-Elbary, N.A. & Stock, S.P. 2011. Diversity and distribution of entomopathogenic nematodes (Heterorhabditidae and Steinernematidae) in Egypt. Journal of Nematology 43: 224.
Barker, K.R. & Campbell, C.L. 1981. Sampling nematode populations. In; Plant-parasitic nematodes, V. III (B.M. Zuckerman & R.A. Rohde Eds.). pp. 451-472. New York, USA, Academic Press.
Batalla-Carrera, L., Morton, A. & Garcia-Del-Pino, F. 2010. Efficacy of entomopathogenic nematodes against the tomato leafminer Tuta absoluta in laboratory and greenhouse conditions. BioControl 55: 523-530.
Bechinski, E.J. & Pedigo, L.P. 1981. Population dispersion and development of sampling plans of Orius insidiosus and Nabis spp. in soyabeans. Environmental Entomology 10: 835-841.
Bedding, R.A. & Akhurst, R. 1975. A simple technique for the detection of insect parasitic rhabditid nematodes in soil. Nematologica 21: 109-110.
Cabanillas, H.E. & Raulston, J.R. 1994. Evaluation of the spatial pattern of Steinernema riobravis in corn plots. Journal of Nematology 26: 25-31.
Campos-Herrera, R., Duncan, L.W., Stuart, R.J., El-Borai, F. & Gutierrez, C. 2009. Entomopathogenic nematode ecology and biological control in Florida citrus orchards. In: Integrated Management of Arthropod Pests and Insect Borne Diseases (A. Ciancio & K.G. Mukerji Eds.). pp. 97126. Dordrecht, the Netherlands, Springer Science + Business Media B.V.
Caparros Megido, R., Haubruge, E. & Verheggen, F. 2013. Pheromone-based management strategies to control the tomato leafminer Tuta absoluta (Lepidoptera: Gelechiidae). Biotechnology, Agronomy, Society and Environment 17: 475-482.
Desneux, N., Wajnberg, E., Wyckhuys, K.A.G., Burgio, G., Arpaia, S., Narvaez-Vasquez, C.A., Gonzalez-Cabrera, J., Catalan Ruescas, D., Tabone, E., Frandon, J., Pizzol, J., Poncet, C., Cabello, T. & Urbaneja, A. 2010. Biological invasion of European tomato crops by Tuta absoluta: ecology, geographic expansion and prospects for biological control. Journal of Pest Science 83: 197-215.
Desneux, N., Luna, M., Guillemaud, T., & Urbaneja, T. 2011. The invasive South American tomato pinworm, Tuta absoluta, continues to spread in Afro-Eurasia and beyond: the new threat to tomato world production. Journal of Pest Science 84: 403-408.
Duncan, L.W., Ferguson, J.J., Dunn, R.A. & Noling, J.W. 1989. Application of Taylor's power law to sample statistics of Tylenchulus semipenetrans in Florida citrus. Journal of Nematology (supplement) 21: 707-711.
Elliott, J.M. 1971. Some methods for the statistical analysis of samples of benthic invertebrates. Freshwater Biological Association (scientific publication no. 25). UK, the Ferry House, Ambleside. 144 pp.
Erian, W.F. & Yacoub, R.K. 2000. The use of the ordinary kriging techniques in measuring the sustainability in sugar beet area, Nubariya, Egypt. International Archives of Photogrammetry and Remote Sensing. 33: 403-419.
Faleiro, J.R., Ashok Kumar, J.A. & Rangnekar, P.A. 2002. Spatial distribution of red palm weevil Rhynchophorus ferrugineus Oliv. (Coleoptera: Curculionidae) in coconut plantations. Crop Protection 21: 171-176.
Ferris, H. 1984. Probability range in damage predictions as related to sampling decisions. Journal of Nematology 16: 246-251.
Ferris, H. 1985. Population assessment and management strategies for plant-parasitic nematodes. Agriculture, Ecosystems and Environment 12: 285-299.
Gaugler, R. & Kaya, H. 1990. Entomopathogenic Nematodes in Biological Control. USA, CRC Press Inc. 365 pp.
Goncalves-Gervasio, R. de C.R. & Vendamim, J.D. 2007. [Bioactivity of aqueous neem seeds extract on the Tuta absoluta (Meyrick, 1917) (Lepidoptera: Gelechiidae) in three ways of application.] Ciencia e Agrotec 31: 28-34 (in Portuguese).
Goodell, P. & Ferris, H. 1980. Plant-parasitic nematode distribution in an alfalfa field. Journal of Nematology 12: 136-141.
Grewal, P.S., Ehlers, R.-U. & Shapiro-Ilan, D.I. 2005. Nematodes as Biocontrol Agents. UK, CABI Publishing. 528 pp.
IBRAHIM, S.A. & HASAN, E.A. 2012. [Integrated Control of Tomato Leafminer, Tuta absoluta]. Egypt, Media Support Center Press of Malawi. 52 pp (in Arabic).
Jacobson, R.J. & Martin, G.A. 2011. A potential role for entomopathogenic nematodes within IPM of Tuta absoluta (Meyrick) on organic tomato crops. IOBC/WPRS Bulletin 68: 71-74.
Marchiori, C.H., Silva, C.G. & Lobo, A.P. 2004. Parasitoids of Tuta absoluta (Meyrick, 1917) (Lepidoptera: Gelechiidae) collected on tomato plants in larvas, state of Minas Gerais, Brazil. Brazilian Journal of Biology 64: 552-554.
McGraw, B.A. 2009. The ecology and the biological control of the annual bluegrass weevil, Listronotus maculicollis Kirby (Coleoptera: Curculionidae) using entomopathogenic nematodes (Rhabditida: Steinernematidae and Heterorhabditidae). PhD Thesis. 182 pp.
McSorley, R., Dankers, W.H., Parrado, J.L. & Reynolds, J.S. 1985. Spatial distribution of the nematode community on Perrine marl soil. Nematropica 15: 77-92.
Miranda, M.M.M., Picanco, M., Zanuncio, J.C. & Guedes, R.N.C. 1998. Ecological life table of Tuta absoluta (Meyrick) (Lepidoptera: Gelechiidae). Biocontrol Science and Technology 8: 597-606.
Mohamed, M.A. & Dosuki, S.M. 2009. [Tomato Production and Trading]. Egypt, Media Support Center Press of Ismailia. 56 pp (in Arabic).
Pearson, E.S. & Hartley, H.O. 1966. Biometrika: Tables for Statisticians. UK, Cambridge University Printing House. 238 pp.
PELCZAR, M.J. & Reid, R.D. 1972. Microbiology. USA, the McGraw-Hill Companies, 164 pp.
Pereyra, P.C. & Sanchez, N. 2006. Effect of two Solanaceous plants on developmental and population parameters of the tomato leafminer Tuta absoluta
(Meyrick) (Lepidoptera: Gelechiidae). Neotropical Entomology 35: 671-676.
Poinar, G.O. Jr. 1990. Biology and taxonomy of Steinernematidae and Heterorhabditidae. In: Entomopathogenic Nematodes in Biological Control (R. Gaugler & H.K. Kaya Eds.). pp. 23-61. Boca Raton, USA, CRC Press Inc.
Rodriguez, S.M., Gerding, P. & France, A. 2005. [Selection of isolates of entomopathogenic fungi for control of tomato moth Tuta absoluta (Meyrick) (Lepidoptera: Gelechiidae)]. Agricultura Técnica 66: 151-158 (in Spanish).
Salama, H.S.A. & Abd-Elgawad, M.M.M. 2010. Spatial patterns of the red palm weevil and applied entomopathogenic nematode Heterorhabditis bacteriophora. Archives of Phytopathology and Plant Protection 43: 689-699.
Sánchez, N.E., Pereyra, P.C. & Luna, M.G. 2009. Spatial patterns of parasitism of the solitary parasitoid Pseudapanteles dignus (Hymenoptera: Braconidae) on Tuta absoluta (Lepidoptera: Gelechiidae). Environmental Entomology 38: 365-374.
Shamseldean, M.M. & Abd-Elgawad, M.M.M. 1994. Natural occurrence of insect pathogenic nematodes (Rhabditida: Heterorhabditidae) in Egyptian soils. Afro-Asian Journal of Nematology 4: 151-154.
Siqueira, H.A., Guedes, R.N. & Picanco, M.C. 2000. Insecticide resistance in populations of Tuta absoluta (Lepidoptera: Gelechiidae). Agricultural and Forest Entomology 2: 147-153.
SOUTHWOOD, T.R.E. 1978. Ecological Methods with Particular Reference to the Study of Insect Populations. UK, Chapman & Hall. 524 pp.
Spiridonov, S.E., Moens, M. & Wilson, M.J. 2007. Fine scale spatial distributions of two entomopathogenic nematodes in a grassland soil. Applied Soil Ecology 7: 192-201.
Stock, S.P. & Hunt, D.J. 2005. Morphology and systematics of nematodes used in biocontrol. In: Nematodes as Biocontrol Agents (P.S. Grewal, R.-U. Ehlers & D.I. Shapiro-Illan Eds). pp. 3-43. Wallingford, UK, CABI Publishing.
Stuart, R.J. & Gaugler, R. 1994. Patchiness in populations of entomopathogenic nematodes. Journal of Invertebrate Pathology 64: 39-45.
Stuart, R.J., Barbercheck, M.E., Grewal, P.S., Taylor, R.A.J. & hoy, C.W. 2006. Population biology of entomopathogenic nematodes: concepts, issues and models. Biological Control 38: 80-102.
Taylor, L.R. 1961. Aggregation, variance and the mean. Nature 189: 732-735.
Taylor, L.R. 1970. Aggregation and the transformation of counts of Aphis fabae Scop. on beans. Annals of Applied Biology 65: 181-189.
Taylor, L.R. 1984. Assessing and interpreting the spatial distribution of insect population. Annual Review of Entomology 29: 321-358.
Torres, J.B., Faria, C.A., Evangelista, W.S. & Pratissoli, D. 2001. Within plant distribution of the leafminer Tuta absoluta (Meyrick) immature in processing tomatoes, with notes on plant phenology. International Journal of Pest Management 47: 173-178.
White, G.F. 1927. A method for obtaining infective nematode larvae from culture. Science 66: 302-303.
Woodring, J.L. & Kaya, H.K. 1988. Steinernematid and Heterorhabditid Nematodes: a Handbook of Techniques. USA, Arkansas Agricultural Experiment Station (South Cooperative Bulletin 331). 30 pp.
Zadji, L., Baimey, H., Afouda, L., Houssou, F.G., Waeyenberge, L., de Sutter, N., Moens, M. & Decraemer, W. 2013. First record on the distribution of entomopathogenic nematodes (Rhabditida: Steinernematidae and Heterorhabditidae) in Southern Benin. Russian Journal of Nematology 21: 117-130.
Zappala, L., Bernardo, U. Biondi, A., Cocco, A., Deliperi, S., Delrio, G., Giorgini, M., Pedata, P., Rapisarda, C., Garzia, G.T. & Siscaro, G. 2012. Recruitment of native parasitoids by the exotic pest Tuta absoluta in Southern Italy. Bulletin of Insectology 65: 51-61.
M.M.M. Abd-Elgawad. Особенности пространственного распределения Tuta absoluta и гетерорабдитидных нематод.
Резюме. Томатная минирующая моль Tuta absoluta представляет собой опасного вредителя, который может быстро размножаться и существенно снижать урожай томатов. Понимание пространственного распределения этих вредителей и их естественных врагов несомненно поможет формированию интегрированной борьбы с ними. Параметры дисперсии в личиночной популяции T. absoluta, вычисленные с помощью правила степени Тейлора, были изучены на уровне растений, листьев и листовых чешуек на трех плантациях томатов в Египте. Значение коэффициента b -параметра, отражающего характер распределения в правиле степени Тейлора, возрастало по мере уменьшения размера исследованного участка (от растения до листовой чешуйки), указывая на возрастание степени аггрегированности распределения при уменьшении размеров участка, для которого оценивается характер распределения. Трансформация данных в соответствии с параметрами правила степени Тейлора давала результаты, сходные с эмпирическими данными, полученными по трем изученным полям. Не удалось выявить энтомопатогенных нематод на трех изученных полях с использованием гусениц вощинной моли Galleria mellonella в качестве приманок, однако, нематоды были обнаружены в 29 из 30 почвенных проб, собранных под плодовыми деревьями манго, неподалеку. Смертность личинок T. absoluta от нематод из культур варьировала от 90,0 до 96,7%, указывая на имеющийся потенциал этих нематод. Оценка пространственного распределения гетерорабдитид в почве сада показала, что каждая отдельная проба почвы содержит достаточное количество нематод для успешного заражения нескольких гусениц вощинной моли. Распределение числа зараженных нематодами G. mellonella в пределах каждой из проб почвы было случайным в 27 из 29 исследованных образцов и аггрегированных в двух оставшихся. В то же время, вероятность обнаружения, зараженных нематодами G. mellonella, в положительных пробах не превышала уровня, ожидаемого при случайном заражении.
