1. Thripidae), within pepper fields on St. Vincent.
Distribution of the Chilli Thrips, Scirtothrips dorsalis Hood (Thysanoptera:
Thripidae), within pepper fields on St. Vincent.
S. Dorsalis damage
D. R. Seal, M. Ciomperlik, M. L. Richards and W. Klassen
1University of Florida-IFAS, Tropical Research and Education Center, Homestead, FL 33033; 2USDA APHIS PPQ CPHST, Pest Detection Diagnostics and Management Laboratory, N. Moorefield Rd., Bldg. 6414, Edinburg, TX ; 3Ministry of Agriculture and Fisheries, St. Vincent, Richmond Hill, Kingstown, St. Vincent and the Grenadines
S. dorsalis
S. dorsalis
S. dorsalis
Abstract: Spatial distribution patterns of Scirtothrips dorsalis Hood adults on `Scotch Bonnet’ pepper in St. Vincent in 2004 and 2005 were determined using Taylor’s power law and Iwao’s patchiness regression. The distributions of S. dorsalis adults on the terminal leaves were aggregated irrespective of plot size (6, 12, 24 and 48 m2). The optimum sample size in `Scotch Bonnet’ pepper plots, when the estimated population density is 0.5 per terminal leaf sample, is 9 with a 40% precision level; and this sample size is sufficient at a 10% precision level when the estimated density is 2 per terminal leaf sample.
KEYWORDS. Scirtothrips dorsalis, spatial distribution, within plant distribution, pepper, alien invasive species, Caribbean
Table 1. Distribution of S. dorsalis adults on terminal leaves in a scotch bonnet
pepper field on William’s Farms, St. Vincent during October, 2004
Table 2. Distribution of S. dorsalis adults on terminal leaves in a scotch bonnet
pepper field on William’s Farms, St. Vincent during March, 2005
Taylor’s Power Law Iwao’s patchiness regression
Introduction. Scirtothrips dorsalis Hood is a major pest of various vegetable crops, cotton, citrus and other fruit and ornamental crops in eastern Asia, Africa, and Oceania (Ananthakrishnan 1993, CABI/EPPO 1997, CAB 2003). This pest occurs on all above-the-ground plant parts of its hosts, and by feeding causes scarring damage (Chang et al. 1995). In India, S. dorsalis is a severe pest of chilli pepper and hence it is known as the chilli thrips (Thirumurthi et al. 1972). In Japan, S. dorsalis is known as the yellow tea thrips (Toda and Komazaki 2002). The Florida Nurserymen and Growers Association considers S. dorsalis as one of the thirteen most dangerous exotic pest threats to the industry (FNGA 2003). Venette and Davis (2004) indicate that the potential geographic distribution of S. dorsalis in North America would extend from southern Florida to north of the Canadian boundary, as well as to Puerto Rico and the entire Caribbean region. It appears that most of Latin America is suitable for colonization by this alien invasive species. S. dorsalis is a key vector of tomato spotted wilt virus (TSWV) which causes bud necrosis disease (BND), an important disease of peanut in India (Amin et al. 1981, Mound and Palmer 1981, Ananthadrishnan 1993).
There are no published reports on spatial distribution patterns of S. dorsalis. Such information is essential in the development of tactics and strategies for managing this pest. Here we report on the distribution of S. dorsalis within pepper fields.
Taylor’s Power Law Iwao’s patchiness regression
Plot size m2 r2 a b 6
0.38
-0.05
1.02RAN
0.07
-0.13
1.04RAN 12
0.22
-0.17
0.79REG
0.01
0.44
0.43REG 24
0.99
0.36
0.40REG
0.32
0.48REG 48
-0.48REG
Plot size acre) r2 a b
.001
0.28
0.09
0.63ag
0.17
0.70
0.56ag
.002
0.37
0.08
0.38ag
0.34
0.83
0.39ra
.005
0.99
0.04re
1.28
-0.03re
.011
0.01
1.29ag
-0.28
AGG, aggregated distribution, b significantly > 1; REG, regular distribution, b significantly < 1; RAN, random distribution, b not significantly different from 1. These distributions are significant at P < 0.05 based on Student’s t-test. Numbers of plots (n) are 2, 4, 8, and 16 for the 6, 12, 24 and 48 m2 sized fields, respectively.
Materials and Methods. Within field distribution of S. dorsalis was conducted in `Scotch Bonnet’ pepper fields in William’s Farms (one field: Field 1) in October, 2004; and in William’s Farms (one field: Field 2) and Baptist’s Farms (one field: Field 3) in March, All fields were located in George Town, St. Vincent, and were about 3,035 m2 each. Each field was planted to `Scotch Bonnet’ pepper into a deep soil. The plants were spaced 90 cm within a row and 1.2 m between rows. Plants were maintained by using standard cultural practices recommended for Saint Vincent. The pepper plants were not treated with any insecticides, but they received the recommended fungicide and fertilizer applications. Plants were treated with Manzate and Bravo at 7-10 d intervals. Plants were drip-irrigated weekly. In both years, the studies were initiated mo after planting the crop.
For the purpose of studying distribution patterns of S. dorsalis, each field was divided into 60 equal plots, each 4.6 m long and 1.2 m wide and contained 5 plants. The within field distribution of S. dorsalis was studied in plots of four different sizes- 6, 12, 24, and 48 m2. Spatial distribution of S. dorsalis in ‘Scotch Bonnet’ pepper fields was studied in both years by collecting terminal leaves contained in a group of 3-4 leaves at the tip of a branch. From each of five randomly selected plants, one such group of terminal leaves was excised and placed in a zip lock bag. The sampled material from a plot was placed in a zip-lock bag to prevent escape of S. dorsalis. All samples were transported to the laboratory for further processing.
The Spatial distribution patterns of S. dorsalis were determined by using Taylor’s power law (Taylor 1961) and Iwao’s patchiness regression (Iwao 1968). Taylor’s power law parameters were obtained by the regression of log10-transformed variances, s2, on log10-transformed mean numbers of S. dorsalis adults and larvae per sample by means of the linear regression model: log s2 = log a + b log x (Taylor 1961). Accordingly a population with an aggregated distribution has a b value > 1, while this value is significantly less than 1 for a regular distribution, and not significantly different from 1 for a random distribution. The fit of each data set to the linear regression model was evaluated by calculating the r2 value. The Student t - test was used to determine if the slopes (b values) obtained by means of the linear regression procedure were equal to 1, significantly greater than 1 or significantly less than 1 (Neter & Wasserman 1974). Separate regression equations were calculated for different sample types and plot sizes. Variation in plot size was achieved by pooling data from related plots to obtain a range in plot size from the smallest plots used (6 m2 ) to substantial fields ( 3,035 m2). In a similar manner, Iwao’s patchiness regression was calculated for each data set. Iwao’s patchiness regression (x* = a + bx), which may be seen as parallel to Taylor’s power law, is the regression of mean crowding, x*, on the mean x (Lloyd 1957, Iwao 1968), a is a sampling factor depending on the size of the sampling unit and b is the index of aggregation in the population.
Sample Size Requirements. In order to estimate the population density at a given level of reliability (Wilson & Room 1982), the number of samples, n, required for a particular plot size was determined by the equation:N = c2t2axb-2
where c is the reliability (half of the width of the confidence interval as percentage of the mean), a and b are the coefficients of Taylor’s power law, x is the mean density, and t is Student’s t-value determined with n-1 degrees of freedom. This t value approximates to 2.0 when n is large. Sample sizes were determined at three levels of precision (0.10, 0.20 and 0.40) for densities of 0.5, 1.0, and 2.0 adults or larvae per sample, respectively.
Table 3. Distribution of S. dorsalis adults on terminal leaves in a scotch bonnet
pepper field on Baptist Farms, St. Vincent during March 2005
Taylor’s Power Law Iwao’s patchiness ression
Plot size m2 r2 a b 6
0.61
-0.25
0.12REG
0.64
-0.95
1.72AGG 12
0.69
-0.32
2.63AGG
0.58
-1.18
1.89AGG 24
0.95
0.48
3.30AGG
0.99
-2.83
2.79AGG 48
0.59
3.93AGG
-3.47
3.20AGG
Table 4. General distribution patterns of S. dorsalis adults on terminal leaves of `Scotch Bonnet’ pepper based on cumulative data collected from three fields, one on Baptist Farms (March 2005) and two on William’s Farms (October 2004 and March 2005).
Plot size (m2) n r Equation
Taylor’s Power Law
log s2 = log x
log s2 = log x
log s2 = log x
log s2 = log x
Iwao’s Patchiness Regression
x* = x
x* = x
x* = x
x* = x
In the equations x represents the mean of the samples, s2, and x* is the mean crowding index. Slopes of all equations are significantly different from 1.0 (P > 0.05), indicating aggregated distribution of Scirtothrips dorsalis adults on terminal leaves
Terminal leaf sample
References Cited
Amin, P. W., and J. M. Palmer Identification of groundnut Thysanoptera. Tropical Pest Management 31: , 340, 344.
Ananthakrishnan, T. N Bionomics of thrips. Annual Review of Entomology 38:71-92.
CAB Crop protection compendium: global module. CAB International, Wallingford, UK.
CABI/EPPO Quarantine Pests for Europe, 2nd Ed. CAB International, Wallingford, UK.
Chang, N.T Major pest thrips in Taiwan. In T. Lewis [ed.], Thrips Biology and Management. Plenum Press, New York.
Cho, K., et. al Spatial distribution and sampling plants for Thrips palmi (Thysanoptera: Thripidae) infesting fall potato in Korea. J. Econ. Entomol. 93 (2):
Gowda, G., E. Ramaiah, and C. V. K. Reddy Scirtothrips dorsalis (Hood)
(Thysanoptera; Terebrantia: Thripidae) a new pest on cashew (Anacardium occidentale L). Current Research 8:
Iwao, S A new regression model for analyzing the aggregation pattern of animal populations. Res. Popul. Ecol. (Kyoto) 4:
­Mound, L. A., and J. M. Palmer Identification, distribution and host plants of the pest species of Scirtothrips (Thysanoptera: Thripidae). Bull. of Entomol. Res. 71:
Taylor, L.R aggregation, variance and the mean. Nature (London) 189: Assessing and interpreting the spatial distributions of insect populations. Annu. Rev. Entomol. 29:
Toda, S., and S. Komazake Identification of thrips species (Thysanoptera: Thripidae) on Japanese fruit trees by polymerase chain reaction and restriction fragment length polymorphism of the ribosomal ITS2 region. Bulletin of Entomological Research 92:
Vanette, R.C. and E. F. Davis Chilli thrips/yellow thrips, Scirtothrips dorsalis Hood, (Thysanoptera: Thripidae) Mini Pest Risk Assessment. Univ. of Minnesota, St. Paul, MN. 31pp.
Results and Discussion
Distribution of S. dorsalis in 2004.
For adults the values of r2 obtained with both Taylor’s power law and Iwao’s patchiness distribution were higher for plot sizes of 24 and 48 m2 than for 6 and 12 m2 plots (Table 1). This indicates a good fit of both models to the data on adults using terminal leaf as sampling unit in 24 and 48 m2 plots. For larval populations on terminal leaves (Data not shown) the r2 values obtained with both Taylor’s Power Law and Iwao’s Patchiness regression showed a good fit to the data ( r2 = 77 – 99) from all of the plot sizes. The slope in either model was significantly higher than 1.00 (P > 0.05) indicating that the distribution of larval populations in all plots, irrespective of size, was aggregated.
Distribution of S. dorsalis in 2005
With respect to data collected at William’s Farms the results from both Taylor’s power law and Iwao’s patchiness regression were in agreement that the distribution of adults on terminal leaves in the 6 m2 plots was random and that it was regular in the 12, 24 and 48 m2 plots (Table 2). The values of r2 from Taylor’s power law ranged from 0.22 – 0.99 indicating moderate to good fit to the data collected from the various plots, while those from Iwao’s patchiness regression were low for 6 and 12 m2 plots (indicating poor fit to the data), and 0.99 for the 24 and 48 m2 plots (indicating a very good fit to the data in larger plots (0.99). The distribution pattern of S. dorsalis adults in various plot sizes at William’s farms was fairly similar in two seasons in 2004 and 2005.
On Baptist Farms, the adult distribution on terminal leaves based on Taylor’s power law was regular in the smallest plots (6 m2) and aggregated in all of the larger plots (Table 3). However Iwao’s patchiness regression showed an aggregated adult distribution in all plot sizes including the smallest. The high r2 values indicated a good fit of the data from all plot sizes to both Taylor’s power law and Iwao’s patchiness regression.
When the cumulative data on adult from three fields, two on William’s Farms (October 2004 and March 2005), and one on Baptist Farms (March 2005) were considered, the distribution of S. dorsalis adults on terminal leaves was aggregated irrespective of plot size (Table 4). The value of b ranged from 1.10 to 1.63 for Taylor’s power law and from 1.28 to 1.82 for Iwao’s patchiness regression. Hence both methods consistently gave slope values corresponding to an aggregated distribution from cumulative data involving one set of samples collected in the rainy season and two collected in the dry season.
Numbers of leaf samples necessary for reliable estimate of S. dorsalis. S. dorsalis densities ranged from 0.5 to 1.0 per terminal leaf sample in fields that suffered economic damage. This information was used to determine the optimum sample size (OSS) for estimating densities of S. dorsalis adults in pepper fields based on the method of Wilson & Room (1982). The OSS becomes large when the desired level of precision is high (i.e., within 10% of the mean) and when the average density of insects per sample is low. For example, the number of leaf samples required to estimate S. dorsalis abundance in 24 m2 is 140 at the 10% precision level and the average adult thrips density is 0.5 per sample (Table 5). It would be prohibitively labor intensive to collect such a large number of samples from a small pepper field. By relaxing the precision level to 40% in this example, the number of samples required to obtain an estimate of the population density can be reduced to 9 for the same size field (24 m2), and this is a practical number of samples. Population density strongly affects the number of samples required. Thus if the density was assumed to 2 adults per sample then number of needed samples for an estimate at the 10% precision level would be 9.
From our results, we draw the following conclusions: (1) Generally S. dorsalis populations tend to be aggregated irrespective of plot size; 2) the optimum sample size in `Scotch Bonnet’ pepper plots, when the estimated population density is 0.5 per terminal leaf sample, is 9 with a 40% precision level; and this sample size is sufficient at a 10% precision level when the estimated density is 2 per terminal leaf sample.
Table 5. Number of terminal leaf samples required for reliable estimates of S. dorsalis adult population densities in `Scotch Bonnet’ pepper fields on St. Vincent.
Field size Mean adults Sample size at different levels
(m2) per sample of precision
Acknowledgments. We are very grateful to the Plant Quarantine Division, Ministry of Agriculture, Industry and Labour, Kingstown, Saint Vincent and the Grenadines for the use of laboratory facilities, local transportation and arrangements with growers. This study could not have been accomplished without the facilitation and encouragement of Mr. Philmore Isaacs, Chief Agricultural Officer. Also we are grateful to Mr. Emil Williams and Mr. Lauron Baptist for allowing us to conduct the studies on their farms. Financial resources and guidance were provided by the Animal and Plant Health Inspection Service, USDA through the leadership of Dr. Daniel A. Fieselmann, National Science Program Leader and Ms. Carolyn T Cohen, Caribbean Area Director. In addition financial support was provided by the Florida Agricultural Experiment Station and the University of Florida’s Center for Tropical Agriculture.