Spatial distribution and sequential sampling plans for the banded sunflower moth (Lepidoptera: Cochylidae) eggs on sunflower

Transcription

1 University of Nebraska - Lincoln From the SelectedWorks of Gary J. Brewer 1996 Spatial distribution and sequential sampling plans for the banded sunflower moth (Lepidoptera: Cochylidae) eggs on sunflower C Peng Gary J. Brewer, University of Nebraska-Lincoln Available at:

2 Entomologia Experimentalis et Applicata 79: , Kluwer Academic Publishers. Printed in Belgium. 235 Short communication Spatial distribution and sequential sampling plans for the banded sunflower moth eggs in sunflower Chengwang Peng & Gary J. Brewer* Department of Entomology, North Dakota State University, Fargo, ND 5815, USA *Author for correspondence Accepted: August 29, 1995 Key words: Cochylis hospes, Helianthus, spatial distribution, population sampling i 1 I I~ Introduction The banded sunflower moth, Cochylis hospes Walsingham (Lepidoptera: Cochylidae), is distributed in most parts of North America (Beregovoy et al., 1989) and in the northern Great Plains, and is a major pest of cultivated sunflower, Helianthus annuus L. (Charlet et al., 1987). This moth is univoltine with adult emergence beginning in early July (Charlet & Gross, 199). Beregovoy & Riemann (1987) found that the R3 sunflower stage (a plant with an immature bud more than 2. cm above the nearest leaf, Schneiter & Miller, 1981) is the most attractive for oviposition by the banded sunflower moth. Females oviposit on the outer surface of the bracts and on the back of sunflower heads (Beregovoy & Riemann, 1987; Charlet & Gross, 199). Early instars feed on the inner surface of the bracts or florets which may reduce the total number of mature sunflower seeds (Charlet & Busacca, 1986). Later instars feed on developing kernels, damaging an average of 6-7 seeds per larva (Charlet & Gross, 199), which may be largely or entirely consumed (Peng & Brewer, 1995). Management strategies for the chemical control of banded sunflower moth are to target either adult moths or the early stage of larvae. Because adult moths aggregate at field margins, in weedy areas, or in adjacent crops before their diurnal movement into sunflower, they are difficult to control (McBride et al., 1984). Charlet & Busacca (1986) reported that insecticidal control of larvae was effective, but the economic threshold has not been determined. One moth per two plants is recommended by McBride & Charlet (1991) as a treatment threshold. However estimating adult moth density in fields is difficult because moths rest on the undersides of the lower leaves of sunflower plants during the day, are small, and difficult to see, especially in fields where the plants are dense and tall. Attempts to use pheromone catches as a treatment threshold failed to correlate between larval infestations and moth catches (Wilde et al., 1994; Blodgett et al., 1994). Evidently, moth counts are not reliable for making control decisions. We propose using counts of the eggs per head to make control decisions. The objective of this study was to determine spatial distribution of banded sunflower moth eggs which would allow the development of sequential sampling plans to estimate banded sunflower moth egg populations. Materials and methods Field sampling. Six commercial sunflower fields (2-5 ha) and a research plot (rv 1 ha) were sampled at bud stages (R3 and R4) for eggs of banded sunflower moth in North Dakota in Plant stage R4 refers to a plant with visible immature ray flowers when viewed from directly above (Schneiter & Miller, 1981). Samples were taken from four sampling routes in each field, which were approximately centered on the four sides (E, S, W, and N) of the field. Each sampling route had three sampling sites which were 6,37, and 67 m from the field margin in the six commercial

3 236 fields and 6, 18, and 37 m from the field margin in the research plot. Ten plants were chosen at random from each sampling site, and heads were removed and sealed in a plastic bag. Heads were stored at 2-3 C in the laboratory for later egg counting. We counted the number of eggs on the outer surface of the involucral bracts (bracts surrounding the floral bud) and on the back of the heads under a dissecting microscope. The bracts grow as a series of overlapping layers around the periphery of the bud and to get distribution on the different layers, the eggs on each were counted separately. The preliminary study indicated that eggs laid on the bracts beyond the fifth layer accounted for < 1% of the total, so we did not examine subsequent layers. Egg density prediction. The edge-effect coefficient (C g /h) defined the ratio of mean number of eggs per head atg m from field margin (EDg) over mean number of eggs per head at h m from the field margin (EDh) (g"?h). It was calculated using: (1) The denominator, ED h, always referred to the mean number of eggs per head at the sampling sites nearest the field margin, i.e., 6 m. So, EDh is ED6 and Eq. 1 can be rewritten as: Data from the research plot were excluded for edge-effect analysis because the field was small, next to other sunflower plots, and banded sunflower moth pheromone traps were placed nearby. Counting the number of eggs on all bracts in a head (about 6 bracts per head) and the back of the heads is time consuming, and is not feasible in field sampling. The alternative is to count the number of eggs on only a few bracts and from that estimate the total number of eggs per head. This required determining the relationship between the number of eggs on a few bracts and the total number of eggs per head. We determined the relationship between the number of eggs on two, four, or six bracts in the first layer and the total number of eggs per head using a linear regression technique (SAS Institute, 1987). Distribution. The mean egg density (m) per head and variance (s2) were calculated for plants in each sampling site. Taylor's power law which relates the variance (S2) to the mean (m) by S2 =am b (Taylor, 1961) (2) was used to determine the spatial distribution of egg counts. To estimate a and b, the values of In(s2) were regressed against those of In(m) using the model In(s2) = In(a) + bln(m) (3) The slope (b) indicates a uniform, random, and aggregated distribution when b«; 1, b = 1, and b> 1, respectively. The hypothesis, Ho:b = 1, was analyzed using the t-test (Sakal & Rohlf, 1981). Sampling plan. Sequential sampling plans were designed based on Karandinos' (1976) sample size formula: where Zo./2 is a standard normal deviate such that P(Z> Zo./2) = a/2, d is the predetermined half width of the confidence interval (CI) as a proportion of the mean (d = C1/2m) and S2 is the variance of estimated m. Replacing S2 in Eq. 4 with Taylor's variance (am b ) and m with Tn/n (Tn is the cumulative number of individual pests in n sample units) and rearranging the terms, Tn can be solved (Nyrop & Binns, 1991) as: In(Tn) = [In(d 2 jz~/2a)l/(b-2) + [(b-i)j(b-2)]lnn (5) which, in the form of antilogarithm, generates a stop line for sequential sampling based on numerical counts. Three stop lines were determined for numerical sampling of banded sunflower moth eggs using the fixed d values of.2,.3, and.5 (which are equal to sampling error levels [sm/m, where s-«is the standard error of m] of.1,.15, and.25, respectively). These precision levels are a reasonable range for a pest management purpose (Southwood, 1978). Results and discussion Egg density prediction. Based on the Eq. 2, three edge-effect coefficients were calculated for data sets from the six commercial sunflower fields. For sampling sites at 6 m from the field margin, the edgeeffect coefficient was obviously 1 ± (x ± SEM) (n = 24), while those at 37 and 67 m were.468 ±.4 (x± SEM) (n=24) and.391 ±.4 (x± SEM) (n=24) respectively. The relationship between C g /6 and the distance (g) was described by a logarithmic model (Fig. 1): (4)

4 .... s:: <li.... u <li <li o U 1.2., , ~ ~ A 3 2 o r----r----, o w w w w g, Distance from Field Margin (m) Fig. 1. Relationship between edge-effect coefficient (C g I 6 ) of banded sunflower moth eggs and distance (g) from the field margin: Cd6 = In(g) (,-2=.99, P<O.Ol). C g /6 = In(g) (6) Replacing C g /6 with Eq. 2 and rearranging the terms, EDg can be expressed as: EDg = [ In(g)JED6 (7) Eq. 7 can be used to estimate the number of eggs per head at any distance (g) from the field margin using the estimated number of eggs per head at 6 m. The total number of eggs per head can be estimated by sampling a few bracts per head because the number of eggs per two, four, or six bracts in the first layer was significantly related to the total number of eggs per head (Fig. 2) with r? values of.57,.71, and.8, respectively. When an economic threshold is determined for banded sunflower moth, use of Eq. 7 can determine whether an entire field or proportion of a field should be treated. Distribution. From sunflower heads examined, about 8% of the banded sunflower moth eggs were laid on the outer three layers of bracts (Table 1). The number of eggs found in the fifth layer of bracts was only 2.3% of the total number of eggs in the heads. The Taylor power law fit the data sets well, yielding a high coefficient of determination (? =.82, P<O.OOOl). Taylor's parameters a and b were estimated to be and , respectively. The slope b was significantly greater than 1 (t= 7.27, df = 7, P<O.Ol), indicating an aggregated distribution of banded sunflower moth eggs per head. 4 8 B " 3 nl <U ::c.. 2 <U Q.. <Il 1 ~ Eggs per Two Bracts(X2) Eggs per Four Bracts (X4) Eggs per Six Bracts(X6) Fig. 2. Regression of total number of banded sunflower moth eggs per head (Y) on number of eggs per two bracts (X2), four bracts (Xa), or six bracts (X6) in the first layer of bracts. A. Two bracts: Y = X2 (,-2=.57, P<O.OOOI, n = ); B. Four bracts: Y = ~ (,-2=.71, P<O.OOOl, n n = ); C. Six bracts: Y= X6 (,-2=.8, P<O.OOOI, n=). Sampling plans. Sequential sampling plans with fixed precision levels (Fig. 3) were developed based on Taylor's parameters. Sampling stops if the cumulative number of total eggs per head exceeds a previously established stop line and the egg density is estimated based on the cumulative number of eggs found in the number of sunflower heads sampled. Sampling for a precision of.2 is not feasible in the field where egg

5 238 Table 1. Distribution of banded sunflower moth eggs in suntlower heads collected in North Dakota, 1994 Location n Mean no. of bracts ± SEM Mean % eggs ± SEM First layer of bracts (outermost) Second layer of bracts Third layer of bracts Fourth layer of bracts Fifth layer of bracts Back of heads 7.2 ± ± ± ± ± ± ± ± ± ± ± ~ 8 1 (IJ " «l <lj 5 ::r:: l::.5 (IJ on on f.i.l «l r d =.2 d = I d = Sample Size (n heads) Fig. 3. Stop lines for sequential sampling of banded suntlower moth egg populations with fixed d levels (alpha =.5). densities are low (i.e., <1 eggs per head) because excessive samples are required. For example, at the.2 level of precision about 7 samples (sunflower heads) are necessary to estimate a mean of nine eggs per head. To estimate a mean of nine eggs per head, the sampling plan at the precision of.5 requires 11 samples, and the sampling plan at the precision of.3 requires 31 samples. Both sample sizes are manageable in field sampling. Because the basis for the sampling plans is a confidence interval based on a normal model, a minimum sample size should be taken before comparing Tn and stop limits (Nyrop & Binns, 1991). We suggest using a minimum sample size of 1 to limit sampling error rates. In summary, the distribution of banded sunflower moth eggs was aggregated in sunflower. The number of eggs per head decreased from the field margin inward, which was described by a logarithmic model. Use of this model can estimate the number of eggs per head at any distance inside fields from the margin of the field by estimating number of eggs per head at 6 m from the field margin. Number of eggs per head at 6 m can be estimated using the sequential sampling plans. These findings are useful for making control decisions on banded sunflower moth when the relationship between number of eggs per head and damage is determined. We are currently studying the economic threshold for banded sunflower moth. We believe the prediction model and sampling plans presented in the paper will be in use soon for making treatment decisions on banded sunflower moths. Acknowledgments Research support was provided by the USDA-CSRS under project number We thank L. D. Charlet, M. J. Weiss and H. J. Meyer for reviewing the early draft of the manuscript. The comments and suggestions of the anonymous reviewers improved this manuscript and were appreciated. References Beregovoy, V. H. & J. G. Riemann, Infestation phenology of suntlowers by the banded suntlower moth, Cochylis hospes (Cochylidae: Lepidoptera) in the Northern Plains. Journal of Kansas Entomological Society 6:

6 239 Beregovoy, V. H., G. L. Hein & R. B. Hammond, Variations in flight phenology and new data on the distribution of the banded sunflower moth (Lepidoptera: Cochylidae). Environmental Entomology 18: Blodgett, S., S. Pilcher & F. Peairs, Sunflower insect pest situation during 1993 and prospect for In: Proceedings, Eighth Great Plains Sunflower Insect Workshop. USDA, ARS, Northern Crop Science Laboratory, Fargo, ND, USA, p. 9. Charlet, L. D. & 1. D. Busacca, Insecticidal control of banded sunflower moth, Cochylis hospes (Lepidoptera: Cochylidae), larvae at different sunflower growth stages and dates of planting in North Dakota. Journal of Economic Entomology 79: Charlet, L. D., D. D. Kopp & C. Y. Oseto, Sunflowers: their history and associated insect community in the northern Great Plains. Bulletin of Entomological Society of America 33: Charlet, L. D. & T. A. Gross, 199. Bionomics and seasonal abundance of the banded sunflower moth (Lepidoptera: Cochylidae) on cultivated sunflower in the northern Great Plains. Journal of Economic Entomology 83: Karandinos, M. G., Optimum sample size and comment on some published formulae. Bulletin of Entomological Society of America 22: McBride, D. K., L. D. Charlet & c. Y. Oseto, Banded sunflower moth. North Dakota State University Extension Service Bulletin E-823. McBride, D. K. & L. D. Charlet, Banded sunflower moth. North Dakota State University Extension Service Bulletin E-823. Nyrop, J. P. & M. Binns, Quantitative methods for designing and analyzing sampling programs for use in pest management. In: D. B. R. Pimentel [ed.], CRC Handbook of Pest Management in Agriculture. CRC, Boca Raton, FL, pp Peng, C. & G. J. Brewer, Description of achene damage by the red sunflower seed weevil, the banded sunflower moth, and the sunflower moth. Journal of Kansas Entomological Society 68: SAS Institute, SAS User's Guide: statistics. Cary, NC. Schneiter, A. J. & J. F. Miller, Description of sunflower growth stages. Crop Science 21: Sokal, R. R. & F. J. Rohlf, Biometry, The Principles and Practice of Statistics in Biological Research. 2nd Ed. W. H. Freeman and Company, San Francisco. 859 pp. Southwood, T. R. E., Ecological Methods with Particular Reference to the Study of Insect Populations, 2nd Ed. Wiley, New York. Taylor, L. R., Aggregation, variance and the mean. Nature 189: Wilde, G., G. Hein, S. Pilcher & F. Peairs, Regional project for In: Proceedings, Eighth Great Plains Sunflower Insect Workshop. USDA, ARS, Northern Crop Science Laboratory, Fargo, ND., USA, p. 91.