The actual number of insect pests present in a particular field is not possible to count. Therefore, survey is conducted to record the approximate pest population in the crop. The expertise of the surveyor is the primary factor on which the efficiency of the survey depends.
The surveyor must have thorough knowledge of morphological characters of insect pests, based on which these are identified and biology of the pest. For example, one must be able to differentiate the larvae of spotted, pink, and American bollworms in the field.
Some other factors which affect pest survey are discussed below:
Factor # 1. Nature of Sample:
The mechanics of sampling a given species depends upon its life-history and habits, which determine among other things the best time to sample. When only a single stage is being sampled in an extensive study, it is most important to coincide this operation with the peak of population. The timing of sampling is still more critical in cases where the rate of development is faster. Regular sampling is required throughout the season in case of intensive studies designed for the construction of life-tables.
The time of the day may also affect sample pattern considerably. The insects are known to move from one part of the habitat to another because of their diurnal rhythms, as in the case of Moroccan locust, Dociostaurus maroccanus (Thunberg). Many grassland insects move up and down the vegetation not only because of changes in weather, but also at certain times of the day and night.
A majority of the insects may be airborne during the day, as in case of adults of Pieris brassicae (Linnaeus) and other butterflies, but there are still others that are active at night, e.g., the defoliating beetles. It is thus necessary to resort to sampling at the time when the pest is active.
The developmental stage to be sampled depends on the objectives of the study. If the objective desires only one estimate in a generation, a quiescent and easily approachable stage is often selected. When the objective demands frequent sampling during each generation, it is often necessary to sample active stages that are repeatedly appearing in the population rapidly.
If the purpose of sampling is to determine the necessity for taking up control measures, the timing should be such as to give advanced information of an outbreak. The estimation of population levels at any one point in the life-history of an insect would give only the total effect of various factors on all the earlier stages in the generation. The total effect may often be confusing and it is desirable to estimate the population level at as many suitable points in the life-cycle as desirable or practicable.
Factor # 2. Size of Sample:
The sampling unit for its size in population estimation should meet the following criteria:
i. The size of sample is such that all units of the universe have an equal chance of selection.
ii. The sampling unit must be easily delineated in the field.
iii. The sampling unit should be of such size as to provide a reasonable balance between the variance and cost.
The optimum size should be based on variance components and cost function formulae, the latter being generally expressed in terms of man hours and would include the time required to select and take a sample, and do the counting together with the time spent in moving from one sampling site to another. The determination of the size of the sample unit is not always easy because the most efficient size would also depend on the way in which the sampled organisms are spatially distributed.
Factor # 3. Number of Samples:
The total number of samples to be taken for estimating pest population depends on the degree of precision required. For all practical purposes, an error of 10 per cent of the mean is usually considered as the standard. Accordingly at the chosen probability level, the estimated mean will have the chance of being within 10 per cent of the true mean.
Within an ecologically homogeneous habitat, the number of samples required can be worked out by the formula:
N = (ts/Dx̅) 2
Where,
N = Number of samples
t = It is a constant, depending upon the number of samples and is obtained from the statistical tables for ‘student t’ (for more than 10 samples, t approximates to 2 at 5% level)
s = Standard deviation
D = Desired level of accuracy expressed as a decimal (normally 0.1)
x̅ = Mean number of individuals per sample.
Generally, a fixed number of samples have to be taken every time the population density needs to be determined. For assessing pest density in relation to need based control measures, we can resort to sequential sampling in which case the total number of samples taken is variable.
Very few samples would be required in the case of extremely high or extremely low density populations and this would result in considerable saving of time and effort. For devising such a plan, however, extensive preliminary information is required about the type of distribution pattern and the level of densities that can be tolerated and those that are associated with the economic damage to the crop.
Factor # 4. Sampling Habitat:
The different insect pests survive and feed in various parts of the plants, e.g., leaves, stem, inside the stem, inside fruiting bodies, inside flowers, in soil, etc. Thus the specific habitat must be defined for different pests before starting survey. For example, the survey of white grubs should be done by digging out the soil sample and for whiteflies one should take the samples from leaves of plants.
Depending upon the potential damage, the ETLs of various insect pests vary in different habitats, viz. root (rice)-root weevil, Hydronomidus molitor Faust (2 grubs/hill); stem (rice)-yellow stem borer, Scirpophaga incertulas (Walker) (5% dead hearts or white ears); leaf sheath (rice)-brown planthopper, Nilaparvata lugens (Stal) (10 hoppers/hill); leaf (cotton, sucking pest)-whitefly, Bemisia tabaci (Gennadius) (6-8 adults/leaf); leaf (groundnut, chewing pest)-tobacco caterpillar, Spodoptera litura (Fabricius) (20-25% defoliation); floral form (cotton)-American bollworm, Helicoverpa armigera (Hubner) (5-10% infestation); shoot (sorghum)- shoot fly, Atherigona soccata Rondani (15% dead hearts); pod (pigeonpea)-pod borer, H. armigera (5 eggs or 3 small larvae/plant); fruit (citrus)-fruit fly, Carpomyia vesuviana (Costa) (1-2% incidence)
Factor # 5. Sampling Pattern:
To collect samples, the surveyor follows a particular path which is called as sampling pattern. Various sampling patterns are followed to fulfill different objectives; however, the important thing is that the pattern should be unbiased.
Following patterns are followed for insect pest monitoring:
(i) Random Sampling:
This is the simplest and most widely used design and provides an unbiased estimate of the population. The visual selection of sample is difficult and may result in non-randomness. For example, it is a natural tendency of human being to select damaged plant or plant part. Therefore, some pre-determined pattern such as U, X, W, etc., should often be used to walk in a field.
(ii) Stratified Sampling:
When there is great variability in sampled area, the habitat is divided into different strata based on variation with respect to a particular character, from which random samples are taken. This design is powerful alternative to simple random sampling. This technique assures coverage, use sampler’s knowledge and gives precision.
(iii) Systematic Sampling:
This pattern of sampling provides the best combination of reliability and cost. The first sample is taken at one reference point and subsequent sampling is done at fixed intervals. A particular route is usually followed through the growing area to collect the samples.
Factor # 6. Types of Sampling:
To estimate the insect population in a particular area, sampling is done from a limited area.
However, different types of sampling are described herein:
i. Sequential Sampling:
An increasingly important type of sampling programme in IPM is the sequential sampling or sequential decision programme. Such programmes are based on insect dispersion pattern and economic decision levels. In this type of sampling, the total numbers of samples are variable and depend on the population density and distribution pattern.
A sequential sampling plan (SSP) requires only a few samples to be taken at very high or low populations and hence expenditure in terms of time and efforts is minimized. More samples are required when pest population density is near the dividing line between endemic and outbreak levels so that a management decision can be taken confidently.
Besides reducing the cost of sampling, sequential sampling has also helped to reduce insecticide use by 25-35 per cent for the control of pink bollworm on cotton. However, the utility of the method diminishes as the number of species or age classes being monitored increases. The greater these numbers the more likely are one or more species to be close to the action threshold requiring that the scouts or farmers sample upto an upper sample number limit.
ii. Variable Intensity Sampling:
Variable intensity sampling (VIS) provides either a classification of density or an estimate of density if the density is within a fixed distance from threshold value. It is designed to ensure a representative sample throughout the sample area and sampling is done more intensely when mean pest population is close to economic threshold level. Because it is designed to meet both objectives, it is very appealing and when possible should be more widely used.
The biggest drawback of VIS is that it is more difficult to implement than other sampling procedures because the decision to continue sampling is not simply yes or no; instead the number of samples to be taken at a particular sampling location in a field is a function of the cumulative number of samples and the current estimate of mean density. This generally requires the use of programmable calculator or computer.
iii. Double Sampling:
Double sampling is occasionally more convenient and useful than sequential sampling. The basic principle is to take an initial sample and use the information so obtained to decide about the size of a second sample, if required.
iv. Binomial Sampling:
Sampling can sometimes be made easier and less time-consuming by substituting binomial counts for complete counts. Binomial sampling is found on defining a relationship between the density of organisms (m) per sample unit and the proportion of sample unit with more than T organisms (1 – PT), where T = 0, 1, 2.
Historically, T = 0 has been used most often, but there are compelling reasons to use some other tally threshold. Binomial sampling is the most economical and feasible field sampling method for many organisms. The most critical aspect of binomial sampling is the knowledge of the formulae relating to the binomial proportion (PT) obtained from field sampling to the mean density (m).