Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
CHAPTER 3 Insect Surveys The detection of insect and related pests and surveys of their distribution and abundance are essential prerequisites to rational control programs. The details of detection and survey activities vary according to ways in which the resultant data will be used to support specific kinds of control programs, but certain aspects of these activities are basic to all such programs. The first principle of pest detection and surveys as related to control is that no control measures should be undertaken against a pest unless it is known that the pest is actually present. In many instances, particularly in agriculture, this principle is not followed. A farmer often follows a control schedule blindly, without bothering to determine whether the pest is truly present on his crop. The economic implications of this are obvious, and such wasteful procedures may also result in the needless destruction of beneficial species and the addition of undesirable chemical contaminants to the environment. After the presence of a particular pest has been confirmed, attention may be turned to the second principle of pest detection: No control measures should be undertaken unless it is known that the pest is present in sufficient numbers to cause economic loss. This presupposes that adequate background research has been done with the particular pest and its plant host to establish the "economic threshold": the population level of the pest that will cause sufficient damage to make control economically desirable. Below this thresh- old, the cost of control exceeds the value of the portion of the crop protected from damage, and the net result is a financial loss. Above this point, the value of the portion of the crop protected exceeds the cost of control, and money will be saved. The economic threshold varies in time and place throughout the season, and is sensitive to weather, agricultural practices, and market and 23
24 INSECT-PEST MANAGEMENT AND CONTROL labor conditions. The determination of a threshold requires economic studies as well as those in biology. It also requires the ability to predict the short- term trend of pest populations; a given population may be increasing toward its economic threshold, but control activities need not be initiated if it can be predicted with confidence that the population will peak and decline below this point. Detection and surveys are especially important to quarantine and eradica- tion programs. They are discussed in detail in Chapter 4. They are also essential to the success of integrated control systems (Chapter 17). KINDS OF SURVEYS All insect surveys are alike in that they attempt to assess some property of the insect to be found over a reasonably large area, and they are extensive in nature. Therefore, they differ from an intensive program carried out on a single population in a limited habitat. The actual property of the fauna mea- sured in a survey will depend on the type of survey. The following broad types may be recognized: qualitative surveys, involving the identification of the different species; quantitative surveys, involving the estimation of the popula- tion of one or more species of insect. QUALITATIVE SURVEYS Qualitative surveys may cover the whole fauna of a region, as does the "California Insect Survey," or they may be limited to one or more of the following categories: potential pests, actual pests, and natural enemies. It is theoretically desirable that, when extensive agricultural development is proposed, the potential pests be surveyed along with the soil, vegetation, and climate. Some work on these lines has been done in Russia. However, in the present state of our knowledge it is difficult to determine which species will adapt particularly well to the proposed new crops and conditions. Also, pests that can become economically important in the future may be scarce at the time of the survey and may be easily overlooked in a natural ecosystem. Nevertheless, certain broad generalizations can be made; e.g., the stalk borers of wild grasses will probably attack graminaceous crops. Such information is often valuable in retrospect in interpreting the origins of pest species and in adding to our basic understanding. Thus, some useful information can be gained from a survey of potential pests. Qualitative surveys of actual pests, although an important initial step in the rationalization of peasant agriculture, seldom provide significant new informa-
INSECT SURVEYS 25 tion in regions where any form of agricultural advisory service has been established for some time. In contrast, a knowledge of the natural enemies of a pest in different parts of its range is an integral and initial component of any program for biological control by importation. In many early biological- control projects, an entomologist made a qualitative survey of the natural enemies of a pest in its supposed country of origin; if any selection of potential control agents for introduction was made, it was based on his intuitive assess- ment of their roles. As shown below, additional relevant evidence may be obtained by a quantitative assessment of the roles of various parasites and predators. Consequently, whenever possible, a survey of natural enemies for biological-control potentialities should be quantitative as well as qualitative. QUANTITATIVE SURVEYS Quantitative surveys involve measuring the abundance of the insect. There are two types of measuring: absolute and relative. Absolute estimates give the number of insects per fixed unit; the unit may be part of the habitat, e.g., leaf or stalk, and give a measure of population intensity, or it may be a unit of square measure, e.g., acre or square meter. The units in which popu- lation intensity is measured will change from field to field as well as from season to season. It is always important to measure the number of habitat units per unit of area, so that estimates in terms of population intensity can be turned into strictly absolute measures and vice versa at any time. Popula- tion intensity is the correct expression of density for some purposes, but for others absolute population per unit area is more appropriate. Relative population estimates are obtained by measuring the population, or samples from it, in qualitative units, on the assumption that these units allow comparison from time to time and place to place. Faith in the strict com- parability of relative measures is not always justified. For example, differences in the numbers of insects caught in two identical light traps in two areas could be caused by real differences in the actual populations, by variations in activity due to weather, or by variations in the efficiency of the trap. The efficiency of a light trap will be reduced by any lack of contrast between it and its sur- roundings. Lack of contrast may be due to light from adjacent tall vegetation or to light from other sources, including moonlight. The same types of inaccuracies can influence almost all types of relative measurements, e.g., variations in efficiency and, when the method depends on the movement of the insect, variations in the level of its activity. Weather and climatic variation will influence both. Relative estimates do, however, play an important role in survey work; compared with absolute estimates, they can usually be obtained at fairly low
26 INSECT-PEST MANAGEMENT AND CONTROL cost both in terms of apparatus and in hours of work-important considera- tions when large areas are to be covered. The principal methods of obtaining relative estimates are the use of various traps (light, water, bait, sticky, and interception) and the making of catches per unit of effort, e.g., the number of larvae per 10 minutes of search. Another rather special type of relative estimate is the population index; this is based on either the products or the effects of the insects, not on the insects themselves. Population indices have been widely used in survey work. One of the earliest examples was the regular survey of the fall webworm, Hyphantria cunea (Drury), made by counting the number of caterpillar tents seen by the observer while driving along given roads; data obtained in this way have recently been of value in population research. A more widely available criterion is the amount of damage caused by an insect. Certain types of insect damage to forests can now be readily surveyed over a large area by aerial photography, in black and white or in color. At the outset of quantitative surveys, it is important to select the ap- propriate stage of the pest for sampling. Since surveys are usually conducted in different areas at different times, the stage selected should be present in the field for as long a period as possible; i.e., its development time should be long. Otherwise, there is the risk that a low count in an area will be caused, not by a low population, but by the insect having already passed through or not having reached that particular stage. However, for most survey purposes, the insect stage selected should be related in abundance to that causing the economic damage to the crop. This is usually the larval stage. The extent to which the egg stage may be used to predict the numbers of larvae will normally be a matter of research in which investigators use methods of analysis described later. USES OF SURVEYS The numbers of a pest species over a wide area may be surveyed in relation to (1) damage; (2) various factors influencing population size, e.g., climate and natural and artificial control; and (3) the future size of the population- prediction and the need for control measures. The special aspects and techniques of these types of quantitative surveys are discussed below, but some consideration must initially be given to the various advantages and disadvantages of the survey method for investigating pest populations. To study the role of different population factors by using the methods of R. F. Morris or of G. C. Varley and G. R. Gradwell, it is necessary to have at least 10 observations, and preferably many more, with extensive work on
INSECT SURVEYS 27 a univoltine insect. This will mean more than 10 years of work. It is to some extent possible to use the data from different areas obtained in surveys and thus to substitute space for time and obtain an indication of the roles of various factors in population control much more quickly. Another advantage of surveys is that any justifiable conclusions based on a wide geographical area may be claimed to be sound generalizations; in contrast, if an intensive study is made of one area, this area may turn out to be peculiar in some way, and generalizations from it may be unreliable. The disadvantage of surveys is that, while gaining so much from their breadth, they of necessity lose depth, and the detailed mechanism of popula- tion regulation is not directly revealed, as it may be by the study of a single cohort. Therefore, the role of different factors is assessed by correlation analysis of, for example, the size of the population against a climatic factor or of the amount of damage against population size. Correlation is not necessarily equivalent to causation; thus, although reduction in yield may be closely correlated with increase in insect population, the two may be linked to a third factor, such as drought, which affects the plant adversely and the insect favorably. Therefore, in the ideal program the interpretation of surveys should be confirmed by carefully designed field tests and, in some instances, by laboratory experiments. DAMAGE SURVEYS Damage surveys are often important both in the initial intensive stage of an investigation and in routine control projects. The object of the survey may be to determine the extent to which an insect species causes significant economic damage over a wide area. This is sometimes referred to as "pest assessment." The criterion of damage can be reduction in economic value or in yield, or it can be some measure of destroyed or damaged parts, e.g., a "dead-heart count" or a "leaf-tatter index." Damage is measured in one of these ways, and the numbers of the pest in the area are also determined- most appropriately as population intensity, or the numbers per plant or per leaf, although relative estimates will often suffice. The two sets of data, damage and insect numbers, from various areas may then be tested for correlation. As already emphasized, this correlation is not the same as causa- tion, and more precise evidence should therefore be obtained from field ex- perimentation. The fields or forests are divided into plots, and in some the insect is reduced to a very low level by treatment with an insecticide; the differences in damage between the treated and untreated plots can then be compared with their insect populations. Other factors, including weather, will be the same in both plots, so that any correlation is likely to be meaning-
28 INSECT-PEST MANAGEMENT AND CONTROL ful. Damage surveys may also serve in the initial phases of a program of work to lay down the population intensity at which a pest reaches its economic- injury level-the level at which control measures are indicated. Accurate determination of this level is basic to any sequential sampling design, or to any integrated control program. Qualitative and quasi-quantitative estimates of damage obtained rapidly and regularly from a large area, as with the Canadian Forest Insect and Disease Survey, can be useful in indicating long-term changes in a pest's importance, the onset of periodic outbreaks, the success or failure of current control measures, and which pests are of sufficient significance to warrant more-intensive work. When collected over a long period, such data may also allow a more sophisticated analysis from a variety of ecological aspects. FACTORS INFLUENCING POPULATION SIZE Surveys can be used as a research tool to investigate the effects of different factors on the size of the pest population. Information about these factors is necessary to give a fully rational basis to a control program. The simplest approach is similar to that utilized in damage studies, the correlation of insect numbers with a factor, e.g., weather or parasitism. The weaknesses of this method have already been indicated, and whenever possible an attempt should be made to gather life-table-type data; that is, absolute population estimates of at least one stage of the insect, in such a way that, where appropriate, the techniques of Morris or of Varley and Gradwell can be applied for analysis. Morris' techniques may be used on populations in perennial crops or forests where one generation («) may be expected to have some relationship to the next (n + 1). The relationship between log n and log n + 1 is then investigated, and the various other factors, e.g., weather and parasitism, are added into the equation for log n, which is then tested to see if it gives a better prediction of log n + 1. The factors giving the best predic- tion are those associated with the major part of the change in numbers from generation to generation and are referred to as the "key factors." Often, as with annual crops, successive annual generations may have no significant direct biological relationship, but for such crops it may be particularly important to determine the role, if any, of natural controlling factors in influencing the size of the economic population. Put another way, Is natural control having any significant effect that we should try and preserve when planning the use of pesticides? This can be answered by estimating the number of eggs laid and the size of the resulting economic population in a number of different areas or years, or both. The resulting population is plotted against the egg population, and the extent to which variations in egg
INSECT SURVEYS 29 numbers influence the numbers of the later stage will be shown by the coef- ficient of determination (r2). A high value of r2 will indicate little significant natural control; a low value may indicate very striking changes in numbers after the egg stage, which may be brought about by natural control. PREDICTION AND THE NEED FOR CONTROL MEASURES One of the fundamental aims of the entomologist is to be able to predict outbreaks of pests, so that appropriate pesticide treatments may be applied when required, but not otherwise. The survey is the basic method enabling him to do this. In general, some preliminary work is necessary before a predic- tion survey can be started. This work, outlined previously, should indicate the numbers of insects associated with the economic-injury level and the extent to which subsequent populations can be predicted from previous generations or from egg numbers. The relative roles of natality and mortality are important, as are the reproductive potential of the female adults, which can often be determined from their size or the size of the pupae. The entomologist's decision to exercise control measures is determined by the numbers per unit of habitat reaching a certain level. Often it is not just the level of the population but the time at which this level is reached that is important, that is, time relative to the development of the crop. There- fore, the statement will be in the form: If the numbers of eggs reach or exceed x per bush before 50% of the flowers have opened, ory per bush after more than 50% of the flowers have opened, control measures should be applied. The difficulty arises of determining when the population level is truly x. This involves problems of sampling, which are discussed in more detail else- where, but one approach is to take a sufficient number of samples to insure that the fiducial limits will lie within 10% of the mean. With very high or very low populations, this method will mean that many more samples are taken than are really needed to be sure that the population level is in excess of x or far below it. This situation can be overcome by the use of sequential sampling, which provides an ideal and efficient plan for a survey. A large amount of initial work is necessary to determine the type of distribution occurring in the field, i.e., whether it is negative binomial, Poisson, or another type, and also the variance-mean relationship. When these statistics are known, and using the equations set out by Morris, Southwood, and others, a sequential sampling chart can be prepared. With these sequential sampling charts on hand, survey operators can immediately read off from each sample taken whether further samples are required or whether control measures should be taken.
30 INSECT-PEST MANAGEMENT AND CONTROL METHODS OF SURVEYS Many methods of assessing insect numbers are available, as already indicated. Some measure absolute population per unit area; others, such as traps, mea- sure only relative numbers. Traps may be influenced by their efficiency and the weather, as well as by numbers of insects present. In survey work, the first decision concerns the stage of the insect to be sampled. When pest control is the object, the aim should be to provide infor- mation sufficiently in advance to allow the control measures to be applied before the damage is done. Therefore, if measures are to be taken against a pest that causes damage as a young larva and has a short developmental period, e.g., the codling moth, it may be necessary to predict from a survey of ovipositing females. When damage is being surveyed, the greatest accuracy will be obtained by counting the damage-causing stage. ADULTS More or less absolute estimates of flying adults can be obtained by using the types of suction trap developed by C. G. Johnson and L. R. Taylor. If the traps are exposed to very different wind speeds, a correction needs to be applied, and Taylor has published tables of correction factors. There are various interception traps that are simpler to construct, but they give results that are more difficult to interpret. Suspended cone nets may be used to trap aphids (and other insects with low flight speeds) in exposed situations, and sticky traps and sticky nets, because of their greater ability to retain the insects, may be used in a wider range of situa- tions. The numbers taken by such traps will be greatly influenced by wind speed and by the flight speed (roughly equivalent to the size) of the insect. Taylor has shown how sticky-trap catches may be corrected to allow for these effects, but it is necessary to have continuous records of wind speed at the site of the trap. Water traps also give records of aerial abundance; like sticky traps, they may be painted various colors, and this will affect the numbers and type of insect caught. Because of the shape of water traps and the resultant wind eddies around them, catches from such traps are less easily corrected for the effect of wind speed than are those from sticky traps. Furthermore, water traps cannot be left unattended for long periods, because the water may evaporate, or, if there is heavy rain, it may overflow. Water traps do, however, have the advantage over sticky traps of preserving the insects in fairly good condition for identification. Light traps also are widely used to trap adult insects, but they can intro- duce sources of error; e.g., the chances of capturing a female moth at light
INSECT SURVEYS 31 may vary with her age, and the efficiency of the trap will vary with the amount of moonlight or other illumination reducing the contrast between the trap and its surroundings. Some adult insects are attracted to bait, at least at certain times in their lives. Bait traps have been useful in surveys of fruit flies, fruit moths, and house flies, but the attractiveness of the bait can vary with its age, the position of the trap, and the number of insects already in the trap. Useful measures of colonization and oviposition can be obtained by placing insect-free "trap plants" or parts of plants, e.g., logs, in the field and recording the numbers of adults or eggs, or both, present after a given time. Shelter traps, consisting of pieces of grooved wood held together with screws, may be used for bark-dwelling and other thigmotactic insects or mites. Hibernating or resting insects will enter these traps and give a measure of their abundance. When trapping adult insects for survey purposes, it is important to remember that although larger catches may be obtained by using a colored trap or including a bait along with visual attraction, "more" may not be "better." The additional biological errors that are introduced can more than outweigh the statistical advantages of larger catches. Adults of many insects, such as beetles and bugs, may be surveyed by sweeping, beating, and other methods referred to later under Larvae and Pupae. EGGS When eggs are laid exposed on the foliage, a count of the number of eggs per leaf or stem provides a reliable, if tedious, measure of abundance; sometimes the eggs may be washed off with gasoline or some other solvent. Eggs in soil may be sampled by the traditional soil-washing methods. Eggs embedded in plant tissue are the most difficult to assess. Attempts to count them by the use of x ray or differential staining have not been very successful, and careful visual examination, often under the microscope, along with the splitting of the stems, is often the only satisfactory method. Therefore, for survey purposes, some stage other than the embedded egg should be used whenever possible. LARVAE AND PUPAE These stages are frequently present on the foliage of plants or trees. The traditional and satisfactory method of sampling arboreal larvae is beating, to dislodge the insects onto a tray. Some workers have found it advantageous
32 INSECT-PEST MANAGEMENT AND CONTROL to have a funnel and a collecting jar attached to the center of the tray, which can be covered with wire netting to prevent the entrance of twigs. The insects can then be rapidly shaken into the collecting jar and their escape prevented. With more-conspicuous, less-agile insects, a plain cloth-covered beating tray is often adequate, the insects being counted directly on the tray. This can be particularly useful in survey work when a sequential sampling program is being followed. Differential falling of the various larval stages, variations in the prevailing weather, and the tendency to sample only the lower branches are possible sources of error with the beating method. Larvae on forage crops are more difficult to survey. Use of the sweep net is a convenient and easy method, but it samples only those insects on the tops of foliage, and their presence there will be influenced by weather condi- tions and the time of day. However, with some standardization in this respect, sweeping still provides an easy, although rough, estimate of the abundance of certain insects for survey purposes, which is especially useful for determin- ing the need for control measures. An absolute estimate of the insect popula- tion on a forage crop can be obtained by delimiting an area of it and sucking the insects out with a vacuum-cleaner type of apparatus. A very convenient portable backpack model, driven by a gasoline motor, is now available commercially. Surveys of a number of crop insects involve dissecting the stems or stalks of infested plants and counting the number of larvae and pupae present. This procedure, together with counts of the number of plants infested, has been used for many years over wide areas in the United States to determine the annual abundance of the European corn borer, Ostrinia nubilalis (Hubner), in corn and to provide data for estimating damage by the insect. The population density of relatively immobile or sedentary pests that are fairly abundant, e.g., scale insects, may be measured exactly by counting directly in situ. The immature stages of many insects live in the soil; they may be sampled by sieving (if the insects are rather large), by soil-washing, or by a behavioral extraction method. Behavioral methods can only be used with mobile stages. There are several types of extractors, e.g., dry funnels, wet funnels, and hot- water extractors. The type used will depend on the insect being studied and the soil type. BIBLIOGRAPHY Aldrich, R. C., W. F. Bailey, and R. C. Heller. 1959. Larger scale 70 mm color photogra- phy techniques and equipment and their application to a forest sampling problem. Photogramm. Eng. 25:747-754.
INSECT SURVEYS 33 DeBach, P., Editor. 1964. Biological control of insect pests and weeds. Reinhold Publishing Corp., New York. 844 pp. Johnson, C. G., and L. R. Taylor. 1955. The development of large suction traps for airborne insects. Ann. Appl. Biol. 43:51-61. Morris, R. F. 1954. A sequential sampling technique for spruce budworm egg surveys. Can. J. Zool. 32:302-313. Morris, R. F. 1960. Sampling insect populations. Annu. Rev. Entomol. 5:243-264. Morris, R. F. 1963. Predictive population equations based on key factors. Mem. Entomol. Soc. Can. 32:61-62. Morris, R. F. 1964. The value of historical data in population research, with particular reference to Hyphantria cunea Drury. Can. Entomol. 96:356-368. Southwood, T. R. E. 1966. Ecological methods, with particular reference to the study of insect populations. Barnes & Noble, London. 410 pp. Strickland, A. H. 1961. Sampling crop pests and their hosts. Annu. Rev. Entomol. 6:201-220. Taylor, L. R. 1962<z. The absolute efficiency of insect suction traps. Ann. Appl. Biol. 50:405-421. Taylor, L. R. 19626. The efficiency of cylindrical sticky insect traps and suspended nets. Ann. Appl. Biol. 50:681-685. Uvarov, B. P. 1964. Problems of insect ecology in developing countries. J. Appl. Ecol. 1:159-168. Varley, G. C., and F. R. Gradwell. 1963. The interpretation of insect population changes. Proc. Ceylon Ass. Advan. Sci. 18(D): 142-156.
