Migration, the major movement behavior of insects, allows them to escape deteriorating habitats, to colonize new areas, or to seek temporary shelter such as overwintering sites. It involves a complex of traits that include development, physiology, morphology, and behavior, and it is a major component of the life histories of many species. These trait complexes or syndromes are adjusted by natural selection in complex ways that increase the fitness and therefore the success of migrants.
MIGRATION AND OTHER MOVEMENTS
The movements characteristic of organisms can be roughly divided into two broad categories, those that are immediately responsive to resources and those that are not (Box 1 ). Within the category of immediate responses to resources are two further broad types. The first type consists of “station-keeping” responses that serve to keep the organism on its territory or within the home range in which it carries out most of its life functions and spends most of
its time. Included within station-keeping movements are resource-sensitive behaviors crucial for survival. Examples are foraging, territorial behavior, and commuting, which is a periodic, often daily, round trip for resources. Foraging may be for any resource, including food, shelter, or mates; and commuting, which can also be considered to be a form of extended foraging, may involve travel over considerable distances. The commuting trails of leafcutter ants, for example, may extend for hundreds of meters both horizontally along the forest floor and vertically into the canopy. Foraging, commuting, and territorial behaviors are all readily responsive to resources: thus a female butterfly stops searching (foraging) upon encountering a host plant on which to oviposit, and a territorial forest drosophilid fly is bounded by the borders of its leaf display ground.
Box 1 Types of Insect Movement
Migration is a type of movement displayed by insects, but it differs from all other types because migratory insects (and other migrants as well) are unresponsive to suitable resources. There are two broad categories of movements:
I. Movements that are directed by resources and/or home range.
A. Station keeping: examples are foraging, commuting (periodic, usually daily, movements), and territorial behavior.
B. Ranging: movement to explore an area, often for a new home range or territory.
II. Movement not directly responsive to a resource or home range: here migration is undistracted movement with cessation primed (thresholds lowered) by the movement itself. Responses to resources are suppressed or suspended.
Ranging movements take organisms on exploratory journeys beyond the current home range and serve to locate a new home range or territory. Like station-keeping movements, ranging movement is a facultative response to resources, and like foraging or commuting, it ceases when a new resource (here in the form of previously unoccupied living space) is encountered. Ranging movements may also extend considerable distances but, like station keeping, still belong in the category of activities that are proximately resource sensitive.
Movement that is not immediately responsive to resources constitutes the distinct sort of behavior that is migration. Taking an organism beyond both its current home range and neighboring potential home ranges, migration is physiologically distinct from all other movements. It is so distinct because sensory inputs from resources that would ordinarily cause movement to cease do not stop migration. Thus, a characteristic of migration is that the organism undertaking it is undistracted by and fails to respond to food or mates, otherwise so necessary a part of life functions. Furthermore, migration is usually triggered by environmental cues, such as photoperiod, which forecast habitat change rather than being directly responsive to the change itself (usually a deterioration in the quality or availability of resources). Other characteristics of migration include distinct initiating and terminating behaviors. Many insects climb to the top of a bush or tree branch to take off on migratory flights, behavior they show at no other time. Sensory responses may also change, as in aphids that are sensitive to blue light from the sky during the take-off phase of migratory flight but become increasingly sensitive to yellow light, the characteristic wavelength of young host plants, as migration proceeds. Thus, migration is not defined by the distance traveled or by whether it is a “round trip.” Rather, it is defined in terms of the physiological and behavioral responses to resources; this behavior is true of all organisms, not just insects.
The movement behavior of individuals also has an outcome for the population of which those individuals are a part. This outcome involves displacement for a greater or lesser distance, but at either distance it involves removal from the home range. It can also result in the scattering or dispersal of individuals within the population; thus “dispersal” is a population phenomenon, not an individual movement. Movement can also result in congregation by mutual attraction or aggregation in a habitat. Both tendencies result in a decrease in the mean distance among individuals and contrast with dispersal, which increases mean distances. Note that all the movements just described can contribute to aggregation, congregation, or dispersal, depending on species and ecological circumstance.
Three examples of the sorts of population outcome attributable to migratory behavior occur in locusts, aphids, and moths of the genus Heliothis (and the very similar Helicoverpa). Locusts are a group of grasshopper species that under crowded conditions undergo remarkable behavioral and morphological changes known as phase transformation. In the desert locust, Schistocerca gregaria, perhaps the most extreme example of the phenomenon, crowded nymphs (hoppers) are strikingly black and yellow, whereas isolated individuals are pale brown or green. Crowded adults are larger and display differences in body proportions that readily distinguish them from their isolated counterparts. It is behavior, however, that most distinguishes the two forms. Isolated individuals display no mutual attraction, forage independently, and migrate at night. Crowded individuals show a high degree of mutual attraction and form large swarms that can number in the millions. When a swarm is feeding, locusts at the rear are constantly running out of food, overflying the body of the swarm, and landing at the leading edge. The result is a “rolling” movement across country in extended foraging. If a swarm enters an area with no forage, it may rise in unison and be carried for some distance by the wind. If this happens for a long period or repeatedly, the individuals in the swarm may switch to migratory behavior and cover considerable distances to descend again in regions of fresh plant growth. These aggregated swarms are major pests over much of Africa, occasionally reaching adjacent areas of the Middle East and in several notable instances the New World. The arrival of a swarm can mean that “not any green thing” (Exodus 10:15) is left for human consumption. It is the aggregation and migration that make the desert locust such a notorious pest. It would be much less a pest if its characteristic behavior led to dispersal rather than swarming.
Various species of aphid are also capable of spreading far and wide by migration. In Europe, an extensive monitoring network coordinated by English and French entomologists has mapped the seasonal spread of bean aphid, Aphis fabae. Large concentrations appear first in central France in early to midsummer. The species then spreads westward and northward over succeeding generations so that by late summer the aphid has reached high densities as far north as Scotland. In North America, monitoring of the corn leaf aphid, Rhopalosiphum maidis, has indicated the arrival of large numbers in the cornfields of Illinois. Analysis of weather systems suggests these aphids have come from as far away as Texas and were transported on wind streams. Studies of a number of other insects indicate that the Mississippi Valley is a major spring migration route for wind-transported insects to the agricultural regions of the upper Midwest.
Heliothis moths breed following rainfall in tropical and subtropical arid regions. If productivity is high on the new vegetation on which they feed, large populations of migrating moths are produced and are carried by winds to agricultural areas. In Australia, moths are transported in spring to wheat- and cotton-growing regions in New South Wales from breeding areas in interior regions of New South Wales and Queensland. The location of rainfall in the interior of Australia is unpredictable from year to year, and considerable effort has gone into locating areas in which rain has fallen, determining whether this precipitation is sufficient to produce large moth populations, and forecasting the arrival of migrating moths in conjunction with weather systems so that necessary control measures can be undertaken and unnecessary ones avoided. In the spring in North America, there is similar breeding of Heliothis moths in northern Mexico and southern Texas and migration northward on winds.
HISTORICAL BACKGROUND
In the 1930s C. B. Williams collected and summarized the available information on insect migration. The two topics that resulted were largely responsible for bringing to the attention of entomologists and other biologists the fact that the phenomenon was a common one. Williams focused on large insects such as butterflies and dragonflies, and he adopted the prevailing notion, derived largely from birds, that only round-trip movement could be called migratory. The way entomologists now think about insect migration is primarily the result of the work of four Britons: C. G. Johnson, J. S. Kennedy, T. R. E. Southwood, and L. R. Taylor, beginning around 1960. Johnson and Kennedy stressed that insect movements vary with respect to physiology and function, and their ideas revamped notions concerning the behavioral and life history aspects of migration. Southwood showed that the type of habitat determines the likelihood of migration among insects, and Taylor noted the importance of movement to the dynamics of populations in both time and space. Combined, the work of all four made explicit that migration is a distinct behavior with consequences for populations.
The distinct nature of migratory behavior was precisely outlined by Kennedy in his studies of the flight of the summer parthenoge-netic females of A. fabae. He used a flight chamber that allowed him to analyze the responses of free-flying aphids (Box 2). Key aspects of migratory flight that distinguished it from other types of flight were revealed by the flight chamber experiments. The aphids tested were the winged or alate form produced under crowded conditions. The uncrowded wingless females larviposit (bear live young) as soon as they make contact with a suitable host leaf. In contrast, the winged migrants do not larviposit until they have completed at least some flight. Furthermore, landing responses are primed by migratory flight: the longer the flight, the lower the threshold for landing. Finally, there is reciprocal interaction between flight and settling, since settling responses (i.e., probing a leaf to test its suitability and subsequent moving to the underside of the leaf to larviposit) can prime flight if they fail to be completed by attaching via the mouth-parts and producing offspring. This flight after incomplete settling may actually be stronger than that occurring at the beginning of migration. Migration thus is qualitatively different from other movement because station-keeping responses such as landing and probing (foraging) are inhibited by flight, but flight also primes them and promotes their later recurrence. Based on the behavior of migrating aphids, Kennedy provided a complete predictive definition of migration, as follows: migratory behavior is persistent and straightened-out movement effected by the animal’s own locomotory exertions or by its active embarkation upon a vehicle; it depends on some temporary inhibition of station-keeping responses, but promotes their eventual disinhibition and recurrence.
Box 2 The Kennedy Flight Chamber
J. S. Kennedy used this device in studies of insect migration. His experiments analyzed the performance of free-flying aphids and their landing and foraging responses. The lever arm can be twisted to shake the aphid off the platform, forcing it to fly, and it can be rotated out of the light and presented to the flying aphid again at will. Host plant leaves of different ages and leaves of different species of plant can be presented to permit investigators to observe variation in landing responses. Free flight is maintained by wind from the top of the chamber, and the wind speed is varied with the butterfly valve, whose setting thus indicates the strength of flight as measured by rate of climb which is balanced by the downward wind speed. [Figure reproduced from Dingle, H. (1996). “Migration: The Biology of Life on the Move.” Used by permission of Oxford University Press, Inc.l.
By explicitly focusing on the interaction between growth and reproductive behaviors (station keeping) and migratory behavior, Kennedy put migration firmly in the context of life histories. This context was also emphasized by Johnson, who identified the “oogenesis-flight syndrome” as characteristic of insect migration. Johnson noted that in a high proportion of migratory insects, especially in females, flight is limited to individuals with immature reproductive systems. It thus seemed that migration was based on an interaction between flight and the maturation of reproduction. Implicit was the assumption that migration and reproduction were alternative physiological states, with trade-offs in the mobilization of energy and materials. Johnson further postulated that this life history syndrome would be mediated by juvenile hormone (JH), a postulate now amply demonstrated (see later: Migratory Syndromes).
The population dynamical aspects of insect migration were assessed by Southwood and Taylor. Southwood placed migration into an ecological and evolutionary context by summarizing evidence that migration is characteristic of insects living in temporary habitats such as seasonal pools or early successional fields. This condition is in contrast to that of insects with more permanent habitats such as forests or large lakes; such insects are nonmigratory and often even wingless. This pattern of migration as a response to transitory environments was later formalized by Southwood in 1977 in the ratio H/t. with H the duration of the habitat and t the generation time of the insect. The frequency of migration in populations or species increases as the ratio shrinks toward unity, as later nicely demonstrated, for example, in leafhoppers (Homoptera) by Denno and colleagues. Taylor stressed the role of migratory behavior in the spatial dynamics of insect migration. He noted that migrants can disperse or coalesce, depending on whether individuals attract or repell one another and on atmospheric dynamics. This behavior can create a mosaic of insect densities over the landscape. He also initiated the Rothamsted Insect Survey, an array of traps to sample insects in the air, from which data on numbers were taken and analyzed at the Rothamsted Experiment Station near London. This network allowed tracking and forecasting of insect pests such as aphids with major practical implications for insect control.
THE CAUSES OF MIGRATION
Since Southwood’s original statement of the relationship, it has become apparent that the impermanence of habitats is indeed the primary selective force driving insect migration. Much of this imperma-nence is a function of season, and as with other well-known migrants such as many fish or birds, seasonality is a common factor in insect migrations. Most seasonal migrations are on a relatively small scale, with distances traveled only a few hundreds or thousands of meters, but others cover much greater distances. Examples of short-distance migrations to overwintering diapause sites include the Colorado potato beetle, Leptinotarsa decemlineata. several species of common seed-feeding hemipterans, including Lygaeus kalmii, a milkweed bug of Europe (in Sweden it often flies to lighthouses to diapause), and several species of lady beetles (Coccinellidae). In some lady beetles migration extends to several kilometers. In California, Hippodamia convergens, the convergent lady beetle, overwinters at high altitudes in the Sierra Nevada and migrates to agricultural areas in the Central Valley in March. Beginning in June, offspring of the early spring migrants fly back to intermediate altitudes and form aggregations. There is then a later movement to higher altitudes to overwinter so that the migration to overwintering sites is a two-step process.
Other insect migrants make spectacularly long journeys. The best known of these is that made by the eastern North American population of the monarch butterfly, Danaus plexipypus, studied extensively by Lincoln Brower and F. A. Urquhart. This butterfly cannot overwinter in the temperate zone, so it must migrate to southern overwintering sites. The short days of autumn cause adult butterflies to enter reproductive diapause, and they undertake a southward journey of 3000 km or more. The majority of the eastern population overwinters en masse in a very few high-altitude protected sites in the Transvolcanic Range of central Mexico, where they arrive in the late autumn. Beginning in
February, the aggregations break up, mating occurs, and the butterflies begin to move northward. Identification of the chemical cardeno-lide “fingerprints” of the milkweeds eaten by monarchs when they are caterpillars and stored in the adults has revealed that the overwintering generation stops and breeds on the spring flush of milkweeds along the coastal plain of the Gulf of Mexico. It is the offspring of these individuals that invade regions farther north beginning in late May and early June. Thus, as with fall California convergent lady beetles, the spring migration is at least a two-step process. A very similar migration pattern occurs in the same region in the large milkweed bug, Oncopeltus fasciatus, and it, too, occurs in two stages in the spring.
The migration of western populations of the monarch is more complicated. These populations overwinter in aggregations along the coast of southern California, where winter climatic conditions are similar to the aggregation sites in Mexico. When the aggregations begin to break up, as early as late January, the butterflies move to early sprouting milkweeds in the Coast Ranges and breed there. The next generation moves both to more coastal milkweeds and to milkweeds that grow farther inland as far east as the Rocky Mountains and as far north as the Canadian border, so that as in the eastern populations, the spring migration occurs in two stages over two generations. A very similar pattern occurs in the monarch population introduced into eastern Australia, with overwintering near the coast, a migration inland in the spring, and a return to the coast in the autumn. In more northern parts of the Australian range, there may be year-round breeding in coastal and subcoastal regions.
One way to assess the influence of ephemeral habitats on the evolution of insect migration is to survey across species and populations occurring in different kinds of habitats. In Europe, a number of species and populations of water striders (Hemiptera: Gerridae) occur over an array of habitats, from small, temporary ponds to large lakes and permanent streams to isolated permanent bogs. Species in the more temporary bodies of water have wings and undertake regular migrations to locate their aquatic habitats as they appear and disappear in the landscape. At the opposite extreme in permanent lakes and bogs, there are species that are wingless. Across habitats with varying degrees of permanence are populations and species of water striders with varying proportions of winged and wingless individuals determined primarily environmentally (polyphenisms) where habitat change is predictable, and primarily genetically (polymorphisms) where change is increasingly random with respect to the life cycle.
A second example of the influence of habitat duration on the occurrence of migrants within a fauna occurs in Australian butterflies. Often, latitude predicts the amount of migration that will occur, especially where there is adequate rainfall and seasonal change is largely a function of temperature. In eastern North America, for example, 98% of the variance in the proportion of migratory birds is explained by latitude, with a higher proportion of migrants at northern latitudes. Similarly in eastern Australia, where the climate is warmer overall but still temperate with adequate rainfall, 72% of the variance in the proportion of butterfly migrants is explained by latitude. The situation is quite different in the dry regions of Australia west of the Great Dividing Range. Here latitude accounts for less than 1% of the variance in proportion of butterfly migrants, and climate variables that indicate rainfall patterns, such as soil moisture, which accounts for about 50% of the variance, are much better indicators of migration. The amount of rainfall is not correlated with latitude, and so latitude does not predict migration. In this dry climate it is the availability of erratic rainfall that counts, and only migrants that can take advantage of the ephemeral flushes of vegetation that follow such rainfall. Thus, as with Heliothis moths, migration allows some butterflies to exploit a dry and ephemeral habitat.
MIGRATORY SYNDROMES
Accompanying migratory behavior is a syndrome of traits that act in coordination to increase fitness. These traits vary from enzymes to life history characters and, being influenced by subsets of the same genes, are genetically correlated. At the physiological level, insects (like most other migrant organisms) use fat as fuel, primarily for two reasons. First, fat metabolizes to produce about twice as much energy as carbohydrate or protein; and second, it requires no water for storage (in contrast, storage of 1 g of carbohydrate requires 3 g of water). Insects such as the monarch butterfly and the large milkweed bug shift lipids from yolk formation to fat storage under the influence of the shorter photoperiods of autumn and just prior to migration. The flight muscles of migrants are also adapted to the energetic demands of lengthy flight. Enzymes active in oxidative metabolism, such as citrate synthase, and in fatty acid oxidation, such as ( -hydroxyl coenzyme A dehydrogenase, or HOAD, tend to show higher levels of activity in the flight muscle of migrants compared with that shown in nonmigrants. This difference is most apparent where there are wing polymorphisms and migrants have longer wings.
The most important hormone involved with insect migration is JH. It has influence not only on the coordination of the various relationships of the oogenesis-flight syndrome but also has direct effects on migratory flight. In many insect migrants such as the monarch butterfly, short photoperiods result in reduced outputs of JH from the corpus allatum. This reduction in JH output in turn leads to a reduction in ovarian and egg development, which is then accompanied by migratory flight. In several species of migrant insects, prolonging of the prereproductive period by reduced JH titers results in the triggering and maintenance of migratory flight. At the same time it has been demonstrated in several migratory species (such as the large milkweed bug, the convergent lady beetle, and the monarch butterfly) that JH directly stimulates migration. Implants of corpora allata, the source of JH, or topical application of JH or some of its chemical analogues, are effective in increasing flight in migrants. In some insects such as the monarch, adipokinetic hormone (AKH— involved in promoting fat metabolism) also stimulates additional flight. The effects of JH and AKH in the monarch are illustrated in Fig. 1 .
FIGURE 1 Influence of JH and AKH on tethered flight in the monarch butterfly. Flight (%) longer than 30min is the index of migratory flight. The butterflies received topical application of hormones or the acetone control, and subsequent flight duration was determined. Both hormones, singly or together, increased flight over controls.
Because migratory flight occurs when reproduction is delayed by reduced JH concentrations, it is logical to inquire what level of JH determines migration. This question was answered for the large milkweed bug by M. A. Rankin. She selected for delayed onset of flight, which also resulted in delayed reproduction. Rankin measured JH titers in the blood during the prereproductive period and showed that JH titers were low when there was no flight or reproduction; intermediate titers stimulated flight, and high titers stimulated oog-enesis. Thus, if titers rose only to intermediate levels, as might occur under short days, for example, flight but not reproduction would be triggered. These JH titers may also be regulated by JH esterase, the enzyme that breaks down JH. In wing-polymorphic crickets, high concentrations of JH result in short-winged individuals. Artificial selection experiments that increased the frequency of long wings also resulted in increased amounts of JH esterase in the blood and so reduced amounts of JH. Selection also demonstrated that it was possible to change both mean and threshold JH esterase activity. The possible role of JH esterase in fully winged migratory insects remains to be studied.
An additional behavioral aspect of the migratory flights of many insects is the ability to maintain a more or less constant direction during migration. Mostly, this directionality has been studied in butterflies, although some other large insects such as dragonflies and larger Hymenoptera also seem to maintain a constant direction when migrating. The monarch butterfly in eastern North America flies in a steady southward or southwestward direction in the autumn, flight directions that lead to the overwintering sites in central Mexico. In the spring, the migratory flight is to the north. Compilations of observations of several species of Australian butterflies, including the monarch, reveal that the insects fly south or southwest in the spring and north or northeast in the autumn. The apparently migratory flights of a few species occur in the same direction no matter what the season, a phenomenon that has yet to be explained. In Europe migratory flights of the butterfly Pieris brassicae are consistent with both season and geography. Autumn migrants from northern Germany fly south or south by southeast, whereas migrants in the south of France fly to the southwest, which takes them to Spain rather than over the Mediterranean. Further experiments have demonstrated that butterflies that have diapaused, as they do during the winter, fly north when they migrate, but those emerging from nondiapause (summer) pupae fly toward the south. Seasonal winds also frequently carry migrating insects in the “correct” direction. Monarch butterflies in eastern North America frequently soar and are carried southward by northerly winds, and simulations of the migration of the large milkweed bug from the same region indicate that a portion of the population reaches southern overwintering areas regardless of whether they orient.
Where the mechanism of orientation has been studied, the evidence suggests that it is a time-compensated sun compass. To use the sun effectively for orientation, organisms must be able to compensate for its daily passage across the sky by reference to a “biological clock.” To demonstrate that an organism is using a time-compensated sun compass, it is necessary to clock-shift it by maintaining it in a daily light cycle that is out of phase with the ambient cycle and then to show that its orientation is displaced by an amount consistent with the clock-shift. A displacement of 6 h, for example, should lead to a directional change of 90° ; whether the change is plus or minus depends on the direction of the clock-shift. Experiments with southward-moving monarch butterflies suggest that this pattern is indeed followed. In tropical Panama, two species of migrating pierid butterflies, Aphrissa statira and Phoebis argante, regularly maintain
a directional flight across Gatun Lake in the Panama Canal. Clock-shift experiments resulted in changes in the direction of orientation that were consistent with a sun compass, even though there was also a component imposed by wind drift. The sophisticated orientation mechanisms of honey bees, ants, and certain other insects incorporating a sun compass, imply that orientation is probably widespread. The future will undoubtedly reveal the presence of a sun compass in other migrants, as well as the presence of other mechanisms, especially in nocturnal migrants. Radar observations indicate that the passage of many species of nocturnal migrants is specific to winds of a certain direction, but the means by which this preference is enforced are unknown.
In addition to behavioral and physiological characters, migration syndromes often include life history traits such as the age at first reproduction and fecundity, particularly in many wing-polymorphic insects. Typically, in these species the short-winged or wingless forms reproduce earlier and display higher fecundities than their long-winged counterparts. This dichotomy is at least in part because of trade-offs between flight and reproduction. The metabolically active flight muscles that accompany long wings and migration are costly to maintain, requiring considerably more maintenance energy than the thoracic musculature of wingless or short-winged individuals. In contrast, the later reproducing individuals, with lower egg production, are often longer lived.
Migration syndromes that include life history traits are the result of underlying genetic mechanisms, as revealed in artificial experiments using the large milkweed bug. Like all flying insects, this migrant can be induced to fly by removing substrate contact. Bugs that are glued at the prothorax to a tether will fly if contact with the tarsi is removed. An individual in the migratory state can fly on the tether for several hours, and the duration of flight can be used as an index of migration. Artificial selection can be used to increase the proportion of individuals making long (or short) flights, with the duration of flights also affected. Selection was used to both increase and decrease the proportion of bugs undertaking long flights. In addition to flight, wing length and fecundity responded to this selective regime. The bugs of the line with a higher proportion of long flights also had longer wings on average, and the females of this line produced more eggs during the first 5 days of reproductive life. This means that the genes influencing flight also influenced wing length and fecundity, most likely via pleiotropic effects. Longer term selection experiments on wing length, which also resulted in higher fecundities and increased flight as wing length increased, suggested that linkage disequilibrium is unlikely. Parallel selection experiments on a population that did not migrate failed to reveal genetic correlations among wing length, flight, and fecundity, indicating that the genetically based syndrome of correlations among these traits is unique to migratory populations. The selection experiments reveal that natural selection has produced an adaptive migratory syndrome that includes wing length and fecundity. Interestingly, the age at first reproduction is unaffected by selection.
The conclusion from the brief survey of insect migration is that this behavior involves more than simply extended movement to escape to a new habitat. Rather, migration is a trait of considerable complexity, requiring knowledge of behavior, development, ecology, physiology, and genetics to provide a full understanding of its evolution and function.
