The Relation Between Age and Flight Activity in the Milkweed Bug, Oncopeltus

Recent reviews by Johnson (1960), Kennedy (1961) and Southwood (1962) have represented a change in outlook toward insect migration. Previously emphasis had been placed on mass movements and orientation of flights; the subject from thisstandpoint has been reviewed by Williams (1957 and 1958) who was largely responsible for its development and for demonstrating that migration was a widespread phenomenon among insects. Williams defined migration as ‘a continued movement in a more or less definite direction, in which both movement and direction are under the control of the animal concerned’ (1958), but the difficulty with this definition is that such well-known migrants as aphids and locusts are not included. Migration flights in these insects may or may not be in a definite direction and are not under direct control since both use the wind as a transporting vehicle (Kennedy, 1961). Further, return movements of the same individuals, long considered a fundamental aspect of migrations (e.g. Allee et al. 1949) now must be considered dubious for any insect. But rather than abandon the term ‘migration’ altogether when considering insect movements, it seemed more reasonable to attempt a reassessment based on other than topographical characteristics.

The new approach crystallized by Johnson, Kennedy and Southwood considers migration in insects a distinct behavioural and physiological syndrome essentially independent of specific directional orientation. Kennedy (1961) defines migration as ‘persistent, straightened out movement that is accompanied by and dependent upon the maintenance of an internal inhibition of those ‘vegetative’ reflexes that will, eventually, arrest movement’. The straightened-out movement, which need have no directionally adaptive significance, is a result of enhanced sensorimotor functions, and the ‘vegetative’ reflexes are those associated with growth, e.g. feeding and reproduction.

The distinction is no longer made between active migration and passive dispersal by wind since large active fliers such as locusts are as much wind-horne as such small weak fliers as aphids (Kennedy, 1951, 1961). To become wind-borne any insect must actively rise above what Taylor (1958) has called the ‘boundary layer’, the altitude, different for each species, above which it no longer fully controls its own orientation. A further characteristic of migrations is that they usually occur pre-reproductively. The major difference between migratory and non-migratory flights is one of relative thresholds, with no sharp line of demarcation. Acceptance of Kennedy’s definition opens the way to the experimental analysis of the behavioural, physiological, and ecological factors initiating and controlling migration (Johnson, 1963 ; Southwood, 1962). The present paper concerns an attempt to study the above factors in the laboratory.

For such a study I needed an insect which would show a peak of flight activity, obtained conveniently using tethered flight, which might be considered migratory. This flight would presumably be a function of age and ideally would be characterized by rather narrow limits. The relation of flight activity to age could then be compared with its relation to other variables important in connexion with migration, such as physiological development, reproductive activity, and population parameters. The insects considered most likely to exhibit a distinct flight peak were the Hemiptera and Coleoptera; these forms usually move about by walking and fly only occasionally, and in these circumstances a long flight could well be migratory (Southwood & Johnson, 1957; Southwood, 1962). Diptera which have been popular insects for studies of tethered flight usually move about by flying anyway, thus introducing some ambiguity.

The insect chosen for the present study was the ‘large milkweed bug’ Oncopeltus fasciatus (Dallas) (Heteroptera: Lygaeidae). Oncopeltus has not been listed as a migrant, but certain considerations indicate that it probably is one. Southwood (1962) has summarized a considerable amount of information concerning the relation between the occurrence of migration in various terrestrial arthropods and the habitats in which they are found. He gives extensive evidence that those frequenting ‘temporary’ habitats such as annual and perennial plants of field and wasteland are more apt to be migrants than those of ‘permanent’ habitats such as woodland trees. On the criterion of habitat, since it occurs on milkweed (Asclepias sp.) which is a perennial of open fields and roadsides, one would predict that Oncopeltus would be a migrant. Some field observations of mine (below) give more direct evidence that it does in fact migrate in the field. Other Lygaeidae closely similar to Oncopeltus are known migrants (Southwood, i960).

The insects used in the present study were all descended from a culture of Oncopeltus maintained in the Sub-Department of Entomology at the University of Cambridge. They were raised at 23 ± 1° C. at approximately 65 % relative humidity on a daily schedule of 16 hr. of light and 8 hr. of darkness and were fed on dry milkweed seeds every other day and given water from soaked cotton wool. Dead individuals were removed daily, and if marked, their age at death in days after the final moult was recorded. Daily checks were also made for newly emerged adults which were then marked on the prothorax with spots of quick-drying paint. At every observation of a particular culture all individuals in the act of copulating or egg-laying were noted. Loose cotton wool was supplied for egg-laying and proved satisfactory.

For flight-testing, bugs were attached at the pronotum to a short stick by means of a small amount of paraffin wax. Once flying, they were clamped into a ring stand about 30 cm. above a black substrate with a 100 W. light bulb about 50 cm. above them. The durations of the first five successive flights were recorded and summed to give the total durations whose means are plotted in Fig. 1. If a bug became flight-refractory after fewer than five flights, the total duration was simply computed from the number of flights obtained. The experimental room was black walled and maintained at a temperature of 25 ±1° C.

Flight was initiated in a bug by subjecting it to a standard sequence of procedures given one after the other until it flew. Each step was numbered to give a relative measure of threshold at which flight began (Table 2). The procedures and threshold values were as follows:

1. Bug lifted off substrate by stick attached to pronotum.

2. Stick clamped in ring stand and tapped firmly with another stick.

3. Bug given tarsal contact ; this suddenly removed and at the same time a jet of air directed at head.

4. Stick to which bug attached held by experimenter and bug moved rapidly back and forth at arm’s length five times.

5. Bug allowed to walk 25 cm. then rapidly lifted off substrate and repeat (4).

6. Wings opened by experimenter (if still unopened) and repeat (5).

7. Repeat (5).

8. Repeat (5).

9. Bug opens wings to any of above, but never flies.

10. Bug never opens wings without aid of experimenter.

Thus if a bug failed to fly when subjected to (1), (2), or (3) above, but did so the first time it was moved back and forth at arm’s length, it would be assigned a flight threshold of 4.

Mean flight-duration as a function of age in days after the final moult is shown plotted in Fig. 1. The sizes and ranges of the samples from which this figure was drawn are given in Table 1. In both sexes a peak period of maximum length of flight occurs from 8 to 10 days, and in males a second peak occurs at 30-35 days. The peaks seem to represent real changes in flight activity, for if the values at 8, 9, 10 and 11 days are compared with either the preceding (days 3, 6 and 7) or succeeding (days 20 and 25) low flight values respectively, the differences are found to be significant statistically at the 0·02 level for a two-tailed Mann-Whitney U test (Siegel, 1956). For males the figures at 30 and 35 days are likewise different from those at 20 and 25 or 40 and 45 days, this time at the 0·002 level.

The occurrence of a peak of mean flight-duration at 8-10 days is of special interest when the flight data are compared with data for cuticle growth in Oncopeltus. Recently, Neville (1963) has discovered that insect endocuticle is, in the adult, formed in paired light and dark layers with one pair deposited daily until the insect is full grown; this daily deposition occurs in many of the insects so far examined including HemipteraHeteroptera (Neville, personal communication). Frozen sections through the cuticle of Oncopeltus adults of different ages indicate daily deposition of paired zones with cessation of such growth at 9 or 10 days. Fig. 2 is taken from a frozen section through the hind tibia of an adult more than 30 days old. It is not possible to distinguish whether there are nine or ten paired zones either in the figure or in the original section, but since there are no more than nine or ten in a bug more than 30 days old, the insect evidently attains full growth at 9-10 days. Since the peak flight-duration occurs at the same time (Fig. 1), there is thus a remarkable correlation between the growth and behavioural parameters. Long flights evidently occur when an adult insect reaches full maturity at least with regard to the formation of the endocuticle.

In addition to the usual flight experiments which were run on bugs coming from cultures including both sexes, 15 virgin males at 20 days and 12 virgin females at 30 days also were tested. The respective ages were chosen because they represented low points for the two sexes in the curves given in Fig. 1. These experimentals were kept separate from members of the opposite sex as soon as the final moult was complete. The 30-day virgin females showed no significant difference in performance from their mated counterparts, but the 20-day virgin males were strikingly different (Fig. 1). In fact 20-day virgin males flew longer on the average than males at any other age, and the difference between them and 20-day normal males was statistically significant at the o·oo2 level. The flight threshold was also generally lower in the virgins (Table 2). Reproductive activity in males thus seems to have a decided influence on the time course of the tendency to fly and to make long flights.

As each bug was tested the thresholds of each of its flights, in arbitrary units as described above, were recorded. The thresholds of the initial flights for each age are given in Table 2. There is a general decline of flight threshold in both sexes as bugs become older with no indication that long-flying bugs showed thresholds lower than those of others of the same age. Thresholds at 8-10 days, when flights were of maximum duration, were thus higher than for older individuals, which showed consistently IQW thresholds even when there were no long flights. From the data for threshold and duration there thus seem to be three kinds of flights: (i) long-duration, high-threshold occurring in both sexes at 8-10 days; (ii) short-duration, low-threshold occurring in older individuals; and (iii) long-duration, low-threshold occurring only in 30 to 35-day males.

In respect of long-duration flights there was a further difference between the sexes. In males if a long flight of more than 30 min. occurred in the five test flights, it was usually preceded by one or more short flights of less than 1 min. This was not true of females. In the latter, long flights when they occurred were more frequently the initial flight of a series. These differences are shown in Table 3 and are statistically significant (P < 0·05 using χ²). For both sexes, in those animals in which long flights occurred, the initial flight if not of less than 1 min. was greater than io min. with only two exceptions; these were males with first flights of about 3 and 8 min. respectively. The pattern of activity associated with long flights is thus different in the two sexes.

The pattern of a long flight itself was, however, generally similar in both. At onset and shortly thereafter flight was unsteady with changes in wing-beat frequency and amplitude and often oscillations of the abdomen presumably caused mechanically by the uneven wing-beats. After this initial period the wings settled down to a steady frequency and amplitude with few changes; this ‘cruising’ phase lasted until shortly before the end of flight when another period of unsteadiness combined with a general decrease in frequency and amplitude ensued. One could usually predict from the pattern of the wingbeats when a bug was about to cease flight.

In spite of the general tendency to long flights at 8-10 days, some individuals never flew. Some of these were tested repeatedly in the event of late development of flight readiness, but no flight could be elicited. As a check on the general health of these individuals they were followed throughout the remainder of life along with those which did exhibit long flights. The ‘flightless’ bugs were no different with respect of length of life, general reproductive activity, and in the case of females, frequency of egglaying. There is thus the possibility of a behavioural polymorphism with respect to flight activity.

In order to determine the relation, if any, between flight-duration and reproduction, three criteria of reproductive activity were recorded: copulation by males and by females and egg-laying by females, all as a function of age. Two of these are shown plotted in Fig. 3. Egg-laying by females is first observed at 15 days after becoming adult, rises to a maximum at 22-27 days, and then declines to a last recorded observation at 55 days. As one would expect, copulatory activity in males starts somewhat earlier, but otherwise shows much the same pattern except for a few males copulating at more than 55 days. The copulatory activity of females gave essentially the same picture.

By comparing Fig. 3 with Fig. 1 one can see that the decline of mean flight-duration in females and the initial decline in the males coincides with the rise in reproductive activity. The second flight-duration peak of males occurs after reproductive activity has begun to decline. At the peak of reproduction, 22-27 days, the flight-duration curve for both sexes is at or near its minimum. As a result of the foregoing, therefore, the 8-to 10-day flight-duration peak for both sexes can be considered pre-reproductive. The relation of the male second peak to reproduction is less clear. Also, as described above, males can be excited to much greater flight activity by depriving them of sexual partners.

Whereas physiologically one can meaningfully speak of the flight-duration peak as pre-reproductive, ecologically it has greater meaning in terms of the ‘reproductive value’ of an individual. This concept has been discussed by Fisher (1958) and Slobodkin (1962) in terms of the worth of the individual to the future growth of the population. Slobodkin defines it as ‘the diminution of future population increase produced by removing a single animal of a given age from a population..Reproductive value is a measure of sensitivity to natural selection; it follows logically that peak reproductive value, since it represents maximum contribution to future population increase, also indicates that period of the life history at which natural selection exerts its greatest influence.

The reproductive values (V_x/V₀) for females, plotted in Fig. 4, were determined by substitution in the equation

where V_x is the reproductive value at age x, V_o is this value at age o( = 1), e is the base of natural logarithms, x is age in days, l_x is the proportion of those alive at age o alive at age x, m_x is the birth-rate expressed as the number of female eggs laid per female of age x (i.e. during the interval to ⁠), and r is an expression of the growth potential of a population usually referred to in the ecological literature as ‘the intrinsic rate of (natural) increase’ and is an instantaneous rate (Slobodkin, 1962). The value of r is calculated by trial and error substitution in the equation

a simple and rapid method of performing this calculation is outlined by Birch (1948). All the above concepts are discussed by Slobodkin (1962).

The values for l_x came directly from the daily records of adult mortality. It was estimated that 75 % of the eggs laid survived to become adult bugs ; mortality among eggs and nymphs could not be calculated precisely because of insufficient data, but the data available indicated that it was certainly no more than 25 %. The final moult took place at 42 days on the average so all adults were taken to be of that age at the completion of moulting. I did not have a direct estimate of m_x but did record the number of females at various ages observed laying eggs (Fig. 4). Johansson (1958) gives values of eggs per female per week from virgin females; these gave an estimate of the number of eggs laid by a female during her reproductive life. The curve for egg production given by Johansson shows a peak later than the peak of egg-laying activity observed by me; but his females were virgin, and my observations confirmed that egg-laying is delayed in virgin females. Johansson also maintained his animals on less than 16 hr. of light. By assuming that maximum egg-laying activity as expressed by number of females laying at age x (Fig. 3) reflected maximum number of eggs laid per female, estimates of m_x for my animals were obtained. These estimates were then halved on the assumption of a 1:1 sex ratio and used in calculating reproductive values for females. The value of r obtained was 0·0806 per day which is similar to estimates of r for other insects of similar generation time all of which approximate to o·1 per day (Evans & Smith, 1952).

The reproductive values calculated for the Oncopeltus females in this study are shown plotted as a function of age in Fig. 4. The peak value occurs 19-5 days after completion of the final moult. The flight peak in both sexes, 8-ro days, occurs prior to the time of maximum reproduction, but at a period when the reproductive value is still better than 75 % of maximum. Since the long flights do occur during a period of high reproductive values, they must have been subject to strong favourable selection. That such flights are still not obligatory, however, is indicated by the fact that some bugs never displayed them.

Several plots of reproductive value were calculated using somewhat different assumptions from those made in calculating Fig. 4. For instance, it was assumed that there was no mortality among eggs and nymphs, that all adults were age o at the completion of the final moult, and that values of r were higher or lower. The curves obtained were of slightly altered shape and the actual values were, of course, not the same as in Fig. 4, but the peak reproductive value always occurred at 19-21 days. Assumptions concerning birth rate remained the same.

For comparison with those calculated for Oncopeltus, reproductive values were also calculated for the alate alienicolae of Aphis fabae. The data for these calculations were taken from Cockbain (1961) using the values given for flown insects. The final moult occurs 4-5 days after birth (Kennedy & Booth, 1950), and the mortality curve of Cockbain was extrapolated back 5 days, on the assumption of a 5-day larval period, to reach an estimate of larval mortality which was thus taken to be 5 %. The resultant curve of reproductive value is shown in Fig. 5. The migration flight of A. fabae usually occurs within the first 24 hr. after the final moult (Cockbain, 1961) which means that it virtually coincides with peak reproductive value. This close correlation is a reflexion of the burst of larval deposition immediately following a flight and indicates that the aphid is maximally sensitive to selection at time of flight. A. fabae is thus relatively more sensitive to selection at its flight peak than is Oncopeltus.

The question that must first be asked with regard to the earlier flight-duration peak of Oncopeltus is: does this peak in fact represent migration? I believe that it does for the following reasons. First, the long flights occurring during that period are of a much greater order of magnitude than flights which usually occur. The latter are of less than 1o min. duration, whereas the former may last over an hour ; a tenfold increase at the peak would not be an unreasonable estimate of the difference. The long flights certainly represent persistent movement, one of Kennedy’s criteria for migration (see above). Secondly, the pattern of an individual prolonged flight is suggestive of migration. Initially it was uneven, but later settled into a ‘cruising’ phase; this pattern is similar to that observed by Kennedy & Booth (1963 a) in aphid migrants in a flight chamber. The cruising indicates that the flight, had it been in nature, would probably have been straightened out. Thirdly, the flight peak occurs pre-reproductively during a limited time in the life history (this is somewhat less clear for males because of the second flight peak, see below), and most migration flights areprereproductive (Johnson, 1963). Fourthly, Oncopeltus would be expected to be a migrant because it is a denizen of a ‘temporary habitat’ (Southwood, 1962). And, fifthly, I have some evidence of active dispersal, at least, in the field. Oncopeltus tends to congregate in groups of 30-50 individuals up to and including the fifth instar. About a week to 10 days after the final moult, however, the bugs disappear from the aggregate. Soon after the last fifth instars have moulted, there are no insects left at the site.

If the flight-duration peak in Oncopeltus does represent migration, then the present results support Johnson’s (1960, 1963) thesis that migration in insects is a postteneral phenomenon. The method of counting cuticular growth rings discovered by Neville (1963) provides a precise, objective, morphological determination of the teneral period; the major flight effort of Oncopeltus, at 8-10 days after the final moult, occurs just at the time when deposition of cuticular layers ceases. Kennedy (1951) has already noted that migration of locusts develops as the cuticle hardens; Taylor (1957) discusses the teneral period in aphids defining it mainly by behavioural criteria, while Bursell (1961) correlates this period with feeding and muscular growth in the tsetse fly (Glossind). In both aphids and tsetse flight activity increases in post-teneral or ‘non-teneral ‘adults. In all these cases, however, exact means of identifying teñerais are lacking with the result that some ambiguity is introduced. This ambiguity is eliminated by the use of daily growth rings, and for most insects post-teneral migration must mean migration occurring shortly after the cessation of cuticular deposition. It is also usually pre-reproductive.

The data for flight thresholds in Oncopeltus (Table 2) at first seem paradoxical, for the lowest thresholds do not occur at the time of peak flight-duration. In terms of the bugs’ behaviour in the field, however, the fact that flight threshold at presumed migration is at least moderately high insures that flight will take place only under favourable conditions—on a warm sunny day, for example, instead of a cold rainy one. Mortality which might result from a long flight under adverse weather conditions would thus be avoided. The fact that many more Heteroptera migrate during the day than at night (Southwood, 1960) may also be a reflexion of relatively high flight-thresholds at the time of migration. Southwood found during extensive trapping operations that Lygaeidae were restricted to diurnal catches. As for the short-duration, low-threshold flights which are common to both sexes, they may be considered to represent trivial flights or ‘flits’ (Southwood, 1960, 1962) which in the field are short flights to change feeding site or to search for a sexual partner.

The long-duration, low-threshold flights which occur only in older males do not fit as well the concept of a post-teneral migration flight. They are not pre-reproductive, although some migrations are inter-reproductive (Kennedy, 1961 ; Schneider, 1962), and these may fall into that category. In. fact it is probably most reasonable to suggest that there are, in males, two types of migration. The first is the ‘typical’ post-teneral, pre-reproductive flight which is characteristic also of females and which occurs when threshold is relatively high. The second type occurring in older individuals, since its threshold is low, may be an ‘explosion’ from trivial flitting activity into migration proper. The general pattern of flight in males in which long flights are apt to be preceded by short flights (Table 3) is further suggestive of this. A possible trigger could be deprivation of responsive sexual partners. Virgin males (Fig. 1) exhibited low-threshold, long-duration flights at an age when mated males were short-flying. If the decline of sexual activity in older males (Fig. 3) is a reflexion of lack of female responsiveness rather than of male senility (experimental animals did come from cultures of similar aged individuals), then the long flights of old males are the product of events similar to those resulting in long flights of virgins. Making such a second migration flight increases the probability of finding females elsewhere with which to mate. Johnson (1963) has indicated that male migration is more variable inter-specifically, and for Oncopeltus this seems to be true intraspecifically as well. It appears that in Oncopeltus the male is the more mobile sex with the advantage of being able to adapt mobility to different environmental conditions.

Since the ‘typical’ long-duration flight, in males and females, occurs pre-reproductively, it is worthwhile to examine what this means ecologically as well as behaviourally and physiologically. Reproductive value, V_x/V₀, since it takes into account fecundity, mortality, and the time course of reproduction, is the most meaningful ecological parameter. Maximum reproductive value represents the time of maximum sensitivity to natural selection. The fact that the flight peak occurs at a time when reproductive value is high (Fig. 4) indicates that there is strong selection favouring it. The maximum V_x/V₀ is reached shortly after the flight peak, so at this time the bugs will have their greatest potential with respect to future population growth. One of the theories appearing often in the literature concerning the function of insect migration is that it is a means of alleviating population pressure or, in other words, that migrants are refugees. In view of the correlation between high reproductive value and long flights this explanation seems unlikely, for one would not expect refugees to have a high V_x/V₀. Settlers, on the other hand, would be expected to have high values, or they would have little worth as colonizers. Migrating Oncopeltus females, then, are presumably settlers, and by implication, so are males migrating at the same time. The foregoing line of reasoning supports the contention that migration is an evolved adaptation rather than a reaction to current adversity (Johnson, 1960). Further one would predict that migration flights in other insects would occur when reproductive value was high.

This prediction is borne out in the case of Aphis fabae (Fig. 5) in which peak flight and peak reproductive value occur virtually simultaneously. The migration flight is thus subject to maximum selection. A reflexion of this strong selection is the fact that the flight of alate alienicolae is in normal conditions a necessary prerequisite for settling and deposition of young (Kennedy, 1958; Kennedy & Booth, 1963b). The production of winged individuals is facultative and in this respect so is migration, but the winged forms themselves must normally fly. The relatively greater sensitivity to selection at the flight peak of A. fabae when compared with Oncopeltus is thus reflected in the tendency of virtually all individuals to fly. There is little variability in the occurrence of migration.

In Oncopeltus, on the other hand, flight peak and reproductive value do not coincide exactly, and the flight is not obligatory. Some of the individuals tested never flew, suggesting, since these were as long-lived and as reproductively active as fliers, a behavioural polymorphism which is undoubtedly of adaptive advantage. Bugs taking part in migration would assure the dispersal of the species to new situations while sedentary individuals would maintain a centre of dispersal in a currently suitable habitat. Johnson (1963) and Southwood (1962) have pointed out that factors which produce morphometric differences in migrants might also be expected to produce behavioural differences.

There are a few instances where other insects have been studied with respect to age and flight. Riegert (1962), for example, studied three North American grasshoppers and found that long flights in both sexes took place mostly between 7 and 20 days. Females of all three species flew longer than males indicating that, in contrast to Oncopeltus, they are the more mobile sex. In the Douglas fir beetle, Dendroctonus, on the other hand, the males seem the more active (Atkins, 1959). Both sexes of this wood-boring beetle flew prior to gallery-construction, but once construction began females ceased flight while some males still flew. After approximately 15 days in the galleries there was a second period during which a relatively small proportion of both sexes flew. Williams, Bamess & Sawyer (1943) studied the flight of Drosophila funebris females and correlated it with the amount of glycogen present in the body. Periods of high flight activity, as measured by duration and wing-beat frequency, and high glycogen concentration both occurred between 3 and 15 days. Various changes in biochemistry have also been correlated with flight maturation in certain other insects; the subject has been reviewed by Rockstein (1957). In Oncopeltus the flight peak is correlated with the completion of developmental changes in the form of deposition of cuticular growth rings, and there are no doubt biochemical changes as well. But the crucial question is still, as Williams et al. pointed out so long ago (1943): what is the basis, presumably in part at least hormonal, for the developmental changes including flight? The demonstration by Haskell & Moorhouse (1963) that the hormone ecdysone affects central and peripheral nervous activity is perhaps a start toward an answer to the behavioural side of that question.

This study was supported by a Postdoctoral Fellowship from the U.S. National Science Foundation. I am grateful to Dr J. S. Kennedy for sponsorship, discussion, and criticism; to Dr L. B. Slobodkin for valuable advice and comment; to Prof. V. B. Wigglesworth for facilities; and to Geraldine Dingle for criticism and typing. Dr A. C. Neville kindly prepared the section from which Fig. 2 is taken.