Co-Ordination of Successive Activities in An Aphid. Depression of Settling After Flight

The background to this work on the parthenogenetic winged form of Aphis fabae Scop, is given in two earlier papers (Kennedy & Booth, 1963 a, 6: referred to hereafter by date only), where it was shown that aphids allowed to fly and alight freely in a laboratory flight chamber become increasingly ready to settle down, feed and larvi-posit on a leaf the longer they have flown. Flight and settling are the two main activities of these insects and they are nervous antagonists, each a complex of responses which inhibits the other. But the strengthening of the settling responses, following flight, is not in the first place a consequence of some negative feedback weakening flight, for it occurs even when flight is still gaining in strength. This type of coordination of two successive activities was therefore thought to be a central nervous process and named antagonistic induction.

Settling responses are in large measure inhibited before flight occurs and totally inhibited when it starts ; and afterwards, when the settling responses that have been primed by flight are initiated on an unsuitable leaf, or even on a suitable one if the priming has not been sufficient, they are soon inhibited by flight once again (1963 a, b). This suggested a method of confirming that the co-ordination of flight and settling is a central nervous function, with peripheral feedback having only a reinforcing effect. If so, then it might be possible, by appropriate external stimulation alone, to tip the balance against the excitatory, priming effect of flight on settling and in favour of the inhibitory effect, even to the point of reversing the normal course of events and producing a decrease instead of an increase in settling readiness after flight. This paper gives the evidence obtained on that point.

The materials and methods were as already described (1963 a, b) except where otherwise indicated below. Some of the results now reported were obtained from two lengthy experiments previously designated I and II. Table 1 summarizes for reference purposes the main treatments these aphids received. Two additional treatments employed in Expt. II are described under Results. The treatments employed in a further lengthy experiment, III, are summarized in Table 2 and detailed below under Results. The ‘host’ leaves used throughout these experiments were either growing or fully expanded (‘mature’) leaflets from young broad bean plants (Vicia faba L., var. Seville Longpod or Colossal). The ‘non-host’ leaves were mature ones from Fuchsia in Expts. I and III and mature leaflets from potato in Expt. II. A piece of card, cut to the shape and size of a leaf and painted with ‘leaf green’ (Ostwald No. 23) watercolour paint, was also used in Expt. III.

The arrangements used in Expts. I and II (1963 a) for maintaining a continual supply of fresh adult aphids, taking flight for the first time, were modified for Expt. Ill in order to reduce the variability among the individual aphids by improved nutrition during their development and behavioural selection afterwards. One hundred young apterous females were confined for larviposition under a small (1 × 1 in.) muslin-topped cover enclosing a just-emerged broad bean seedling. The cover and mother aphids were removed 48 hr. later and the larvae left to develop on the growing seedling. The intense crowding of the larvae resulted in a high proportion of winged individuals. As soon as moulting to adult began (about 6 days later) each plant, now bearing two stunted leaves, was cut into four pieces carrying roughly equal numbers of aphids and each piece was laid upon a fresh seedling bean that had been decapitated just above the lowest leaf and had had the leaf stipules and any basal shoots removed.

The aphids left the wilting cut pieces in the next few hours and settled down on the underside of the two leaflets of the decapitated plant, where they moulted to adult, passed through the teneral period and then came up to the leaf’s upper surface and took off from one of the leaf tips which were kept oriented toward the lights. The result of this structural simplification and orientation of the plants was that the ‘flightmature’ aphids usually effected a clean departure from the plant on the first occasion they took up the ‘tip-toe’ take-off posture and raised and vibrated their wings. Previously, when the aphids matured on intact plants bearing clusters of leaves (1963 a, Fig. 3) departures were often long delayed by the disorderly disposition of the substrate with respect to the light source. The walking insect was re-orienting continually to both these factors and was liable to spend much time wandering, clambering, raising its wings, ‘teetering’ and closing them again, or over-balancing, or flipping merely from one leaf to another, until it arrived at a secure, unobstructed, right-way-up stance facing the light, from which the effective take-off occurred.

As soon as take-offs began, the aphid-bearing decapitated plants were transferred from the culture room kept at c. 16° C. to the non-return box in the flight chamber room at c. 19° C. The plants were jettisoned after 1 day, before the majority of the aphids had become flight-mature, so as to exclude from experimental use the latermaturing, less robust individuals that had developed under the less favourable conditions of more acute crowding and a more stunted plant. The non-return box (1963 a, Fig. 3) was re-designed as shown in Fig. 1, leaving a gap between the leaves from which the aphids took off and the sloping glass sheet that intercepted their flight toward the lights. This served to eliminate all specimens that merely fell off the leaves or took off without developing enough lift to clear this gap. An aphid that flew strongly enough to reach the glass was not collected from there with a brush as before but allowed to walk directly on to a detached leaf laid on the glass, and immediately carried on the leaf into the adjoining flight chamber to begin its flight treatment.

The aphids reared and selected in this way were more uniform in flight behaviour than those used in Expts. I and II. At the same time they were less ‘flight-mature’ than those aphids when collected, suggesting that flight excitability mounts during a period of clambering, ‘teetering’ and wing-raising before the effective take-off, for that period was curtailed by the modified arrangements. Thus in Expt. Ill, the initial unsteady phase of each aphid’s flying in the chamber (1963 a) was generally more pronounced and prolonged than in Expts. I and II, and, as an aspect of this, more the Expt. Ill fliers ‘ranged’ away from the space under the chamber lights into the surrounding dark and had to be abandoned before they had entered upon the steady cruising phase of flight.

The procedure for stopping flight in order to measure an aphid’s responsiveness to a leaf was in principle the same in Expt. Ill as in Expts. I and II (19636, pp. 353-4), except that the leaf was not at first placed in the holder on the adjustable arm in Expt. Ill, but held in the (black-gloved) hand, because some of the treatments required jolting the alighter off the leaf again after only one probe and this could be done more assuredly in the hand. For alighters that were to be left to take off in their own time the leaf was placed in the leaf holder after the first probe. Flying aphids normally landed within seconds of the leaf being presented alongside them. Occasionally the flier still had not landed after a minute and the leaf was then removed and presented again after a further minute, and so on.

Where a significance level is quoted for the difference between two sets of data with no mention of the statistical test used, this was the ‘one-tailed’ Mann-Whitney U test as described by Siegel (1956).

As reported in the previous paper (19636, Fig. 5), Expt. II showed that flying exerted a cumulative ‘priming’ effect on settling even when the flying was repeatedly interrupted, at long or short intervals, by landings on a leaf where settling was initiated but then inhibited again. As an aphid’s series of flights and landings lengthened, the chances that it would make a settling response stronger than any it had made before increased, at a rate depending on the type of leaf used. Yet there was certainly no progressive increase, with each successive landing, in the strength of the individual’s settling responses on any leaf, as can be seen from the examples in Fig. 2. Actual regressions were common, when the sequence of settling responses was broken off at an earlier stage than the aphid had reached at its previous landing. The settling responses at successive landings often fluctuated apparently at random and it was not possible to predict when a new advance would occur (cp. Johnson, 1958, Table VII). To discover whether or not the regressions reflected anything more than random fluctuations not affecting the overall priming of settling as the flying time was added to, comparisons were made of the settling responses elicited by a given surface after flights with and without interruption by prior landings, in Expts. I, II and III. In order to obtain the maximum effect from the prior landings, relative to the priming effect of flight which was known to increase with flight duration, the flights allowed between prior landings and between the last of those and the test landing were kept brief, the leaf being presented 1 min. after take-off in each case. For the same reason the total duration of flying done before the tests were completed was kept within 25 min. in Expt. III.

In Expt. I (Table 1) most of the aphids that took off again after their first landing, on host or non-host, were kept flying for 1 further minute and then allowed to land on a host leaf. The previous paper (1963b, Figs. 2E and 3) showed a sample record from one of these aphids and their collective behaviour at the first landing only.^* For present purposes, Fig. 3 adjoining compares the collective behaviour of the aphids that landed directly on the host at their first landing, with that of the other set of aphids when they landed on the host at their second landing, after a previous landing on the non-host and 1 further minute’s flight. Fig. 3 is simplified by pooling the results from the aphids that were given very brief first flights of 15-30 and 60-75 sec. The results for each of the six groups are expressed as a frequency distribution obtained after placing each individual into one of five behavioural classes, representing the furthest stage to which it advanced, at that landing, through the sequence of settling responses as previously described (1963b).

Fig. 3 shows that some of the ‘priming’ of settling due to the first flight persisted when the settling was inhibited again on the non-host and flight was resumed, for at their second landing this set of aphids still showed stronger settling responses the longer their first flight had been. But, as Fig. 3 also shows, there was some loss of settling readiness after a first landing on a non-host in spite of the additional minute of flying done after it. The aphids that flew for 6 or 40-60 min. gave weaker settling responses to the host when they had landed on a non-host before it, than when they landed on it directly (P < 0-025 for difference between the two lots given a 6 min. first flight). Johnson (1958) made a similar comparison, giving tethered fliers a long flight and then releasing them on a host leaf either directly, or after prior release on a non-host followed by 20 sec. untethered flight. In one of his experiments a notably smaller proportion of the aphids that had encountered the non-host leaf beforehand settled down to larviposit on the host (in his other experiment, Table IX, the difference appears insignificant when the columns are re-totalled). When the first flight was very’ brief, however, as for one lot in our Expt. I (Fig. 3), the aphids that had landed on the non-host beforehand did not give weaker responses to the host than aphids that landed directly on the host the first time (this was true also of the aphids given a first flight of 60-75 sec. considered alone). And when both landings were made on the host the collective results revealed no consistent effect of the first landing on settling readiness at the second; but these aphids will be reconsidered when individual behaviour is analysed below (Fig. 9 and Table 3).

In Expt. II all the treatments involved serial flights and landings and the aftereffect of a number of prior landings on a non-host was then examined, using a treatment additional to those listed in Table 1. Eight aphids were given a series of 8-14 successive landings on a non-host leaf between 1 min. flights, then given 1 further minute’s flight and tested for settling readiness on a young host leaf. All of them now took off again without settling down to larviposit on the leaf. They were then given four more 1 min. flights and landings on the non-host followed by another 1 min. flight and another test landing on the host leaf, and so on, repeatedly. Two of the aphids settled down and larviposited at their second landing on the host leaf ; two other aphids took off from it a second time without larvipositing, another did so four times, another five, another eight and another ten times. The total time flown without settling down thus varied from 16 to 66 min. among these eight aphids, although they had landed on the host leaf after about 10 min. of flight and again after every additional 5 min. of flight.

This unreadiness to settle down on the host contrasted sharply with the behaviour of the aphids which never encountered a non-host and landed only on a young host leaf between their i min. flights, in the same experiment (Table 1 ; and 19636, Fig. 5). Fourteen of the twenty aphids in that group settled down to larviposit after flying for a total of 5 min. or less, and, of these, four settled after 2 min. and six after 1 min. of flight.

The question now arose as to whether repeated landings on a host could also reduce subsequent readiness to settle, or whether the reduction was an effect peculiar to the non-host. The above results did not settle this point because fewer landings were usually made on the host than on the non-host, before settling readiness was tested; most of the aphids soon settled down on the host. Another group of eight aphids were therefore given repeated 1 min. flights and landings on a young host leaf in Expt. II and these were prevented from settling down by the operator tapping the tube holding the leaf, lightly but sharply, whenever the aphid went below the leaf and before it started to probe there. This jolted the aphid off the leaf into flight again, to continue the series.

Sooner or later all these aphids took off spontaneously from the leaf upper surface, without either going below or being jolted. They continued to do this, intermittently, through long series of flights and landings, even to the ‘full ranging’ stage of flight (1963 a). Their unreadiness to settle contrasted with the behaviour of the aphids which, in the same experiment, reached the stage of ‘full ranging’ by uninterrupted flying, and only then landed on a young host leaf; these aphids went below in almost every case (1963b, Fig. 4). The results obtained later, in Expt. Ill (p. 813), made it unlikely that the eventual unreadiness of the jolted aphids to settle in Expt. II was due to over-exciting flight by repeatedly depriving them of tarsal contact.

Expt. Ill was designed to provide a direct comparison of the after-effects of interrupted and uninterrupted flights. This was done by giving aphids approximately the same total duration of flight, about 20 min., either uninterrupted, or interrupted in various ways summarized in Table 2; and then giving them a test landing on one standard leaf. Their response to it could then be compared with that of other aphids allowed to land on the same standard leaf after only i min. of flight, to determine how much, if any, ‘priming’ of settling the additional flight had brought about. Interrupted flights were allowed to continue until they totalled slightly more than 20 min. for each aphid, to ensure that any bias toward readier settling, due to longer flight, would favour these aphids as against those flown uninterruptedly.

Two measures of each individual’s settling behaviour at its test landing are used in presenting the grouped results of Expt. Ill in Figs. 4-7 : a qualitative measure, viz. the furthest stage reached in the sequence of settling responses as in Fig. 3 (also 1963b); and a quantitative one, viz. the total duration of the aphid’s stay on the leaf between touch-down and take-off. Aphids that settled down and larviposited, and hence stayed longest of all, are combined in one last category in both parts of the figures. It can be seen that the results, according to these two measures of settling, display a similar pattern of differences among the treatment groups, and it is the stay durations that have been used for statistical tests. The advantage of this is that the stay durations fall on a continuous scale so that the individuals can be ranked with fewer ‘ties’ and fuller use is made of the available information.

The aphids in the control group A were flown uninterruptedly for the 20 min. before their test landing on a young host leaf. Groups B and C were both given repeated 1 min. flights and landings on a non-host leaf before their test-landing. The B aphids were left undisturbed on the non-host leaf after each landing and took off when they were ready; the C aphids were regularly stimulated into renewed flight prematurely, by jolting the leaf as soon as they started walking after completing their first probe on the upper leaf surface, or as soon as they took up the pre-take-off posture if they did not probe first. Both B and C aphids were of course left undisturbed at the final test-landing on the young host leaf. Two further groups, D and E, were treated like B and C, respectively, except that all their landings were made on a young host leaf. Aphids in group D that settled down to larviposit before they had completed the desired total of 20 min. flying were abandoned. This ‘wastage’ was avoided in group E by the jolting off of the aphids. The aphids in one further group, F, made repeated undisturbed landings like B and D, but on the green card ‘leaf’, before their test-landing on the young host leaf. The behaviour of the group D aphids at their first landing provided a measure of settling on the standard young host leaf after only 1 min. of flight (treatment D₁).

The results from treatment groups D₁ and A in Fig. 4 show once again that settling readiness was increased by previous uninterrupted flight, and the A results provide a standard by which to gauge the effects of the other treatments in which the flight was interrupted. By that standard, a series of prior landings on a non-host weakened the settling responses to a young host leaf considerably (B v. A in Fig. 4; P < 0-01). An extra group of twenty aphids (group AA, not shown in Table 2 or Fig. 4) were flown uninterruptedly for only 10 min. before being given a test landing on the young host leaf. They responded to it rather less strongly than group A, as expected, but nevertheless more strongly (P < 0-03) than the interrupted group B which had done twice as much flying in all.

Settling at the test-landing on the young host by the aphids in group C (Fig. 5) did not differ significantly from that of group B and was in fact slightly stronger ; being barely significantly weaker than that of group A (P = 0-053). Thus the repeated stimulation of flight by jolting the aphids off the leaf (treatment C) did not in itself reduce settling readiness afterwards. Settling on the host by the aphids in group E (Fig. 5), after they had been jolted off a host leaf repeatedly, likewise did not differ significantly from that of group B and was distinctly weaker than that of group A (P < 0-04). If anything, group E settled less than group C, but the difference was small. Evidently it made little difference to the aphids’ subsequent settling on a host whether their prior landings were made on a host or on a non-host, provided the responses made at those prior landings were kept, artificially, at about the same level. Settling on the host was reduced by prior landings on both types of leaf.

The type of leaf used for the prior landings made a marked difference when the responses to it were not kept artificially at about the same level. The aphids in group D had made their prior landings on a host and been left to take off when they were ready, without jolting, and had therefore stayed longer before taking off again than the aphids in groups B, C or E. Yet Fig. 5 shows that the eventual settling, at the testlanding on a host, of those D aphids that flew for the prescribed period like groups B, C and E was the weakest of all.

The D histograms in Fig. 5 refer to nine individuals only, because the eleven others in the group had settled down to larviposit on the host leaf before they had completed the prescribed total of 20 min. flight, eight of them within the first 5 min. Thus the nine that did complete the D treatment were selected for unreadiness to settle, unlike the aphids that completed the other treatments. To make allowance for that, these nine D aphids that had not larviposited on the host after 20 min. flight were compared with the aphids that did not larviposit on the host after 20 min. flight in each other group, thus selecting from these groups also the aphids that were less ready to settle. Even so, the nine non-larvipositing D aphids stayed for less time at the test-landing on the host than did the non-larvipositors in groups A (P < 0 01), AA(P < 0-01), E(P < 0-01), B and C (P > 0-05 for both). This sub-group of nine D aphids was unique in showing no increase in settling readiness after 20 min. of flight.

Repeated previous landings on the green card ‘leaf’ had a somewhat ambiguous effect on behaviour at the final test landing on the young host leaf. Fig. 5 shows that settling by this group F, taken as a whole, was then insignificantly weaker than that of group A, which had made no previous landings at all. But the proportion of aphids that settled down to larviposit on the host was smaller in group F than in group A and, in the result, settling by the F group as a whole was not significantly stronger than that of groups B, C and E, which had made their previous landings on true leaves. Yet it is clear from Figs. 4 and 5 that the F group included fewer brief stayers (0-1 and 1-5 min.) at the test landing on the host than groups B, C, D or E. All the F aphids probed at least twice on the test leaf, whereas 10-35 % of the B, C, D and E aphids probed only once or not at all before taking off again. Thus, on balance, flight that was repeatedly interrupted by landings on a card, where probes and stays were the briefest, provided the nearest approximation to uninterrupted flight in its ‘priming’ effect on the settling responses to a host leaf.

Fig. 6 shows the effect on eventual settling responses to a non-host leaf, of interrupting the flight with prior landings, as observed in Expt. III. Thirty aphids forming an additional group, G, not so far mentioned, were flown uninterruptedly for 18 min. and then given a test landing on a non-host leaf, Fuchsia. As expected, they responded to it more strongly than aphids landing on a similar leaf after only 1 min. of flight (B₁ in Fig. 6). But, when these latter, group B, aphids had completed a total of 19 min. flight interrupted repeatedly by landings on the same non-host leaf, they responded to it (B₁₉) by much shorter stays than the uninterrupted fliers of group G (P < 0-001). The B aphids’ stays on the non-host leaf were in fact shorter by that time than their own responses to the same leaf had been after only 1 min. of flight (BJ(P < 0-05).

Fig. 7 shows the effect on eventual settling responses to a green card leaf of interrupting the flight with prior landings, again in Expt. III. Group H comprised only six individuals, which were flown uninterruptedly for various periods ranging from 20 to 76 min. and then given a test landing on the card; they are included in Fig. 7 merely to confirm that aphids so treated would respond to the card somewhat more strongly than aphids landing on it after only 1 min. of flight (F₁). But, when these latter, group F, aphids had gone on to complete a total of 20 min. flight between repeated landings on the same card, their stays on it (F₂₀) were briefer than they had been after 1 min. of flight (P < 0-01).

Note that these same B and F aphids were given their test landing on the host leaf immediately (i min.) after their last take-off from the non-host surface. On the host leaf they showed clearly that their preceding interrupted flight had had the usual excitatory after-effect on (i.e. had ‘primed’) their settling responses, for these were now to some degree stronger than the responses of aphids landing on a host leaf for the first time after only i min. of flight (DJ, as already described (Figs. 4 and 5). But Figs. 6 and 7 show equally clearly that an opposite, weak, inhibitory after-effect had been developing concurrently.

Figs. 4 and 5 give no such direct evidence of an inhibitory after-effect of landings on a host. The responses of the nine non-larvipositing aphids of group D, taken together, were not less at their arbitrarily chosen ‘test’ landing on the host than at their first landing. But there were wide fluctuations in the individuals’ settling responses from landing to landing on a host (Fig. 2); and the other aphids in group D that settled down to larviposit before that had completed 20 min. of flying have so far been left out of account. The behaviour of group D along with the other groups given interrupted flights was therefore examined in detail throughout their series of landings.

Because of the individual fluctuations some method of expressing in a single figure the collective behaviour of each group at each successive landing was required to reveal any general trends. The simplest method was to attach an arbitrary numerical rating to each grade of response (no probe, 1 ; one probe, 2 ; more than one probe, 3 ; going below, 4; and larviposition, 5), then ‘weight’ the numbers of individuals responding in each way accordingly, and take the average of the weighted figures ta give an overall settling ‘score’ to each group at each landing.

The serial scores for three of the treatment groups in Expt. Ill are plotted in Fig. 8. Group F shows, collectively, a fairly continuous slow decline in settling readiness from the first landing on the card to the last. Group B shows a slight initial decline and then remains almost steady. These declines in groups B and F were ultimately significant in terms of stay duration, as already mentioned (pp. 814-15). The curves for groups C and E (regularly jolted off host and non-host leaves, respectively, after one probe) are not shown in Fig. 8 but remain almost steady at the B level throughout, with no hint of a trend. Group D shows a sharp initial rise and then increasing variability and what appears to be an overall decline. Inspection of the detailed results suggested the apparent decline was mainly due to the two strongest responses (going below the leaf, either with, or without, larviposition) becoming less frequent. The proportion of occasions when the aphids went below the leaf was therefore compared during the first ten and the second ten landings, combining all aphids, and showed a small but significant decrease (from 53/145 to 26/109: X², 4 - 1 ; P < 0-05).

The most conspicuous change during serial landings on a host leaf between 1 min. flights was in the incidence of the maximum response, settling down to larviposit. In Expts. II and III most of the aphids that larviposited did so within the first five landings (pp. 811 and 813); and up to 50 landings were made before all the remaining aphids had larviposited in Expt. II. In Expt. Ill the proportion of aphids that had larviposited after repeated landings on the host during 20 min. of flight (11/20, group D) was hardly greater than the proportion that larviposited at their first landing on the host, after 20 min. of flight (14/30, group4). And by that time the remaining D aphids had become significantly less ready to settle than the A aphids that did not larviposit (p. 81 −314). Hence the decline in larvipositions after the first few landings of group D was not merely due to an early elimination of individuals that were ready to larviposit, and reflected a positive depression (inhibition) of settling in the D group as a whole, as in the other groups.

The nine non-larvipositing members of group D are represented by the dotted line in Fig. 8. Unlike the B and F aphids and group D as a whole, these nine D individuals that completed 20 min. of flight did not themselves show any significant depression of settling between the first and last of their landings, as already mentioned. This does not imply that landings on the host leaf (D) had less inhibitory after-effect than landings on the other surfaces. The separate analyses of the serial landings on different surfaces serve only to provide direct evidence, from the overall depression of settling, that there was some inhibitory after-effect in each case. But they permit no comparison of the magnitude of this after-effect in the three cases because the immediate effect of each surface also differed. Comparison of the after-effects required a test of the three groups on one and the same surface.

The results of this test have already been detailed in Figs. 4 and 5, but they are summarized as settling ‘scores’ at the right-hand end of Fig. 8 to bring out the point that the greatest inhibitory after-effect on settling was produced when settling had been most excited at previous landings, and had not shown any overall decline but only a failure to increase as the flying time was added to. If we use the settling ‘score * of the aphids landing on the host after 20 min. of uninterrupted flight (group A) as a measure of the full priming effect of 20 min. flight on settling, it can be seen in Fig. 8 that this priming was apparently unimpaired by prior landings on the card (F), appreciably reduced by prior landings on the non-host leaf (B) and reduced to insignificance by prior landings on the host leaf itself (D).

Most of the aphids that took off again after their test landing in Expts. I and III were kept flying for 1 further minute and then allowed to land a second time on the same surface. This provided an additional set of data on the after-effects of interrupting flight, based on some of the aphids already considered but with what was previously taken as the test landing now serving as the interruption. The aphids concerned were those allowed an uninterrupted first flight of varying duration followed by an undisturbed landing, in fact, the three lots that landed first on the host leaf in Expt. I (Table 1); and, in Expt. Ill (Table 2), the groups that landed on a leaf or card after an uninterrupted first flight.

Each point on each graph in Fig. 9 shows for how long one individual stayed on the given surface at its first and second landings. The distribution of points above and below the diagonal broken line, along which first and second stays are of equal duration, gives a measure of the tendency for settling to increase or decrease after the additional i min. of flight, in the group of aphids represented by that kind of point. Within each such treatment group the individual differences between first and second stay durations were ranked for a Wilcoxon matched-pairs signed-ranks (one-tailed) significance test (Siegel, 1956), the results of which are shown in Table 3.

Fig. 9 and Table 3 show there was sometimes a definite tendency for settling to decrease instead of increase after the added 1 min. of flight. There were two trends in this respect. Compare, first, aphids given a similar length of initial flight (i.e. represented by one kind of point in Fig. 9, or lying in one column of Table 3). The tendency for settling to decrease was then more marked, the lower the ‘host status’ of the surface landed on. Decreases were least frequent on the host leaves (Figs. 9A, B), more frequent on the non-host leaf (9C), and most frequent on the card (9D), where the overall decrease was very significant even when the initial flight had been only 1 min. (Table 3). This confirms the previous results during repeated landings on these three surfaces, as shown in the main part of Fig. 8. In both cases the comparison is between different surfaces, with the same total amount of flight, and the weaker the settling responses typically elicited by a surface, the more likely they were to become weaker still at subsequent landings on it.

Comparing now the responses to any one kind of surface after different lengths of initial flight (shown within any one diagram of Fig. 9, or one line of Table 3), it is clear that there was a much stronger tendency for settling to decrease at the second landing when the initial flight had lasted for a number of minutes, than when it had lasted for only one. The overall decrease in the longer-flown groups was significant on the mature host leaf and on the non-host leaf, although not on the young host leaf, where the minority of individual increases included some large ones. After a i min. initial flight, only the group of aphids landing on the card showed a significant decrease, while those landing on the host leaves showed an overall increase (not quite significant).

The first effect of a longer initial flight was of course a longer stay at the first landing, as is evident once again in the distribution of points along the horizontal axes in

Fig. 9. This was a feature in common with the former comparison of settling at a ‘test landing’ on one kind of surface after previous landings on different surfaces, summarized at the right-hand end of Fig. 8. The after-effects were also similar, for in both comparisons the aphids that had stayed longer beforehand more often showed a decrease of stay duration at their subsequent test landing. The results were reexamined in this light.

Fig. 9 shows wide variation in the duration of the first stay even within each treatment group, i.e. after any one length of initial flight, and if the duration of the second stay is considered simply in relation to the duration of the first, without regard to the length of the initial flight, then a more consistent pattern emerges. Every graph shows the same general trend : the longer the aphids stayed at their first landing (above a certain minimum in Figs. 9C, D), the more likely they were to stay for less time at their second landing than at their first. Conversely, the shorter the first stay (down to a certain point, in Figs. 9 C, D), the more likely was the second stay to be longer.

In order to check this conclusion on a larger sample of pairs of successive landings, the entire series of repeated landings on the host leaf, on the non-host leaf and on the card, in the course of 20 min. flight in Expt. Ill, were analysed on similar lines. Every pair of successive landings by every aphid in the group was first classed according to the duration of the aphid’s stay at the first of the two landings. Then, within each class, the numbers of cases in which the stay duration increased and decreased at the next landing were separately summed, and the ratio of increases to decreases was calculated.

These ratios are plotted in Fig. 10, where a more complete version of the same pattern as was perceptible in Fig. 9 is seen for all three surfaces: increases preponderate after a first stay that is brief, for the given surface, and decreases preponderate after a long one. The three curves differ mainly in position, as the data in Fig. 9 also imply. Thus the increase/decrease ratio fell below unity in Fig. 10, that is to say decreases came to outnumber increases, when the first stay reached 1-2 min. on the host leaf, about 0-5 min. on the non-host leaf and about 0-2 min. on the card.

Most of the aphids in Expt. Ill were given three more 1 min. flights and landings on the host leaf after their first two, and their 2nd − 3rd, 3rd −4th and 4th −5th stays were analysed in the same way. Groups AA and A, which had made only two landings since their uninterrupted first flight of 10 or 20 min., reproduced the same patterns at these later landings as is seen in Figs. 9B and 10D. On the other hand, groups B, C, D and E, which had already made more than 20 landings, now showed no significant preponderance of decreases over increases even when the first stay exceeded 5 min.

Previous discussions of this co-ordination (Kennedy, 1958; Kennedy & Booth, 1963 a, b) were based on Expts. I and II only, and assumed that aphids took off again after landing on a surface because the surface itself eventually inhibited settling. Settling was regularly and most quickly inhibited on non-host surfaces, and the depression of settling at a subsequent landing was therefore regarded as an aftereffect of landings on non-host surfaces especially (cp. Johnson, 1958). The results of Expt. Ill (and of one set of results in Expt. II, p. 811, the significance of which was not grasped at the time) show the situation is otherwise. When proper comparisons were made (Figs. 4, 5, 8, 9 and 10) it was found that, the weaker the previous settling responses and the more quickly they were inhibited again with the resumption of flight, the less effect did they have on the strength of the settling responses at a subsequent landing. The maximum depressing effect on subsequent settling was obtained when the previous landings were made on the young host leaf where the settling responses were most strongly excited, and where flight remained inhibited for longest before settling was at last inhibited again in its turn. The stronger were the settling responses at previous landings on any surface, the greater was the subsequent depression.

The depression of settling after interrupted flight is not therefore connected with inhibition of settling by the surface previously landed on. There was, in fact, no warrant for assuming that settling was inhibited by any of the landing surfaces used in this work. They all stimulated it, albeit to a very unequal extent, and what opposed their excitation of settling and sooner or later inhibited settling entirely were the antagonistic locomotor responses, walking and flight, stimulated by the light. The diminished excitability of settling at a subsequent landing came strictly as an aftereffect of the intervening flight, and was typically observed when that flight had shown heightened excitability. The heightened excitability of flight was itself an after-effect of the previous landing, a point already mentioned by Kennedy (1958) and to be documented in a further paper on the after-effects of settling on flight. These effects make up the reciprocal aspect of the co-ordination of flight and settling as successive, antagonistic activities.

The second point that has emerged with regard to the after-effects of flight on settling is that interruptions can do more than diminish the priming effect of a flight. If the most that occurred as a result of interrupting a flight was that the settling responses were no stronger at the end of it than they were at the start, then it might be supposed (Johnson, 1958) that each interruption had simply allowed the primed state of the settling responses, engendered by flight, to decay again; there would be no need to invoke nervous inhibition. This would fit the assumption (Johnson, 1958, and other references in 1963 b) that the strengthening of settling by flight is due to negative feedback from some peripheral consequence of flight. A recovery process would then be expected when flight ceased. But mere decay of the primed state of the settling responses will not explain the actual decreases of settling from one landing to the next which were obtained regularly, in spite of the priming of settling due to the additional flight in between, when the aphid’s settling responses at the first landing were relatively strong (Figs. 6-10). Nor will decay as a time function explain why, after the same duration of stay at one landing (30-60 sec.), settling increased in strength at the next landing when both landings were made on a host but decreased when they were made on a non-host (Figs. 9-10). These conditions revealed an inhibitory after-effect of interrupted flight upon settling.

The fact that flight can have a depressing instead of a priming effect upon subsequent settling, depending in a roughly predictable manner on external inputs to the system, confirms the previous conclusion (1963 a, b) that these behavioral sequences are co-ordinated centrally in the first place, with peripheral feedbacks playing a reinforcing role on one side or the other but not the determining role often assumed. The strengthening of settling after its temporary inhibition by flight was likened to ‘post-inhibitory excitation’, ‘successive induction’ or ‘rebound’ at lower integrative levels and distinguished at the behavioral level as antagonistic induction (19636). The opposite, inhibitory after-effect now demonstrated may be likened to ‘post-inhibitory depression’ (Bullock, 1957) and called at this behavioral level antagonistic depression. These two types of behavioral co-ordination are shown in idealized form in Fig. 11.

The third and perhaps most important point that emerges from these results is the distinction between the sequences of behaviour actually observed and the types of co-ordination out of which they are built. The antagonistic induction and depression of settling as diagrammatized in Fig. 11 refer to types of change in the excitability of settling after flight and are not literal descriptions of behavioral sequences. Behavioral sequences may take this form in certain conditions but the same type of co-ordination may be at work when the behaviour itself takes a different form owing to interaction with other components.

Thus, Expt. I (p. 809 and Fig. 3) and Expt. Ill (p. 816 and Figs. 4-8) showed that the excitatory and inhibitory after-effects of interrupted flight developed concurrently in the same aphids. The strengthening or weakening of the settling responses actually observed after their flight was therefore the integrated result or resultant of the two opposite after-effects, and not a direct expression of one of them acting alone. For example, aphids that had flown for 20 min. repeatedly interrupted by landings on a non-host leaf, responded afterwards to a host leaf less strongly than aphids that had flown for 20 min. without interruption (Fig. 4). This was an example of the inhibitory after-effect of the interrupted flight: antagonistic depression. But the settling of these aphids on the host leaf was not literally depressed by their interrupted flight. Compared with that of aphids that had flown for only 1 min., their settling was strengthened, this being the usual priming or antagonistic induction of settling by flight. Further, the strength of the settling responses actually obtained depended on the strength of the immediate external stimuli received from the landing surface as well as on the after-effects of previous responses. The priming of settling in the aphids just mentioned was detectable only when they were allowed to land on a host leaf; so long as they were landing on the non-host leaf only depression was evident (Fig. 8).

Although antagonistic induction and antagonistic depression are processes deduced from purely behavioral observations, they are, then, components rather than counterparts of actual behavioral sequences and in that sense belong to a lower level of analysis.

Valuable assistance from Mrs. L. Crawley in preparing this paper is acknowledged with thanks.