TRAIL FOLLOWING AND ITS DEVELOPMENT IN THE NEOTROPICAL ARMY ANT GENUS ECITON ( HYMENOPTERA : FORMICIDAE : DORYLINAE ) BY HOWARD TOPOFF

INTRODUCTION Results o numerous ield studies on the neotropical army ants Eciton hamatum (Fabricius) and E. burchelli (Westwo.od) have shown that newly eclosed callow workers do not participate in adult activities such as raiding until they have matured in the colony environment for .several days after the onset of the nomadic phase. Nevertheless, the callows of both species do emigrate along with the colony, beginning on the very first nomadic dy (Schneirla, 97). We have offered several hypotheses to account for the behavioral deficit of the callows. One hypothesis proposes that the sensitivity of the tallows to their colony’s chemical trail is too low for them to be able to follow along it during the day’s raid. By late afternoon, however, the strength of the trail may be sufficiently increased (as it is constantly reinforced by tens of thousands o.f ants running back and forth over it) so that the callows are able to participate in the emigration. This paper presents the results of a study designed to compare the performance of callow and mature adult workers o( E. hamatum and E. burchelli on their own colony’s trail.


INTRODUCTION
Results o numerous ield studies on the neotropical army ants Eciton hamatum (Fabricius) and E. burchelli (Westwo.od)have shown that newly eclosed callow workers do not participate in adult activities such as raiding until they have matured in the colony environment for .severaldays after the onset of the nomadic phase.Nevertheless, the callows of both species do emigrate along with the colony, beginning on the very first nomadic dy (Schneirla,97).We have offered several hypotheses to account for the be- havioral deficit of the callows.One hypothesis proposes that the sensitivity of the tallows to their colony's chemical trail is too low for them to be able to follow along it during the day's raid.By late afternoon, however, the strength of the trail may be sufficiently increased (as it is constantly reinforced by tens of thousands o.f ants running back and forth over it) so that the callows are able to par- ticipate in the emigration.This paper presents the results of a study designed to compare the performance of callow and mature adult workers o( E. hamatum and E. burchelli on their own colony's trail.

METHODS
This study was conducted on Barro Colorado Island, Panama Canal Zone.Features of the Island that make it suitable or army ant research have been described by Rettenmeyer ( 963).
Field methods consisted of daily patrols of the Island's numerous trails to locate statary colonies.The first indication that a colony was about to enter a new nomadic phase was the appearance of empty pupal cases on the ground near the bivouac, and the presence o lightly-pigmented callow workers inside of the nest.The t]rst nomad- ic day was defined as the day on which the first emigration of the entire colony t)ok place.
Callow and mature adult ants were collected by collapsing a sec-

Psyche
[December tion of the bivouac wall with a trowel, and allowing the ants to drop into a plastic box.When this method did not yield enough ants, we aspirated additional individuals from the area around the nest.
In the laboratory the ants were kept in a glass aquarium.
At the start of each day's test, a petroleum ether extract of whole ants was prepared by adding 6o ml of ether to.a jar containing 3oo intermediate-.sizedmature adult ants.A circular trail of the ether extract (63 cm in circum;erence) was deposited on discs of filter paper, by allowing the extract to flow out of a microburet suspended above a phonograph turntable which was spinning at 78 rpm.The ether evaporated in 5 seconds., and in order to observe whether the test ants were following the trail, the filter paper was placed upon a corresponding circular template.The template circle, as well as the lines dividing it into I6 sectors, were clearly visible through the discs of filter paper.A new disc of filter paper was used for each trial.
In order to insure uniform orientation of each test ant at the start of trail ollowing, a plastic enclosure was used (Fig. DISC OF FILTER PAPER Fig. 1.Apparatus used to insure that all test ants initially orient to the trail in an identical manner.For explanation, see text. consisted o a circular chamber (5.o cm in diameter, t.4 cm high) which opened into a tunnel that was 7.9 cm long.The tunnel was the same height as the adjoining circular chamber, and was .2cm wide.This width was chosen because it was approximately equal to the leg span of the intermediate-sized workers o E. hamatum and E. burchelli (i.e., workers whose overall length was between 7.0 and 9.0 ram).The curvature o the tunnel was equal to that o the trail.
The circuIar chamber was divided approximately in half by a vertical partition attached to its lid.Each test ant entered the cham- ber through a hole in the lid, and the partition prevented it rom contacting the trail during the 5 second period in which it was al- lowed to adapt to the chamber.At the end of this interval the lid with the partition was replaced with another lid without a partition, thus allowing the ant access to the trail.
For each dilution o ether extract that was used, 6.o intermediate- sized ants (3o callows and 3o mature adults) were selected at ran- dom rom the mass of ants in the aquarium.The dilutio.ns were prepared by adding different .amountso pure ether to the extract.
Ants were tested individually, and as indices of trail tollowing we recorded the number of ants following, the distance travelled on the trail (maximum distance 580 mm), and the speed o travel along the trail.I a.n ant did not leave the runwa:y at the end of one minute, a negative response was recorded.

RESULTS
The results of comparing trail-following performances between callow and mature adult worker ants from three colonies o E.
hamatum and one colony of E. burchelli a.re sho.wn in Table I.It is clear that when comparisons are made as to, the relative ability of callow and mature adult ants to follow a.ny .portion of the trail, as well as comparisons of the number ro.m both groups able to com- plete the trail, the differences are not significant in approximately two-thirds of the cases.The third measure of trail following we considered was the median distance (in ram) ollowed by callow and mature adult ants in each test series.Again, these differences were usually not significant.The only measure of trail-following performance that consistently illustrated the behavioral immaturity of the callows was their speed of running over the trail.At all dilutions on all nomadic and statary days in which tests were con- ducted, the running .speed of the callows of both species was slower than that of the mature adult ants.

DISCUSSION
It is clear that the absence o callow army ants rom raiding trals during the irst ew days o the nomadic phase can not be ex- pla.ned on the bass o their nablty to ollow the chemical trail o their colony.Nevertheless, there are sgniicant derences n trail- ollowng performance between callow and mature adult ants, espe- cially in their speed o. running along the trail.I the callows did attempt to participate n ra.iclng activities, their deict in locomctory ability could make them very necent predators.This, together wth the observation that callow army ants placed in raiding columns are behaviorally disorganized (Schneirla, I97I) makes it likely that it is adaptive to the colony for the callows not to participate in raid- ing until they have matured.However, the fact that the behavior of the callows is adaptive to. the colony still does not explain the mechanism that keeps them inside of the nest.
Although there is relatively little data on the behavior of the cal- low workers in the nest immediately after their eclosion, Schneirla   (952) reported that they feed voraciously.Rettenmeyer (I96.3) also observed callows feeding on booty, but pointed out the difficulty of estimating the quantity of food consumed or the duratio,n of the active feeding periods.If the callows are preoccupied in feeding intensively during the first few days after their emergence, from the pupal stage of development, this could account for their absence from raiding columns.
Another hypothesis stems from the observation that callow workers tend to cluster tightly together, both in the bivouac and in emigra- tion columns (Rettenmeyer, 963 Schneirla, 938).This intense clustering may result from an extreme sensitivity of the callows to chemical and tactual stimuli.Inside the bivouac the intensity of tactual stimulation is high, as a result of the continual contact among callows and other ants.The nest is also saturated with trail sub- stance and other chemical stimuli (which arise from the workers, the queen, and even the booty).Accordingly, whenever a callow leaves the bivouac the intensity of stimuli impinging upon it decreases abruptly.Outside the nest the amount of tactual stimulation is de- creased as the adult ants fan out along the trail.The intensity of chemical stimuli is also decreased, as volatile chemicals diffuse and as the concentration of trail chemicals decreases from a dense sphere surrounding the ants inside the nest to only a na.rrow trail leading away from the nest.As a result, .callowsattempting to leave the nest would experience a sudden decrease in chemotactual stimulation, with the result that they would become disoriented.This, in turn, would cause the callows to reverse their direction o] movement and return to the stimulus-saturated interior o the nest.By the end o the day's raid, and immediately preceding the start of an emigration, a large portion of the adult workers are out of the nest.Then, as the t:ull-scale emigration progresses, the remaining adult population, together with the brood and queen, also leaves the nest.Thus, at some time between the beginning of an emigratio.nand its peak, the relative concentration of chemical and tactual stimulation outside of the bivouac may be .sufficientto attract the callows out of the nest.This may occur because the callows are particularly attracted to the chemical secretions and tactual stimuli that arise rom one subpopu- lation of the entire colony (such as the adult workers, the brood, the queen, or even the booty).Or, the stimulative effect on the callows may be quantitative-that is the callows may leave the nest when the magnitude o stimulation outside the nest increases suffi- ciently, regardless xCrom which particular segment of the colony the stimuli arise.
We have recently made an observation in the field that is con- sistent with the hypothesis that high concentrations of social stimu- lation arising inside the nest serve to keep callow army ants inside the nest during the first xCe.wdays of the nomadic phase.During a recent study o] the nocturnal army ant species Neivamyrmex nigre- scens (Cresson) in southeastern Arizona, we located a statary col- ony bivouacked beneath a rock in the bank o a creek.The statary raids were unusually weak, with only several dozen workers partici- pating.On subsequent nights, the raids never increased in strength.One night we were surprised to find several dozen newly eclosed callow ants running sluggishly over the entire raiding route (which extended up to 50 rn t:rom the nest).No emigration occurred that night.The next night the callows were again observed along with the adults on the raiding trails.No emigration occurred the second night either.We then proceeded to dig up the nest, and we ]ound that the entire colony consisted of between IOO and 200 mature adults and approximately 50 callows.There was no large mature adult or callow population, no brood, and no queen.As far as I know, this is the first report of callow army ants participating in raiding immediately after their eclosion ro.m the pupal stage, and it is significant that in this one case there was no.colony to speak of, and hence no large source of so.cial stimulation.
In the army ants, as in social species representing many levels of invertebrate and vertebrate evolutionary history, immature individ- uals are not immediately integrated into the society of their group, Topoff, Lawson & Richards--Eciton but become so gradually, as a result ot numerous maturational and experiential effects that occur within the colony environment.Al- though the literature on the social behavior o ants is vast, espe- cially studies concerning the pheromonal basis ot sociality, behavior- ists have largely ignored the developmental aspects o such studies.
We hope that our contributions will stimulate more research on the development o social behavior in these insects.

SUMMARY
Comparisons were made between the trail-tollowing performance of callow and mature adult ants representing two species ot the neotropical genus Eciton.The behavio.ral immaturity of the callows was evidenced by the fact that their running speeds over the trails were significantly slower than those of the older ants.Nevertheless, the number of callows able to follow along any portion ot the trail, as well as the number able to complete the trail, was not signii- cantly dierent from the older ants.Accordingly, we can not ac- count or the absence o. callow ants ro.m raiding columns during the rst ew days after their emergence rom the pupal stage ot de- velopment as due to their inability to ollow the chemical trail.We hypothesize that the callows' sensitivity to chemical and tactual stim- ulation inside o the nest may serve to prevent them rom participat- ing in raiding activities.
Department of Psychology, Hunter College of the City University of New York, and Department of Animal Behavior, The American Museum of Natural History, New York, N.Y.10024.