AMERICAN MUSEUM NOVITATES

Pubfished by
Number 1253 THE AMERICAN MUSEUM OF NATURAL HISTORY April 8, 1944
New York City

A UNIQUE CASE OF CIRCULAR MILLING IN ANTS, CONSIDERED

IN RELATION TO TRAIL FOLLOWING AND THE
GENERAL PROBLEM OF ORIENTATION'
BY T. C. SCHNEIRLA2

Were insects unable to make and to or return to a food supply, food plants,
follow trails, the problems of, Victory home quarters, or other places by virtue
gardeners, farmers, economic entomolo- of their ability to make and to follow trails.
gists, and housewives would be consider- It is clear that an understanding of trail
ably reduced. For in almost every insect making and trail using is desirable for
order from roaches to the social insects both theoretical and very practical rea-
there are at least a few species that reach sons.

COLLECTIVE FORAGING AND CHEMICAL TRAILS

There are many different forms of trail- explanation was that the newcomers were
ing in insects, from the type of trail that is following a chemical trail left incidentally
followed in a relatively rigid, stereotyped by the first ant. Confirmation of this
manner to routes th'at are used in a com- interpretation was found in a further test
plicated, delicately varying manner, often when a clean floor was substituted for the
by a single individual alone. Among the used one after the finder ant had crossed
ants, the simplest kind is the collective from the food to the home base. Then
trail. Oti such trails the workers of a given other ants came out as before, evidently
ant colony are able to reach a food place in aroused by the returned food carrier, but
large numbers, with a minimum of delay this time they were quite unable to follow
and difficulty once the "track" is estab- her path, instead wandering about and
lished. In this paper we shall be con- reaching the food only by chance.
cerned with collective trails in particular. Other studies throw light on how such
One of the least complicated ways in trails are made. In work with a Tapinoma
which such a trail may be set up was species, Santschi (1911, 1930) noticed
demonstrated by Eidmann (1927), using that in returning from the food the first
a Myrmica colony in a laboratory nest. finder rubbed her gaster against the floor,
With a new paper floor in place, one in- evidently an outcome of her excitement
dividual was permitted to find sugar water which resulted in a trail of some kind. In-
in a distant compartment of the nest, and specting the ground in her wake with a
feed upon it. Shortly after the finder had hand lens, Santschi actually saw tiny
returned to the home compartment, a droplets in the line she had taken, suggest-
number of ants was observed to cross the ing to him that the products of the ant's
floor of the intervening compartment anal gland had been deposited when the
toward the food place, taking the same gaster was rubbed against the floor. Al-
route the finder had used in returning, though he thought this was a "deliber-
even in its detailed turnings. The simplest ate" procedure on the part of the finder,
1 A contribution of the Department of Animal a preferable interpretation is that the lower-
Behavior, the American Museum of Natural History.
2 The American Museum of Natural History and
ing of the gaster and opening of the anal-
New York University. gland orifice occurred merely as incidental
2 AMERICAN MUSEUM NOVITATES [No. 1253

(i.e., as direct reflex) parts of the ant's ability to establish collective trails to new
general excitement, no more intentional food places. A pharaonis worker, upon
than the release of sweat from the skin leaving a place where she has imbibed or
glands of an excited human individual. picked up food, generally lowers her gaster
Similarly, v. Frisch and Rosch (1926) so that its tip rubs against the substratum
have shown that when a honeybee finds a as she moves, thereby leaving chemical
new food place, in her excitement the traces on the return path which new-
products of the "scent glands" are re- comers are observed to follow, usually
leased, greatly increasing the chance that with relative ease. As a result this tiny
other bees will find the place. 4nt, wandering individually over rela-
The present writer has observed that tively large distances around the nest,
the findings of Eidmann and Santschi may appears with surprising quickness in con-
be applied directly to the case of the small siderable numbers at any newly discovered
imported ant, the reddish yellow myrme- source of food. This characteristic un-
cine Monomorium pharaonis, a common doubtedly has greatly assisted the almost
kitchen nuisance which displays a singular unmatched distribution of the species.

CHEMICAL TRAILS IN THE RAIDS OF ARMY ANTS

A strikingly different mode of estab- quickly turns and rejoins the base column.
lishing collective routes is found in the In this way the trails are extended on-
Eciton species, the army or legionary ants ward, by a relay process in which no given
of subtropical and tropical America. Few ant or ants have the "lead" for very long
ants follow their trails in a more slavish and but with many ants participating succes-
stereotyped manner than do the Ecitons. sively in the trail-blazing process.
For example, the raiding columns of E. Eciton trails frequently follow vines, the
hamatum, the golden yellow terrestrial top edges of tree roots, the worn indenta-
pillager, excel in the rapidity with which tions of peccary trails, and other topo-
new offshot columns are formed over un- graphical features along which the ants
raided ground (Schneirla, 1938). If the move readily in antennal contact with a
routes which spring up and survive longest surface. However, a given trail persists
during a day of raiding are mapped in de- and is followed by further ants because it is
tail, the result is a tree-like figure. The marked chemically, first of all by the
trunk line of this trail system leads from pioneer ants (presumably with their
the bivouac or home site, with more and glandular products) as they move excitedly
more frequent branching as the distance forward, each contributing a limited ex-
from the bivouac increases. At the very tension of the track into new terrain.
twig end of the system, usually as much as Experience with a given route is not neces-
150-200 meters from headquarters at the sary in the Ecitons. Any member of the
end of a given day, the newest trails colony can readily follow the trails of each
grow outward into unraided ground from part of the raiding system without having
which insect booty is snatched. Each passed over the particular route previously.
twig column ends in a relatively small Thus in special tests, when workers are re-
group of ants, a few dozen of them in an moved from the bivouac and placed beside
area usually not more than 25 cm. wide, any trail of the day, they readily enter and
behaving in a rather peculiar manner as run along the route, whether or not other
they push onward. Hugging close to the ants are traveling upon it at the time. Or
ground with her antennae flitting close to if Ecitons are held captive overnight when
its surface in quick spasms of wasp-like the colony moves to a new locality, they
vibratory movement, each newcomer are able to follow any fresh trail of the
to the group meanders forward hesitantly next day when they chance upon it after
a few centimeters over fresh terrain, then being released. The chemical traces
19441 CIRCULAR COLUMN AND ORIENTATION IN ANTS 3

evidently remain in existence for some time, tests in which the distinctive influence of "col-
despite repeated washing of the ground by ony odor" has been demonstrated (Fielde,
1902; Morley, 1942), indicate that ants of
beavy rains. This is shown by the fact different colonies of the same species vary in
that when colonies remain in the same their iinmediate effluvia, as well as in the chemi-
bivouac location for as long as three weeks, cal nature of the routes they make. Further-
a periodic occurrence in the terrestrial more, since preliminary tests indicate that an
Eciton worker is able to follow the trail of an-
foragers at least, a long section of trail other species (e.g., transported E. hamatum
may drop out of use for two or more days workers push hesitantly along a section of E.
after it is originally formed, then reappear burchelli trail to which they have been trans-
in use as part of a new raiding system. ported), a chemical similarity is suggested for
the routes of closely related army ant species.'
Evidence for a chemical similarity in the trails
of different colonies of the same species is offered The basis of trail-following thus would
by the ability of Ecitons to run a trail of another appear to rest on the tendency to approach
colony to whose zone they are transferred; how- sources of the ant's own colony chemical or
ever, a certain chemical difference is indicated
by a noticeably more hesitant progress than similar chemical patterns, and thus to
when on trails of their own colony. When a move forward according to the line of this
transferred ant encounters foragers of the strange chemical when on trails. What accounts
colony she is attacked repeatedly and finally
killed unless she flees the trail. Such facts, to- for the ability of the individual worker to
gether with the results of nestmate-recognition respond in this highly adaptive manner?

THE ONTOGENETIC BASIS OF TRAIL FOLLOWING IN ARMY ANTS

In the case of the Ecitons it seems im- the same brood that had lived with the
probable that the mere possession of ol- colony after hatching.
factory sensitivity brings with it this capac- Normally, for some time after a new
ity for trail following. After hatching, brood of mature pupae is removed from
trail following requires a few days to ap- its cocoons by adult workers, these cal-
pear, and circumstances do not justify lows (light-colored and readily distin-
attributing the delay to maturation, guished from regular workers) remain in
i.e., simply to an improved sensitivity based the bivouac without venturing upon the
upon further growth of the olfactory tissues raiding trails to any extent. Then, after a
themselves. The fact that the callows are day or two, they begin to appear on trails
able to run long distances when partici- in the daytime, and are observed in pro-
pating in the nightly bivouac-change gressively increasing numbers at points
movements militates against the possi- farther from the bivouac. There is also a
noticeable change in their
bility that it is merely muscular weakness this period. Upon the trails they move behavior during
which holds them back in the daytime. slowly and hesitantly at first, unable to
The possibility that the irritating effect of move rapidly or to go very far in a given
light may restrain them at first should be direction without stopping. Repeatedly
kept in mind; yet circumstances oppose bumped and shunted about by the regu-
this as the main factor. Normally, E. lars, they generally pass along at the sides
hamatum workers become fairly able trail of the column for limited distances,
followers after four or five days from readily drifting into groups at trail-division
hatching; however, the results of some points. Gradually their deficiency seems
preliminary tests suggest that being with to decrease, until after a few days when
the colony is essential for the improvement. 1 Almost a virgin field awaits investigation in the
After hamatum pupae had been kept away biochemistry of trail secretions in insects, and par-
ticularly ants. To the writer's knowledge, the only
from the colony for a period of five days study of Eciton chemical is the work of Melander
and Brues (1906). These investigators found no
following hatching, when placed indi- various traces of formic acid in whole body distillates of
Eciton (Acamatus = Neivamyrmex) species of
vidually on the raiding trails they seemed the southern United States. However, the outcome
of further
helpless and inept in contrast to others of probably is tests suggested that the basic substance
leucin.
4 AMERICAN MUSEUM NOVITATES [No. 1253

deepening pigmentation has made their join the drift of workers from the cluster;
external appearance difficult to distinguish rather they leave the interior of the bivouac
from that of regular workers, their trail passively when shunted along with out-.
running likewise is close to the general going groups of workers, and huddle near
standard. the exit until incidentally absorbed into
In view of this evidence suggesting the the exodus. Much confusion results
importance of experience, the preferable from the tendency of callows to hesitate,
hypothesis seems to be one centering waver, and turn around when brushed or
around the postulation of a rudimentary bumped by foragers returning to the
learning process of the "habituation" bivouac. By degrees the callows are ab-
type in which there develops an increased sorbed into the outgoing movement, until
readiness to approach the colony chemical. after a few hours they drain from the
In the colony, this chemical dominates en- bivouac in a steady stream, moving along
vironmental stimulation from the time of in a thick column with few interruptions.
hatching; whether resting or feeding in During the early post-hatching period
the bivouac the callow is stimulated by it, callows are seen in numbers outside the
as indeed she was stimulated during her bivouac only when they join bivouac-
active larval period. Thus it is very pos- change movements; then they are able
sible that through a simple conditioning to progress in a regular manner only when
process, perhaps begun during the larval conditions in the column afford an en-
period, the callow moves about readily in vironment of uniform chemo-tactual stimu-
the presence of this chemical and, more lation promoting continuous movement
important, turns toward increases in its in one direction. Running the trail under
concentration.' Gradually, as she moves these "hand-fed" conditions would seem
about more freely, the new worker drifts to contribute fundamentally to an abil-
into the trails in responding to the Eciton ity to make progress in the daytime, de-
chemical and to the movements of depart- veloping as foragers under shifting tactual
ing foragers. Early bivouac life thus may conditions by responding to the chemical
lead directly into trail running, with the trail itself.
initial hesitancy of the callow decreasing The Eciton bivouac-change movement
as she becomes better trained to run sharpens our view of army-ant progress in
freely forward in the presence of the colony columns as a routine stereotyped following
chemical. This hypothesis is consistent of chemical lines. The view is quite de-
with a theory for social insects in general fensible, since under any conditions the
which takes into account not only the ants are oriented mainly by the chemical
original equipment and physiological matu- path, and only secondarily by other fac-
ration of the individual, but also the tors. Through visual stimulation, for
importance of early environment and early example, they are aroused in the morning
activities for the development of social (Schneirla, 1940) and are caused to shift
reaction patterns (e.g., Wheeler, 1928a; their bivouacs away from bright light, yet
Maier and Schneirla, 1935, chap. 7). appropriate tests on their foraging and
One feature of early life which must other column activities give no evidence
hasten the Eciton trail-habituation process that visual factors are involved in orienta-
considerably is participation in the bivouac- tion. On the other hand, the prevalent
change movement of the colony. Regu- importance of an olfactory basis is readily
larly, when a new brood hatches, the E. demonstrated by simple tests of the
hamatum colony shifts to a new bivouac "Finger-versuch" type (see Brun, 1914).
site at the end of each day (Schneirla, Rubbing a finger across a trail causes a
1938). At first the callows do not readily serious interruption, with the ants as-
1 Thorpe's researches (1939) with the ichneumon sembling on each side of the violated zone,
fly in particular have led him to adopt a theory unable to cross until the gap has been
based upon the postulation of a chemically dominated
6onditioning process governing the response of preda- passed by some of them in a manner re-
tory adults to given prey upon which the individual
has fed as a larva. sembling the pioneer process of trail mak-
1944] CIRCULAR COLUMN AND ORlENTATION IN ANTS 5

ing. When a portion of trail on a leaf is long slender limb was quickly turned end
moved laterally so that each end is a few for end, so that the east end touched the
inches from its former position, there is an place formerly occupied by the west end,
interruption which eventuates in activi- and vice versa. After a brief commotion
ties similar to initial trail making, with caused by the lifting itself, traffic soon re-
the difference that the displaced section sumed without much difficulty. Since
may become part of the trail bridge if the relatively long sections of trail may be
pioneers chance to reach it. Such tests show thus reversed without blocking the ants,
that a chemical trail is an indispensable the Eciton path appears to be a simple
basis for the route. However, considerable homogeneous chemical track, not po-
liberties may be taken with sections of larized chemically as are the trails of
trail on movable objects, as long as the many other collective foragers under given
ends join the main trail somehow. In one conditions (e.g., Lasius species, Brun,
frequently repeated test, a trail section on a 1914).

CIRCULAR COLUMNS AND STEREOTYPY IN TRAIL FOLLOWING

CIRCULAR TRAILS IN PROCESSIONARY CATERPILLARS
The psychological limitations of insects and many other lepidopterous larvae of
engaging in this type of trail orientation similar behavior ordinarily follow one
are suggested strikingly when the insects another in winding columns by virtue of
form and continue to follow circular trails. contact with a thread or threads spun by
In volume 6 of his "Souvenirs entomolo- those first to pass over the ground. In the
giques," Fabre reported an interesting case processions formed by the caterpillars of
of this type observed in 1896, perhaps the the American buck moth Hemilenca maia,
first scientific record of the phenomenon for Marshall (1904) found that removing the
any animal.' Some processionary cater- leading individual usually served to halt
pillars (Cnethocampa pityocampa) he had the whole line in place for some time, until
hatched were seen to form a column around the "leader" was returned or until some
the upper rim of a large vase in a circular other larvae chanced to proceed. This sug-
path 1.35 meters in circumference. Fabre gests that in this species only the first
hastened to assist the retention of this individual or the first few to pass along are
column by brushing away all threads re- actively involved in spinning, while the
maining on the sides of the vase after the others merely follow the trail. Contact
larvae had mounted, and by removing all with the thread is not the sole factor in
excess late comers which might have following the path, since when Marshall
swollen the column beyond the limits of removed a portion of thread between
space on the rim. Once started on the end- neighboring caterpillars in the body of the
less trail, the caterpillars circled in this procession those behind were not stopped.
path for seven days, moving more or less However, rubbing across the path with a
continuously except for halts caused by low moistened finger in addition to removing
temperatures during the nights. These the thread brought those following to a
1 A similar case of persistent circling in a Cnetho-
halt, with a delay of a few minutes before
campa species was reported by Dubois in 1899, with a the gap could be crossed. Evidently, in the
brief discussion of the track itself. There is no evi- buck moth at least, the trail is followed
dence to indicate that Dubois did not discover the
phenomenon independently. chemically as well as tactually.
 6 AMERICAN MUSEUM NOVITATES [NO. 1253
WHEELER'S LABORATORY CASE IN Eciton schmitti
That Ecitons may form circular proces- was long enough to accommodate all the in-
sions of a kind very similar to that ob- dividuals. They continued going round and
round the circular base of the jar, following one
served by Fabre in lepidopterous larvae another like so many sheep, without the slightest
was first reported by Wheeler (1910), with inkling that they were perpetually traversing
the following description of the behavior. the same path. They behaved exactly as they
In captivity Ecitons are remarkably restless, do on one of their predatory expeditions. They
at least at certain times during the day. Part kept up this gyration for forty-six hours before
of a fine colony of E. schmitti which I kept some the column broke and spread over the board
years ago, exhibited this restlessness in a striking to the water's edge and clustered in the manner
and ludicrous manner. The colony was at so characteristic of this and the allied species
first confined in a tall glass jar on a square (E. opacithorax, sumichrasti, etc.). I have never
board surrounded by a water moat. The ants seen a more astonishing exhibition of the limita-
kept going up and down the inside of the jar tions of instinct. For nearly two whole days
in files for many hours. Finally I removed these blind creatures, so dependent on the
contact-odor sense of their antennae, kept
the lid. The file at once advanced over the palpating their uniformly smooth, odoriferous
rim and descended on the outer surface till it trail and the advancing bodies of the ants im-
reached the circular base of the jar where it mediately preceding them, without perceiving
turned to the left at a right angle and proceeded that they were making no progress but only
completely around the base till it met the wasting their energies, till the spell was finally
column at the turning point. To my surprise it broken by some more venturesome members of
kept right on over the same circumference which the colony (Wheeler, 1910, p. 265).

A CASE OF ECITON MILLING IN THE OPEN

Although I had observed in detail many circular cluster of ants huddled in the center
Although at this time the ring varies in width
raids of the Central American species of around its circumference and at intervals
Eciton and had frequently set up circular- moves rather eccentrically, these measurements
trail formations with these ants in the are fairly consistent for the period 7:30 A.M.
laboratory, under natural conditions noth- to 8:15 A.M.: outside diameter of the ring,
10-11.5 cm.; width of the ring, 4-5 cm.; diame-
ing but short-lived or ambiguous instances ter
of circular-column behavior appeared until gin of central cluster, 1-2 cm. The outer mar-
of the ring is somewhat irregular, due to the
September 4, 1936, when a striking case frequent straying of a few ants or a ragged file of
was discovered in the laboratory clearing them tangentially in the direction of rotation.
at Barro Colorado Island, Canal Zone. On Usually such deviates continue moving with the
the preceding day a swarm of the small the column, though more slowly, and soon rejoin
principal throng. Less frequently some of
black E. praedator had worked its way from them lose contact with the column, then rest in
the north up the laboratory hill, crossed place or move about hesitantly for a time, at
the grounds near the cookhouse, and after length blundering back where they are swept
passing near the west corner of Haskins into the main activity once more. If an ant
chances to set off in the clockwise direction when
Memorial Library finally dwindled out she reenters the rotation, her movement is soon
toward dusk in the general vicinity of the reversed in the course of a series of contacts and
Chapman cottage. At 7:30 A.M. on Sep- head-on collisions in the throng.
tember 4, Rosa, the laboratory cook, ex- butThe column does not move strictly as a unit,
at times it behaves as though two or more
citedly called my attention to an occur- individual rings of ants were moving more or
rence which beyond doubt was a sequel to less independently around a common center.
the foregoing raid. There follows a sum- Although ants in files near the periphery are
mary of my notes. moving faster than those near the center, the
first impression that the inner files usually com-
At about 7:30 A.M. a singular milling group of plete their revolution sooner than the outermost
Eciton praedator workers is discovered on the ones is readily confirmed. Apparently as a
cement walk in front of the Haskins Library
(fig. 1). There are several hundred of them, consequence of this difference in speed, there are
mainly intermediates and large minor workers, observable indications of friction between ad-
the size classes most frequent in raids. Most of jacent parts of the procession, with here and
them are running in a circular column which there a file in the column deviating somewhat
moves counterclockwise, rotating about a small from its concentric course as contact with ants
1944] CIRCULAR COLUMN AND ORIENTATION IN ANTS 7

moving beside it at a different rate causes some rival to the main action. At 8:15 when I must
of the ants to stray. Jagged irregular gaps ap- leave for forest observations (at the time con-
pear intermittently between bands in the circu- sidered more important) the central cluster of
lar column when one concentric part draws huddled ants, at first piled two or three deep,
briefly away from a neighboring section. Such is visibly dissolving into the movement. The
behavior, occurring at intervals in the general size of the, cluster plainly decreases as more and
movement, may contribute to a general ec- more .of the quiescent ants are stirred into ac-
centric wobbling that characterizes the move- tion through repeated brushing of the hurrying
ment. The deviations of sub-groups are sel- throng against them.

AVIF

Fig. 1. The circular column of Eciton praedator, as drawn from a photograph taken shortly be-
fore 12:00 P.M. At that time the ring was approximately 14 cm. in diameter.

dom marked or lasting, yet occasionally there 12:00 P.M.: The circular column is moving
arises an eddy that may nearly circle upon itself, more rapidly as a whole than in the early morn-
causing confusion until broken down by the ing, rotating as before in the counterclockwise
persistent force of the general rotatory motion. direction. The ring has expanded, now ap-
A sub-rotation of this kind was observed at one proaching 14-15 cm. in diameter. In general
time near the margin of the ring, a completed there are wider intervals between its individual
circling within the general movement which members, and more obvious gaps appear inter-
threatened for a time to become a persistent mittently between concentric sub-groups. The
8 AMERICAN MUSEUM NOVITATES [No. 1253

central cluster has disappeared altogether, and rain noticeably disturbs ants on the exposed
there is now a bare central space about 1-2.5 side of the circle. (Heavy rain from 3:45-4:00
cm. wide, entered only by an occasional ant P.M.)
which has momentarily turned too sharply to 8:30 P.M.: Somehow the group has divided
keep in touch with the innermost file. Since into two distinct circling rings of ants. These
8:15 A.M. the entire ring has shifted its position rings, nearly equal in size, presumably were
so that the center is now about 12 cm. from its caused by the heavy though brief rain which
former place, closer to the wall of the building. came shortly before 4:00. Both groups rotate
At 12:10 the ring is mainly exposed to direct counterclockwise, with their outer margins
sunlight, with only about 600 of arc on the side about 20 cm. apart; both nearer the wall than
toward the library falling within the shade. was the exposed margin of the single ring at
At 12:15 P.M. rain sets in. Ants in the mill are 3:45 P.M. In each group a few ants occupy the
of course disturbed by the drops, particularly central area, where they huddle together or turn
on the side away from the building. This side narrowly in irregular ways. During an hour of
of the mill, although mainly protected from the observation more are shunted into the central
direct downpour by the wide eave of the build- huddle, drawn from ants in the inniermost file,
ing, is still reached by a spattering of droplets where many have difficulty turning rapidly
which strikes the ants and dampens the sub- enough to evade bumping by others in line.
stratum. Ecitons entering the exposed area At 10:00 P.M. the two rings still rotate counter-
soon hesitate, then are forced to move onward as clockwise, and in both, the central clusters have
others push up behind. Forced to run through grown in size. The rate of general movement
the exposed area, they tend to crowd toward the has become very slow.
center of the mill with the result that the more September 5, 6:30 A.M.: On the spot of
exposed side of the ring becomes narrower than yesterday's phenomenon little or no circling
the opposite side, and most of them turn rather is to be seen. The entire area is strewn with
sharply away from the rainy side when this the bodies of dead and dying Ecitons. A few
becomes possible. Notwithstanding such dif- of the survivors wander about slowly, while
ficulties, the rotation continues. no more than three dozen of them form a small
2:00 P.M.: (Fairly heavy rain, 12:15 P.M. (ca. 7 cm. D.) and rather irregular circular
to 1:50 P.M.) The place occupied by the ants at column in which they plod around slowly,
12:00 is now thoroughly wet; however, the counterclockwise. At 7:30 A.M. virtually none
column continues its counterclockwise move- of the ants are on their feet. Circling has
ment in a new spot 15 cm. closer to the building, stopped, and various small myrmecine and
where the cement is fairly dry. The group did dolichoderine ants of the neighborhood are busy
not shift directly toward the building, but at carting away the dead.
angles in the counterclockwise direction. At Circumstances indicate that the Ecitons
2:15 P.M. the ring wobbles as it rotates, so that its
outer margin occasionally enters wet surface; must have died of desiccation, after having
in general, however, there is a small gap between been in nearly constant circular motion
its side and the wet cement. The entire as- over an essentially dry area for more than
semblage moves at a slower pace than at 12:00 24 hours. They might have been able to
P.M. Once again there is a central cluster,
which at 2:10 measures 4 cm. in diameter. survive much longer had they not shifted
3:40 P.M.: (Sunshine after 1:50; now dark, their ring away from the wet each time rain
beginning to sprinkle.) The circling proceeds broke during the day. It is probable that
counterclockwise, with the group about 8 Cm. these ants were marooned from the even-
closer to the building than at 2:00 P.M. Some 60 ing of September 3, since the raid stopped
of the ants have escaped the treadmill, only to
wander in a loose network of columns on the then and no other trace of the colony was
outer side (i.e., farther from the building). observed in the vicinity after dusk on that
Their number increases as the beginning of day.

PROBABLE INITIATION OF THE E. praedator MILL

Although this case may be regarded as a raiding in this locality on September 3,
special instance of trailing, it is far from operating at a distance from the main body
simple, as analysis soon discloses. Un- and connected with it by only one or two
fortunately we lack direct information con- shuttle trails when a heavy rain arose
cerning the initiation of the circling; shortly before 2:00 P.M.1 The downpour
however, circumstances warrant the follow- 1 As
ing supposition. It is very probable that library, measured by the station rain gauge near the
1.12 inches fell within little more than an
hour before 3: 00 P.M.
the group was an offshoot of the swarm that the ants were cutonoffSeptember 3. It is probable
then, since the only other
1944] CIRCULAR COLUMN AND ORIENTATION IN ANTS 9

would cause the ants to cluster, as Ecitons would trail persistently around the lower
commonly do when exposed to rain, and border of their cylindrical bivouac (e.g.,
would isolate them by flooding the area so E. hamatum, housed in tall wire nests),
that the trails connecting with the swarm form an endless column around the inside
could not be followed. (In the behavior of wall of a circular nest, or trek around
these ants on September 4, especially those objects or clusters of ants in flat glass-
that strayed away from the mill, at no time covered nests. In a situation of the last
was there any indication that shuttle trails type, an E. hamatum group consisting of
were in effective existence.) about 40 workers and a queen would move
The circling may have begun in the for hours on end in a circular path -around
following manner. It is reasonable to a 10-cm. square of cellulose sponge in the
assume that the isolated group or one main center of their nesting compartment, at
body of it was huddled in a roughly circular frequent intervals during the 8-day life
mass or confined to a small area during the of the queen in captivity. This would soon
downpour. (Such aggregations of Ecitons begin, for instance, whenever they were
commonly form when rain breaks during a newly exposed to light. The initial runs
raid.) Then, when the rain stopped, the around the sponge were slow and hesitant,
tirst ants able to run freely would be those mainly in contact with the sides and in-
situated on the margin of the huddle. Most volving sharp turns at the corners; how-
of these individuals would tend to follow ever within two days the route had become
around the periphery of the group, through circular, still leading around the sponge yet
their typical response of turning toward the not touching it.'
side of Eciton chemical and toward the side The notion that the E. praedator mill be-
of continued tactual stimulation (i.e., gan through the tendency of Eciton workers
furnished through brushing ants on the always to turn toward the side on which
margin of the mass). A circular trail of weak tactual stimulation is offered, and to
Eciton chemical would thus soon be formed move along beside a surface which main-
which these ants and newly aroused new- tains this contact, is supported by the
comers would follow in a canalized manner. following test. When Eciton workers run-
Favoring this hypothesis, there is a con- ning over a table top have settled into a
siderable body of evidence from laboratory course along the outer edge of the table (a
observations and tests. When light strikes common occurrence with E. hamatum, E.
captive groups of E. praedator or others burchelli, and E. praedator), a smaller cir-
kept in laboratory nests, after a period of cular column of them may be set up by
quiescence in darkness, a common re- placing a glass jar of 10-25 cm. D. near
sponse of those first set into action is to run the table edge so that some of the ants
around the margin of the cluster, with a per- brush it with antennae or legs as they
sistent circular column often resulting. 1 The workers in this case were respondin to the
Occasionally contact alone is sufficient to stimulative effect of the queen as well as to the trail,
as was shown by the fact that when she was present
produce this result, as was the case in a compact group of them always hurried along
Wheeler's observations of E. schmitti closely behind and beside her (many clutching at
her legs and body, or actually riding upon her), but
mentioned above. In his instance the ants always with a clear space ahead of her. When she
was removed they continued to follow the trail,
established their treadmill around a glass but now in an even column and at a slower pace. It
jar and continued moving for 48 hours; may be added that normally the Eciton queen appears
outside the bivouac only in the evening or at night,
under comparable conditions such columns to participate in the movement of the colony to a
new site. Then she follows the trail (used during the
may persist, though not operating con- day as a raiding route) readily under her own power,
tinuously, for as long as a few weeks. I accompanied by a horde of workers which greatly
swells the normal width of the bivouac-change
have observed a number of cases in the column. Far from being dragged along by workers
to the new site, as many writers have assumed is the
laboratory in which, for example, Ecitots case, she furnishes her own power and usually pulls
or carries workers with her. This applies to the
species of Eciton (sensu 8tricto) which have been ob-
rain of the day was a light one (.05 inch in all) served (E. hamatum, E. lucanoides, E. burchelli).
coming just after 7:30 P.M. Before the latter rain In the early part of the rainy season, males have been
came, the praedator raiders (as is usual) would have seen following the bivouac-change route in a similar
mainly withdrawn into their base trails and sub- manner. It is very possible that Eciton mating oc-
terranean avenues leading to the bivouac. curs under these conditions.
10 10AMERICAN MUSEUM NOVITATES [No. 1253

pass. Ants that touch the jar lightly in These results indicate the great impor-
this manner tend to leave the table-edge tance of positive response to weak contact
trail and follow the jar around its circum- in the initiation of circular columns, sup-
ference, and a circular column results porting our hypothesis concerning the
which in the end may absorb nearly all "spontaneous" formation of such a column
the strays on the table. The contact in the open. In that instance, as suggested
must be light, or the response is to turn above, a turning toward Eciton chemical
away from the stimulated side, resulting in must have been involved as well. Such a
an eddy of the table-edge column away combination of weak contact and Eciton
from tlre jar, but with no other change in chemical facilitates the formation of
behavior. If the jar is pushed too far into columnns of this type; we have observed
the initial column, so that the ants actually this repeatedly in the laboratory. Yet the
collide with it in passing rather than merely fact that light contact alone will suffice
brushing its surface, they turn away and to initiate proceedings suggests that it is
thus many get off on the vertical edge of the not so much a matter of the nature of the
table, perhaps ending by the entire group stimulus as it is the intensity of stimulation
trailing away on the floor. that arouses the approach-turning response.

ANALYSIS OF THE PERSISTENT CIRCULAR COLUMN

After the column has been active for a acute angles extending beyond the ends of
time, contact with the surface no longer is the partition in the direction of turning
essential, as is indicated by the case cited (i.e., counterclockwise). (See fig. 2.)
in which an E. hamatum group trailed in a What happened was that after slavishly
circle around a square sponge. In a more following the wall tactually for an hour or
striking demonstration, after Ecitons have two, running speed gradually increased,
formed a column around a glass jar on a and some of the ants began to overrun in
glass substratum, removal of the jar leaves rounding the corner. This brought them
a regular circular procession which usually into contact with the side wall of the com-
is able to maintain its formation for a con- partment, which they followed for a short
siderable time. The chemical trail which is distance before they turned back toward
formed as the ants follow the surface be- the partition (evidently responding to
comes sufficient alone to canalize the chemical stimulation). As their running
column. Thus in a number of instances, speed increased, the ants steadily increased
Ecitons ceased trekking around th6 out- the distance they ran along the side wall
side edge in running around the top sur- before turning back toward the partition,
face of a square wooden block, gradually and the path of their return to the partition
shifting to a course that cut the corners (at first noticeably concave inward) came
more and more, until finally they followed to be a straight line which brought them
a circular course which neared the edge only back to the partition at a point gradually
at the center of each side. closer to its opposite end. Finally the ants
It seems clear that other factors must ran a course parallel to the compartment
complicate such behavior, beyond chemical wall but not at all in contact with it, after
and tactual stimulation as described. An rounding the end of the partition; turning
interesting instance of freedom from the from this line in an arcing course that
original contact which led to new complica- straightened so that they neatly rounded
tions was furnished by an E. hamatum the opposite end of the partition. At length
group first observed running around a they traced out the parallelogram with
simple partition in their shallow glass- blunted corners shown in figure 2, in which
covered nest. After several hours it was they were near the partition only at its
noticed that they were actually describing ends. In this rather complicated example,
a parallelogram around the partition, the original contact with the partition and
1944] CIRCULAR COLUMN AND ORIENTATION IN ANTS 11

Fig. 2. Parallelogram course established by a group of Eciton hamatum workers around a parti-
tion in a shallow observation nest. The solid line represents the path eventually followed by the
ants; the arrows indicate the direction of travel at a given time. (See text.)

newly encountered contact with the side turn back toward the partition through
walls were dispensed with as a chemical what became the long side of the parallelo-
trail formed; however, as increasing satura- gram may be regarded as a result of turn-
tion of the substratum facilitated faster run- ing toward chemical stimulation when the
ning, this augmented the influence of a momentum effect had weakened suffi-
centrifugal-force effect generated in round- ciently.
ing the ends of the partition, accounting for The centrifugal-force factor, disclosed
their modifying the chemical trail to form in the running of geometribal figures ac-
the short sides of the parallelogram. The cording to laboratory conditions, is to be
12 AMERICAN MUSEUM NOVITATES [No. 1253

reckoned with when one studies trailing be- may be understood in terms of the interac-
havior under natural conditions. For ex- tion of three component factors commonly
ample, its influence is disclosed when influential in normal trail movements.
sharp turns around objects gradually shift These are: (1) a pressure, tactually exerted
outward as the trail develops, until the ants according to the principal direction of
finally run at a distance from the previ- movement in the group, which minimizes
ously "hugged" surface in rounding the variability of movement and tends to
corner. Or at other times, when a column coerce "forward" progress in the group;
at first turns off to one side promptly after (2) a drainage factor, the composite effect of
having reached the end of a stick along chemical stimulation from the trail and
which the trail leads, the course shifts in tactual stimulation (as well as chemical)
time so that finally the turn is made in a from travelers ahead and beside the given
wide arc at a distance from the end of the ant; (3) a centrifugal-force factor, the
stick. This type of modification commonly inertia of movement which tends to carry
occurs along the course of trails used for a the ant away from the circle on a tangent
considerable time during the day. Depend- with the circumference. The combined
ing upon: the slope of the incline which operation of these three factors holds the
is mounted or descended, the momentum circular column together and accounts for
which may develop through passage along the principal variations in its behavior.
a fairly straight section of trail according to The term "pressure" for our first factor
the smoothness of the terrain, and the run- is used here in the sense it was used when
ning speed which may be in force in ap- dealing with the phenomenon of organized
proaching and rounding a corner, numerous swarm movement in Ecitons such as E.
features in the initial and the later form of burchelli and E. praedator. "Its basis is the
the route may be understood. In studying proneness of the Eciton worker to turn
trail formation, one can never afford to away from the side of strong or repeated
neglect the simple mechanical principles tactual stimulation and to reverse her
underlying animal locomotion.' direction of progress if such tactual stimu-
Now we are in a position to attempt an lation is encountered in head-on fashion"
explanation of how the E. praedator mill was (Schneirla, 1940, p. 425). Thus through
kept going for more than 24 hours in the forcible bodily contact, an Eciton that hap-
open. It is clear from the facts that this pens to run against the preponderant direc-
behavior involved more than the mere tion of movement in a group is soon brought
following of a circular chemical track; into line. If she happens to lag when in a
the shifting of the entire mill through a hurrying column, forcible tactual stimula-
distance of about 30 cm. within less than 12 tion from the rear spurs her onward. This
hours dismisses that notion. However, at factor must have played a major part in
any given time the ants were responding to accounting for the fact that one direction of
a ring of chemical, as is shown by the locomotion (i.e., counterclockwise) pre-
partial disruption of travel introduced by vailed in our praedator mill. From labora-
occasional "Finger-versuch" tests (e.g., tory observations we may say that a vari-
rubbing clean glass cloth over a portion of able combination of circumstances (i.e., a
ring area vacant at the moment). In this "chance" situation) caused the majority to
phenomenon the chemical factor seems to hit upon this direction of movement at an
have been merged with other influences in a early point, and that deviates were gradu-
rather involved manner. ally forced into the same direction through
The persistence of this circular column ''pressure" until there was no longer any
1 For example, one such effect, the role of gravita- persistent clockwise movement in the
tion, has been studied by Barnes (1930) in relation group. A somewhat similar phenomenon is
to the lateral oscillatory excursions of Lasius in
running. The influence of centrifugal force as con-
sidered here commonly leads Ecitons into behavior
involved in a complex manner when a
situations resembling those described as "centrifugal nomadic movement begins to develop
swing" in the maze behavior of Formica species toward evening in a raiding system of the
(Schneirla. 1929, 1941).
19441 CIRCULAR COLUMN AND ORIENTATION IN ANTS 13

N1- N%
/ N

le/ \
//

/,
/

'N N. __

Fig. 3. Sketch representing the shift of the E. praedator ring as a result of wetting by rain in the
area indicated by wavy lines. 1-7 Indicate the initial courses of concentrically traveling ants; la-7a
their respective courses upon entering the area of disturbance; lb-7b their respective courses upon
leaving this area. 7a Indicates how individual ants are reoriented when they turn back from the
wetted surface. X, the vector of pressure upon the upper portion of the ring; Y, shortest distance to
wall under the eave; Z, direction in which the circle shifted in response to the rain. (See text.)
44 AMERICAN MUSEUM NOVITATES [No. 1253

swarm raider E. burchelli (Schneirla, in throughout the crisis. Then, after the rain,
press). it regained its earlier symmetry.'
Through the effect of pressure, in particu- Although pressure is by no means the
lar, one prominent change in the action of only agency accounting for persistent
the praedator ring may be understood. "forward" motion and the minimizing of
When rain spattered upon one side of the deviations, the "push from behind" plays
assemblage, causing confusion and ac- an important part in most of the group
counting for a noticeable faltering of ants activities of these ants and should not be
entering that area, the circular column was neglected here. Another function of pres-
kept in being especially through the in- sure in the mill, a lateral component operat-
fluence of a forceful mass contact from the ing centripetally, will be considered pres-
less exposed and more smoothly operating ently.
part of the mill. Ants that hesitated and Our second factor, "drainage," also holds
turned when struck by rain were thereby a basic place in all Eciton mass movements,
forced to turn "forward" again and some- especially in the mass organization of swarm
how rush through the damp area. Evi- raiding (Schneirla, 1940, p. 425). This in-
dences of another special pressure effect volves the propensity of the Eciton worker
were observed soon after the rain broke, to turn toward the side of weak tactual and
an effect which seemed to contribute to the weak chemical stimulation, or toward these
shifting of the entire circular mill toward in combination. This positive response to
the wall (i.e., farther under the wide eave of weak tactual stimulation has been illus-
the building) during the downpour. Most trated in the case of a circular column
of the ants running through the exposed starting around the edge of a glass jar.
area (i.e., the side away from the wall) not General observation supports the view that
only speeded up noticeably, but also tended gentle contact with another Eciton ahead
to turn short away from the side of disturb- elicits turning toward the side from which
ance instead of remaining in their respective the stimulation comes; from the same
concentric lines. Presently there were source and from special tests comes sup-
indications that the strong pressure effect port for the tendency of Ecitons to turn
thereby generated, concentrated upon the toward Eciton chemical, favoring the side
inner part of the circle when ants entered of greater concentration.
the drier area in the clockwise direction, For example, a large strip of bristol board
was literally pushing the less exposed part was prepared in an Eciton-chemical gradient by
of the mill toward the wall. By causing regulating the passage of workers across it to
have a maximum of crossings at one end and a
numbers of ants to run more widely out- minimum at the other, with graded frequencies
ward while passing through the marginal between. Then in succession 40 E. hamatum
section on that side, this unilateral pres- workers were released at the edge, 20 at the
sure produced an eccentric torque or "maximal-saturation" end and 20 at the mini-
mal-saturation end, and were permitted to
"twist" in the circling, toward the wall and cross freely. More than 15 of the former
counterclockwise (fig. 3). As a result of crossed at the same end of the card; whereas
this effect and its various secondary conse- the latter group crossed at points extending
quences through the group, the entire mill nearly the full length of the card. Thus the
representative response of ants entering the
gradually shifted its position at angles weakly saturated zone was to turn into parts
(i.e., in the counterclockwise direction, and of the gradient in which Eciton chemical- was
not straight toward the wall) farther under more concentrated.
the eave of the building. As new portions 1 Although the movement of the ring out of sun-
of surface closer to the building became light during late morning was not observed as it oc-
chemically saturated by ants forced out curred, it seems very probable that this shift depended
upon events somewhat similar to those described.
centrifugally on that side, the entire mill The distorting effect-in that case would depend upon
the irritating action of bright light upon the Ecitons,
could move into them, but most important, causing them to speed up when entering the exposed
it was the dynamic interrelationship of the area and turn away from the dazzle as promptly as
circumstances permitted. Such responses are com-
Eciton units in the mass that permitted monly observable in the everyday behavior of E.
praedator swarms which are partially exposed to bril-
the ring to retain its essential pattern liant illumination (e.g., in sun-fleck areas).
1944] CIRCULAR COLUMN AND ORIENTATION IN ANTS .15
The evident influence of the third factor, centrifugal force increasingly effective as
centrifugal force, is disclosed in figure 1, in running speed increases, Eciton paths
which one or more lines of ants may be seen around square or irregularly shaped objects
taking a path tangential to the rotating ring approximnate a circle more and more closely.
in the direction of turning. In the outer Through its effect, a hamatum path around
.portion of the ring, where the ants run a linear partition (influenced by initial

.e ~~~~N,
I.,

-9e4-1
N

/
/
/
/
/ If
I I
/
I;

-, %\
I
ol

/6& /11 /
< ~~~/} I
I
_- _,9

a r
X /,b

, -\e'"
I

I".
1

Fig. 4. Sketch to represent factors influencing the movement of ants in the E. praedator ring.
1, An ant moving on the margin: a, centripetal vector; b, centrifugal vector; r, the resultant eir-
cular course. 2, An ant moving in one of the inner lines: a, centripetal vector; jS, centrifugal vec-
tor, oy, pressure from outside and rear; p, the resultant circular course. (See text.)

most rapidly, the influence of this factor is contact with end walls) became modified
most directly apparent in behavior, in that into a parallelogram (fig. 2). In such cases
there is a persistent tendency for the the interaction of our second and third
marginal ants to "fray out" from the ring factors is recognizable: a centrifugal-force
in the forward direction. We have men- effect leading the ant to run onward after
tioned laboratory cases in which, with turning a corner, and a response to Eciton
16 AMERICAN MUSEUM NOVITATES [No. 1253

chemical when the centrifugal effect moving outer lines and adjacent slower
weakens. moving inside lines, seems to illustrate the
To take stock, ants in the mill are kept in involvement of a bilateral tactual pattern
motion in the forward direction by a pres- for ants within the body of the mill.
sure from others just behind, an effect Assuredly, centripetal pressure must not be
which is mutually exerted throughout the excluded here as a regulator of movement.
group. Upon this basis, the dynamic pat- We have noted that the circular mill ex-
tern mainly responsible for survival of the panded between 8:15 A.M. and noon. It
circular movement itself may be expressed will be recalled that a rise in temperature
most simply in terms of a combination of during this period would increase the run-
drainage and centrifugal-force effects. To ning speed of all ants in the group, thus
illustrate this interactive set of influences, heightening the centrifugal-force effect
let us consider the case of an Eciton moving tending to widen the circling of the re-
at the margin of the ring (fig. 4). Oc- spective concentric rings of ants.' But in
casional contact with ants moving along on addition, in accounting for the expansion
her left (i.e., inside, in the counterclock- of the mill, it seems necessary to consider
wise mill) together with a turning toward the possibility that "inside" ants, by exert-
the side of distinctive Eciton chemical, ing an increased centrifugal pressure,
would account for a centripetal vector in facilitated the wider turning of outside
her movement (fig. 4, 1 a). At the same lines; and that outside lines, in arcing more
time, depending upon running speed, her broadly, thereby reduced centrifugal pres-
momentum would account for a centrifugal sure and permitted freer action to those
force which would tend to impel her directly within.
ahead on a tangent with the circle (fig. 4, Now we may consider the Ecitons caught
lb). (As we have indicated, the effect of in the innermost line, near the hub of the
centrifugal force is directly evidenced at wheel. We have remarked the fact that in
times in actual tangential departures of early morning, near the end of the rain,
marginal ants from the ring, while the effect and toward evening a central cluster was
of the "drainage" factor is shown in the formed. At other times the central area
appearance of hesitancy and turning back was essentially bare, and when empty it
toward the circle after short tangential ex- was also larger. These were the times when
cursions of this kind.) The circular course higher temperatures accounted for a faster
of marginal ants may be considered the re- rotation of the mill as a whole, which
sultant of these two vectors (fig. 4, Ir). checks with our suggestion that the faster
Although the case seems more compli- motion of individuals increased the
cated for travelers on lines within the outer centrifugal-force effect and thereby spread
border of the mill, we may assume that the the ring. However, when the ring con-
same drainage and centrifugal-force pattern tracted, individuals in the innermost line
governs the circular course of all. However, were forced by crowding (i.e., lateral pres-
for ants moving in interior files, there are sure) of ants in the next outer ring to turn
indications that pressure effects are super- very sharply. This meant that some of the
added. In their case, the centrifugal-force central ants were buffeted and crowded so
influence (fig. 4, 2,B) is counteracted not forcibly that they were caused to circle in
only by a contact-chemical drainage effect place. Turning variously at different
(fig. 4, 2a) from ants turning away on the angles these ants interfered with one
left (i.e., "inside," in the present case), another, an organized circling of the inner-
but also by pressure effects through fre- most members thereby became impossible,
quent and at times forcible contact with and they presently were thrown into a
ants running on the right side (i.e., "out- cluster. Very probably the relative slug-
side"). The circular course (2p) may be 1 Shapley (1920) and others (Miller and Gans,
considered the resultant of these three 1925; Pratt, 1925) have shown for various species
vectors. One feature we have mentioned, that the speed of ants traveling an established path
increases with the temperature up to a given maxi-
the indications of "friction" between fast mum.
1944]- CIRCULAR COLUMN AND ORIENTATION IN ANTS 17

gishness of the ants at such times (i.e., in traffic at such points. We may say that
lower temperature) contributed to the Ecitons in the center of the narrowed
clustering tendency. Similarly, in normal praedator mill were passively tossed about
foraging, when the coming of dusk slows by a movement in which they could no
the Ecitons, the nucleus of clusters at trail longer participate as synchronized units,
junctions is formed by disoriented ants and thus in a sense assumed the role of
caught and knocked about in the hectic flotsam in the center of a whirlpool.

MILLING IN RELATION TO TRAILING UNDER NATURAL CONDITIONS

Instances of this kind in animal behavior narrower channel at distances farther in
are arresting not simply because of their the rear, is explicable in terms of gradual
bizarre character, but also because they changes in chemo-tactual response with
may be utilized to review the factors of further use of the ground. A turning of
normal behavior in relief, as it were. In Ecitons toward others, thus as a rule keep-
this example we have seen the operation, ing to circuitous columns and not spreading
under special conditions, of the principal out widely, is noticeable in the rear of the
factors in everyday Eciton mass inove- swarm where forces are reduced in num-
ment, i.e., pressure and drainage, together bers. The center of the area is favored pre-
with the centrifugal-force effect. Studies of sumably on a chemical basis. Because of
other Eciton species have shown how these the way the swarm moves ahead, spreading
factors are involved in the formation and out alternately on the flanks but with a
use of ordinary -raiding trails (Schneirla, fairly consistent crowding in the center,
1938). Furthermore, the same concepts there is evidently a reduction in chemical
make understandable the integrative proc- saturation from the center toward both
esses involved in the surprisingly direct sides in the area pillaged by this body.
advance of the relatively enormous E. Ants coming up in the rear evidently re-
burchelli swarm (Schneirla, 1940) and the spond to this gradient by turning more
smaller swarm of E. praedator. Despite readily toward the center of the wide lane
very complex and at first sight utterly than toward its borders. Their move-
random eddying and cross-current move- ments on the ground by favoring the central
ments among sub-groups in the swarm, by part increase the steepness of the gradient,
virtue of describable relationships of the thus provoking still more precise turns
pressure and drainage factors the whole centrally by ants advancing behind them
body of ants is capable of a fairly straight- in the fan. At the rear of the fan, where
forward advance through vegetation and the gradient must be steepest, all lateral
over broken forest terrain. In the wake trails disappear from use, as a rule, and
of every swarm there is a long fan-shaped only a narrow central route remains. In
network of columns (consolidation fan), the behavior of Ecitons following this
which narrows posteriorly into a single principal trail between bivouac and raid-
column communicating with the bivouac. ing front, there are many indications that
It is a relatively constant pressure, operat- the route is responded to as the region of
ing through this system and applied more or greatest intensity in a bilaterally extended
less evenly across the rear of the swarm by chemical gradient. A crowding away from
ants arriving from the bivouac, which the edges toward the center of the path is
enables the raiding mass to hold its laterally noticeable in moments of traffic congestion,
extended formation and at the same time especially when the route is new, and the
continue its advance. fastest moving ants generally usurp the
Why ants behind the swarm do not con- midline at the expense of burdened ones or
tinue to roam throughout the wide lane slower moving individuals such as the
saturated chemically by the swarm, but in- clumsy major workers. Displaced laterally
stead confine themselves to a progressively from the trail a few centimeters or more, in
18 AMERICAN MUSEUM NOVITATES [No.;1253

the ensuing rambling of the Eciton there is raiding system of any Eciton species is a
a noticeable drift toward the main path, complex network of chemical trails con-
even if it is empty at the time. Displaced necting with the bivouac over well-traveled
on the original ground, ants return to the trunk routes, providing slight opportunity
central route far more quickly than when for any section of traffic to operate inde-
transferred to new ground with the trail pendently for long without being modified
(e.g., a section of it on a narrow leaf) through its connections with adjoining sec-
located at a comparable distance. tions. For this reason, no part of a raiding
Now although Ecitons are undoubtedly system can remain very long in a given pat-
among the simplest of ants, psychologically tern of action. In the case of our praedator
considered, highly stereotyped in their mill the exception occurred, but there
activities, yet to the writer's knowledge a another unusual condition happened to be
true circular mill has been observed in only fulfilled: the ants were cut off from their
this one instance under natural conditions. colony when on or near very homogeneous
The burchelli or praedator swarm as it ad- terrain, the cement walk. Here the sub-
vances exhibits countless instances of vor- stratum was tactually neutral, presenting
texes, whirls, and eddying movements of no irregularities which could intrude upon
short columns in the mass. Also, pushing the symmetrical operation of pressure,
out from the forward and lateral margins of drainage, and centrifugal-force influences
the swarm are "pseudopodic columns" around the ring. Afield, few if any situa-
which frequently take eccentric courses, tions ever approach very close to neutrality
even looping back toward their origin in this sense. To be accurate, the cement
points at times. When the consolidation surface on which our praedator group
fan takes form on ground first swept by the chanced to rotate must, be considered a
swarm, in its complex anastomosis there are quasi-artificial situation.
always numerous remnants of circling We find essentially the same idea in
courses formerly apparent in the swarm Wheeler's (1928b) interesting discussion of
(Schneirla, 1940, fig. 1), which stand out polymorphic specialization in ants, applied
for a time as loops and circular trails in the to the worker major of the harvester
network. At times, on these circuitous Pheidole.
portions of the system, one observes that "Some authors regard the soldiers, the highest
given ants or trains of ants may travel one high-brows of our ant series, as monstrous, or
pathological forms on account of the excessive
or more rounds in a loop before there is a development of their crania. Certain facts
change. However, in these cases, as when might seem to lend support to such an opinion.
comparable circuits happen to form in the If the soldier of Ph. instabilis be placed on its
head on a perfectly smooth, hard, horizontal
tree-like trail systems of E. hamatum, the surface, the insect may be quite unable to right
complexion of events soon modifies the itself and may even die standing on its head.
circular path, so that true milling does not But this is a typical laboratory experiment.
In its natural environment the soldier never
occur. encounters such surfaces. Closer study shows
Actually, although all the chief factors that all these supposedly monstrous forms are
underlying our instance of milling are really exquisitely specialized and adapted for the
readily recognized in trail following under functions they have to perform in the life of
their respective colonies. The soldiers of the
natural conditions, circumstances operat- harvesting Pheidoles and Pheidologetons are
ing in the forest environment militate needed not only as seed-crushers, but those of
strongly against any possibility of Ecitons the latter genus have another very different
function. Several observers have seen groups
being caught in a persistent local circling of the minute Pheidologeton workers sitting
movement. Either the pressure of ants quietly on the huge heads of the soldiers and
entering over connecting trails changes riding to and from the nest. The soldiers of
the insect-eating Pheidoles dismember the tough
matters, or, if few newcomers arrive, the prey before or after it has been carried into the
shuttle trails provide escape routes. The nest... ." (p. 19).
1944] CIRCULAR COLUMN AND ORIENTATION IN ANTS 19

RELATIONS TO MILLING IN OTHER ANIMALS

PARR'S ANALYSIS OF MILLING IN SCHOOLING FISHES
After the foregoing analysis of the Eciton from the known properties of the animal.
mill had been sketched, in a search of the Of course the Parr analysis takes priority
literature for cases elsewhere in the animal as a7 theory of milling behavior in infra-

,~ ~ ~ ~ ~
series my attention was drawn to a similar mammalian animals; however, mention of
phenomenon in schooling fishes, described it has been delayed until this point in order
and analyzed by A. E. Parr (1927). A to emphasize the fact that a second in-
comparison of the Eciton mill in figure 1 vestigator, working independently with

Fi.5.Sethofte
icla is il esrbd yPar (e tx.

with the sketch of one level of Parr's fish very different material, arrived at a theory
mill in figure 5 indicates a marked re- which, notwithstanding the different set-
semblance in these patterns of invertebrate ting, bears a surprising similarity to the
and vertebrate behavior. Even more Parr "bio-mechanics" explanation.
striking are the similar outcomes of analy- The theory offered by Parr to account
sis by the two investigators. Each of us for the formation and maintenance of a
evaluates his phenomenon as essentially one milling group of fish may be represented in
in the mechanics of animal locomotion r6sum6 as follows. When a school of fish
in relation to available sensory patterns; (e.g., mackerel) meets a deflecting surface
each develops an explanation inductively and begins a sharp turn, individuals first
20 AMERICAN MUSEUM NOVITATES [No. 1253

to head in the new direction are presented on the same principles as those causing the
with a potent visual stimulus from those first circular movement of the central file.
directly opposite and traveling in the In each case, a similar turning of the ad-
original direction. This stimulation domi- jacent outward companions in the school
nates over that from adjacent fish which (fig. 5, 3, 4, and 5) prevents their stimulus
have participated in the turn and, are effect on the given individual from oppos-
swimming on the outside in the new direc- ing that of the adjacent interior file. The
tion. Since the stimulative influence of mill is maintained on the same basis:
companions swimming in the same direc- the circular movement of the central file
tion is already adjusted to, the new stimu- acquires the nature of a continuous auto-
lus is "much stronger" because it repre- matic process, and the tendency to turn as
sents a more rapidly shifting pattern. originally transmitted outward is main-
Hence fish on the inside of the wheel con-tained as each fish continues to respond to
tinue to turn toward the inside. The the inward turning of its adjacent inner
companion. Once the mill is formed, the
nucleus of a mill is thereby formed through
this mutual visual stimulation which outermost file exerts a certain condensing
influence upon the moving group, since the
keeps fish in the interior file turning toward
peripheral members are influenced solely
oppositely swimming fish (fig. 5, 1 and 2).
Soon individuals in files next the new by a one-sided attraction toward com-
center become influenced by the central panions turning away on the inside, and
individuals turning away from them in- thus act as a wall restraining any indi-
teriorly, and they respond by turning viduals happening to stray outward from
toward that side, thus beginning to wheel interior files.

A COMPARISON OF MILLING IN FISHES AND IN ARMY ANTS

These phenomena of circular milling in influence of a tactual "pressure" from be-
schooling fishes and in army ants are hind and from the next outer file in the
strikingly similar, as a comparison of the Eciton mill, not demonstrable in the fishes.
respective analyses above suggests. In Although from Parr's analysis this factor of
particular, in both cases a tendency to turn proximal stimulation would seem negligible
toward a dominant centripetally operating or absent in the visually based fish mill, it is
stimulation seems to be paramount. How- barely possible that it may acquire at least
ever, there are a number of differences a limited importance in the exceptionally
worthy of consideration. One is in the dense mills formed by the more tactual
dominant sensory mechanism of the milling, catfishes (Ameirus). When in schools, as
which in the schooling fishes thus far Parr (1927) states, young catfish are
studied (Parr, 1927; Breder and Nigrelli, spaced at intervals governed by the span of
1935; Shlaifer, 1942) has been identified as their tactile barbels.
vision. In connection with his study of This difference in the dominant sensory
mobile aggregations in the herring Jenkin- modality may be involved in accounting for
sia, Breder (1929) has pointed out that an apparent difference in the initiation of
fishes schooling on the basis of vision ap- mills in Ecitons and in fishes. As far as our
proach only as close to one another as the cases go, the circling in visually dominated
distance at which clear optic resolution fish appears to begin centrally and grow
occurs. At shorter distances a negative centrifugally, whereas through various
response is given. In contrast, the army relevant considerations we were led to re-
ants maintain their formation in depend- gard the praedator mill (not observed in its
ence upon a pattern of diverse tactuo- origin) as having begun peripherally and
chemical stimulative factors. This un- developed centripetally. Of course, it is
doubtedly admits certain differences in the quite possible that more extensive studies
detailed mechanism of circling, such as the may show that the mode of initiation, cen-
1944] CIRCULAR COLUMN AND ORIENTATION IN ANTS 21

tral or peripheral, is not an invariable mat- two-dimensional working surface of the

ter for either animal. Of far greater im- ants, in sharp contrast with the great
portance than this is the above-mentioned homogeneity of the aquatic medium of
similarity in the maintenance of the two pelagic fishes. A homogeneous environ-
mills. Both may be understood as cases of ment tends to accentuate and to facilitate
predominantly unilateral stimulation of in- stereotypy in behavior, by permitting the
volved individuals, with Ecitons turning in peculiarities of the animal to establish
stereotyped-manner to the side of light con- relatively rigid patterns which dominate its
tact and more effective Eciton chemical, behavior in given monotonous ways over
the mackerel toward the side of visual considerable periods of time. In contrast,
stimulation. notwithstanding the fact that the properties
Other differences between the milling of underlying Eciton behavior are in them-
these two animals deserve mention here. selves highly conducive to stereotypy
An obvious difference is that although fish (Schneirla, 1940, pp. 447 ff.), the varying
schools and milling when it occurs com- situations of a tropical-forest environment
militate strongly against the persistence of
monly operate in three dimensions, the monotonous patterns such as the circular-
Eciton mill is necessarily a two-dimensional trail formation. It is otherwise with fishes
affair which is more rigidly localized in one such as the herring, which not only possess
plane of space. This suggests a further the organic prerequisites for stereotyped be-
prevalent difference which very probably is havior, but also happen to live in an en-
important for the frequency with which vironment which admits the predominance
milling occurs under natural conditions, of these properties in given patterns for
that is, the relative heterogeneity of the considerable periods of time.

CONTRASTING PATTERNS IN ANT ORIENTATION IN RELATION TO MILLING

Actually the mechanism of Eciton Schneirla (1935, chap. 7) have termed "pro-
orientation, although complex in its own visional orientation," in that the animal's
terms, represents nearly the extreme of behavior in a given field of stimulation is
simplicity in ant way-finding. It bears a dominated in appropriate and plastic ways
close relationship to the stereotyped modes by circumstances which initiated the trip,
of orientation which characterize the in- and by a capacity to modify the route
vertebrates in particular, reaction types through previous experience in the situa-
which, although highly involved, have tion. For example, the common garden ant
been studied very successfully through the Lasius niger americanus runs its trails ac-
experimental examination of direct rela- cording to the prevailing circumstances of
tionships between the energy pattern of chemical, visual, tactual, and muscle-
stimulation and the dynamics of locomo- sense (kinesthetic) stimulation, improving
tion.' The Eciton system of orientation through learning in successive runs over
reduces essentially to a generalized chemo- the same general route. Since Lubbock's
tactually controlled pattern, apparently pioneer studies (1882), it has been known
without much learning involved beyond that Lasius niger follows its trail not on a
the initial approach response to Eciton chemical basis alone, but also according to
chemical. Contrasted with the cases of the direction of light rays. Although there
many other collective foragers, especially are grounds for believing that the Lasius
among the social insects, this army ant forager may learn to move according to
system appears highly rudimentary. these sensory patterns, there remain
Formicine orientation on its higher levels largely unanswered a number of questions
exemplifies the condition which Maier and concerning the ant's ability to adapt her
1 For a survey of evidence on these basic forms of movements to changing outer conditions.
orientation, see in particular the valuable work of One such question concerns how this ant
Fraenkel and Gunn (1940).
22 22AMERICAN MUSEUM NOVITATES [No. 1253

(together with a number of other species of pare simple collective foragers such as
similar foraging activities) is able to use a Eciton with true individual foragers such
given route through a considerable period as ants of Formica or Camponotus species.
with light direction one of her orienting These ants are capable of learning indi-
cues, notwithstanding a gradual change in vidual routes to foraging areas in different
the angle of the sun's rays through the directions from the nest, and can get
period. Part of the answer seems to lie through despite terrain changes (e.g., dig-
in their ability to utilize particular serial ging up the ground, laying down new sur-
tactual cues such as the edge of a boulder faces) which would completely block
or stick in a given section of terrain. Fur- Eciton or most collective-foraging ants.
thermore, in certain localities they may Field tests show that a Formica individual
learn to depend upon slope of the ground may set off from a food place toward the
as one cue, and Brun (1914) showed that nest (50 yards or more away), first travel-
Lasius adapts differently to such properties ing mainly in dependence upon a clump of
according to the nature of its relatively trees (visual stimulation of the ocelli),
broad chemical trail. Also, if the ant re- then pass through a low stretch of terrain
turns repeatedly from feeding on nectar, mainly in dependence upon direction of
there arises an end-for-end (i.e., polarized) light (stimulation of the compound eyes),
difference in the chemical trail to which she then shift mainly to chemical cues upon a
becomes able to respond (for example, she broad "ant road" a dozen or more yards
readily differentiates the two directions from the nest (getting through on a visual
after being displaced from the trail). If, basis if the ground has been disturbed),
however, larvae (Lasius larvae, experi- and finally reach the entrance to the nest by
mentally placed at a distance) are being utilizing tactual cues, slope of ground, and
carried back to the nest, she becomes more a chemical gradient in complex combina-
dependent upon the non-chemical cues tion (see Brun, 1914; Schneirla, 1929).
(e.g., direction of light) and becomes Iin the laboratory, Formica incerta sub-
seriously disoriented if light, tactual cues, jects are quite able to master a fairly com-
or slope of the path is changed. In con- plicated maze even when diffuse illumina-
trast, Ecitons, as we have pointed out, tion reduces visual cues to a decided mini-
follow their foraging trails mainly on the mum and chemical cues are rendered highly
basis of a generalized ability to move unstable through frequent shifting and
toward a common chemical product. In changing of the alley linings (Schneirla,
their case an ability to learn a particular 1943). Attempts to secure a mill phenom-
route according to direction of light or enon comparable to the Eciton praedator
other special cues existing on the given circular trail could never succeed with ants
pathway seems to be absent, as far as the of these species.
writer's tests have shown. Thus among ant species one finds a rela-
Thus the Ecitons represent the primitive tively wide range of foraging patterns,
type of collective foragers among ants, from a stereotyped following of mass
which follow their trails in a slavish, stereo- chemical trails to a highly specialized
typed manner without the involvement learning of individual routes largely inde-
of a special learning of the route, while pendently of any chemical tracking. Of
Lasius represents the more highly special- course, a detailed study of phenomena such
ized type of collective forager which is as the Eciton circular trail shows that even
capable of learning a route in dependence the "elementary" types of trail-following
upon prevailing sensory conditions. Under behavior are far from a simple matter of
appropriate conditions Ecitons follow a running along a line of chemical. However,
circular trail persistently, but we may ven- in Eciton species the milling phenomenon
ture to say that no particular success would discloses susceptibilities for stereotyped
meet an attempt to elicit such behavior behavior which are truly representative
when Lasius individuals are the subjects. of characteristic psychological limitations
The contrast is even greater if we com- appearing in their normal foraging activi-
 194]
1944] CIRCULAR COLUMN AND ORIENTATION IN ANTS 23

ties. We have ventured to predict the species of ants, mainly because of their
impossibility of obtaining a true case of capacity for variable behavior in way
milling behavior in more advanced collec- finding which largely excludes such activi-
tive foragers and in individual foraging ties.

CONCLUDING REMARKS
Studies of mass behavior phenomena the milling pattern may be a misleading
such as milling make their chief contribu- sign of psychological inferiority, since it
tion in the light cast upon the general may well obscure capacities for plastic,
problem of levels of integration in social opportune adjustments not dominated by
behavior. It is not the pattern in and for those native factors which account for
itself, but its relation to other properties of milling.
individual and group behavior in the same It happens that the army-ant circular
animal, which merits special investigation. column really typifies the nature of the
The milling pattern arises when animals animal, even though Eciton milling is un-
moving within the same area yield in com- doubtedly rare under field conditions. De-
mon to the influence of routine stimulus- tailed study of Eciton behavior under
response mechanisms arising through group forest conditions shows that the individual
locomotion. These characteristics mark is subject to fixed and prevalent, intrinsic
it as a very rudimentary form of group (i.e., native) limitations canalizing its be-
organization. In themselves, circular havior quite narrowly. However, in its
columns and other types of milling are complex environment, the animal is pro-
simply monotonous group activities of vided incidentally with innumerable possi-
little immediate adaptive significance save bilities for variation in behavior through
for the expenditure of energy, standing the heterogeneity of forest and jungle ter-
very low in the general scale of group- rain. Such diversity in surrounding stimu-
behavior patterns. lation permits a limited repertoire of
As for the eventual adaptive value of stereotyped individual reactions to func-
milling, in different cases it may be posi- tion in the group setting in highly adaptive
tive (e.g., by increasing protection from ways. Any beginning of monotonous be-
enemies), negative (e.g., by displacing havior, as we have seen, consequently is
essential behavior functions), or may have certain to be interrupted in some manner,
no special significance. The biological so that it becomes merely one of a series of
value of such group behavior, a problem of diversified events in the Eciton raid. Thus
considerable importance, must be worked it is environmental change that frees this
out in terms of the general life economy of animal from the direct consequences of its
the respective animal forms in which it native individual limitations.
appears. In the present connection inter- The frequency with which milling and
est attaches to the nature of milling and its similar rudimentary group functions ap-
relation to other behavior resources in the pear in any given animal must be held as a
given animal. fact of uncertain significance, until the
Analyses of milling in two animals have limitations of the animal and its potenti-
shown us that in such group functions the alities for new adaptive behavior are better
component individuals are dominated by known. For example, the fact that milling
action-engendered stimulus effects to which occurs with considerable frequency in
they respond in very reflex-like ways. schooling fishes is, of course, no certain
Studied in itself, obviously, this type of indication that they stand psychologically
group behavior does not necessarily pro- below army ants, in which we have found
vide us with a true picture of the given the potentialities for such behavior great,
animal's capacities for more versatile although milling itself is highly exceptional
adjustments under different conditions. under natural conditions. To be sure, the
It is apparent that the mere occurrence of capacities of some fishes such as the herring
24 AMERICAN MUSEUM NOVITATES [No. 1253

do appear to be rudimentary and highly subordinate aspect of herd behavior in

limited, yet there are other schooling fishes lower mammals.
in which milling occurs frequently, al- It is a matter of considerable interest that
though at the same time the resources for an animal as limited in its capacity for new
variable and complex behavior are learned behavior as the army ant seems to
definitely wider than in the herring (Breder be may largely escape the non-adaptive
and Nigrelli, 1935). consequences of great susceptibility for
On its face, then, simplified group be- circular-pattern mass activities only by
havior such as milling suggests merely a virtue of the incidental fact of environ-
capacity for a psychologically reduced be- mental heterogeneity. In contrast, the far
havior under the given conditions. With superior resources of mammals for trial
respect to the general problem of social and error learning are available, in addition
behavior, it is a reasonable proposition that to mere ecological variety, for admitting
the appearance of such elementary group appropriate group adjustments that largely
patterns may disclose psychological limita- hold the milling type of behavior in the
tions which are binding lifetime limitations, background. This should be the case in
or may indicate merely a special propensity their group activities in the natural en-
for routine mass activities under particular vironment, where opportunities to learn
conditions, with a possible capacity for new patterns are not subject, to the limita-
higher type adjustments when occasion tions that hold in barnyard and stock pen.
warrants. Unfortunately this principle is It may be observed that the point ap-
frequently neglected in sociological writ- plies even more emphatically to the social
ings. potentialities of man himself. Yet from
In mammalian groups, the milling type the point of view of what may be termed
of behavior has been featured in folklore the "herd instinct" school of sociological
presumably because of its spectacular theorists, certain writers (e.g., LeBon,
character, as in descriptions of massed 1917; Trotter, 1915), operating on a sub-
bison rushing in a frenzied stampede jective and non-experimental basis, have
from prairie fire, sheep jumping fences insisted upon the inevitability of emotional-
blindly in column, and other instances of ized, individually degraded, regimented
pell-niell surging by a horde of animals. patterns in the group behavior of mankind.
This scarcely represents the normal and We find Trotter carrying the notion of
typical state of affairs in an adequate way. "herd impulse" to the extreme of assuming
The scattered literature (Alverdes, 1927; that human subservience to the attitudes
Murchison, 1935) suggests that responses of the group is inborn, with the child
to the movement of other animals keep the instinctively responsive to what Trotter
group together and thus provide a basis for mysteriously calls the "voice of the herd."
its activities, yet the simple stimulus-and- This idea is comparable to that other hoary
response relationships of mass formation mysticism, the "gregarious instinct," which
appear to occupy a subordinate role in the accounts for animal aggregations by
ordinary activity patterns of any mam- postulating an inborn tendency to aggre-
malian aggregation. It is mainly at times of gate.
sudden and great excitement that the On whatever level of the animal series we
highly simplified miffing pattern dominates work, a scientific approach requires us first
group behavior to the exclusion of other to examine the causes of monotonous pat-
factors. For the most part, group activi- terns such as milling when they appear,
ties in lower mammals appear to be then to study their relation to the potenti-
plastically adaptive and rather finely ad- alities of the animal for other types of
justed to the given situation, as is evident behavior. From this point of view, it is
from careful studies such as Darling's apparent that the "herd instinct" school
(1937) work with red deer and Carpenter's has erred not only in evading the question
(1934) with howler monkeys. In general, of how such behavior originates, but also
the milling pattern seems to be a highly in underestimating the resources of higher
1944] CIRCULAR COLUMN AND ORIENTATION IN ANTS 25

animals for group versatility under new It may be observed that while army ants
conditions. Particularly in the case of man, are constitutionally susceptible to the pre-
when milling behavior becomes undesirable dominance of circular-column behavior and
or obnoxious, it is well to consider the can be freed from it only by the incidental
canalizing, straitjacket influence of social fact of environmental variation, man is
institutions fostering group tendencies for by no means susceptible in the same sense,
such behavior. An enlightening study of with his cortical basis for versatile cor-
such influences as prevalent in contempo- rective patterns which under encourage-
rary Germany has been carried out by ment may reduce milling to the minor r6le
Freeman (1940). of an occasional subway rush.

BIBLIOGRAPHY
ALVERDES, F. FRAENKEL, G., AND D. L. GUNN
1927. Social life in the animal world. New 1940. The orientation of animals. London,
York, Harcourt, Brace and Co. Oxford University Press.
BARNES, T. C. FREEMAN, E.
1930. The effect of gravity on the oscilla- 1940. Conquering the man in the street.
tions in the path of an ant (Lasius New York, Vanguard Press.
flavus nearticus Wheeler): A study FRISH, K. V., AND G. R6SCH
of random movements. Jour. Gen. 1926. Neue Versuche flber die Bedeutung
Psychol., vol. 3, pp. 318-324. von Duftorgan und Pollenduft fUr die
BREDER, C. M., JR. Verstandigung im Bienenvolk. Zeit-
1929. Certain effects in the habits of school- schr. vergl. Physiol., vol. 4, pp. 1-21.
ing fishes, as based on the observation LEBON, G.
of Jenkinsia. Amer. Mus. Novi- 1917. The crowd. Translation of Psycholo-
tates, no. 382, pp. 1-5. gie des foules. London, T. Fisher
BREDER, C. M., JR., AND R. F. NIGRELLI Unwin.
1935. The influence of temperature and other LUBBOCK, J.
factors on the winter aggregations of 1882. Ants, bees, and wasps. London,
the sunfish, Lepomis auritus, with Kegan, Paul, Trench and Co.
,
critical remarks on the social behavior MAIER, N. R. F., AND T. C. SCHNEIRLA
of fishes. Ecology, vol. 16, pp. 33-47. 1935. Principles of animal psychology. New
BRUN, RUDOLF York, McGraw-Hill Book Co.
1914. Die Raumorientierung der Ameisen MARSHALL, W. S.
und das Orientierungsproblem im 1904. The marching of the larva of the Maia
allgemeinen. Jena, Gustav Fischer, moth, Hemilenca maia. Biol. Bull.,
pp. 1-234. vol. 6, pp. 260-265.
CARPENTER, C. R. MELANDER, A. L., AND C. T. BRUES
1934. A field study of the behavior and 1906. The chemical nature of some insects,
social relations of howling monkeys. secretions. Bull. Wisconsin Nat. Hist.
Comp. Psychol. Monogr., vol. 10, pp. Soc., new ser., vol. 4, pp. 22-36.
1-168. MILLER, D. F., AND M. GANS
DARLING, F. F. 1925. Some observations on the reactions
1937. A herd of red deer. London, Oxford of the ant Cremastogaster lineolata
University Press. (Say) to heat. Jour. Comp. Psychol.,
DUBoIs, R. vol.5, pp. 465-473.
1899. Sur la soie de la chenille procession- MORLEY, B. D. W.
naire du Pin maritime et sur la 1942. Observations on the nest odours of
mani'ere de la faire filer au fur et a ants. Proe. Linn. Soc. London, 154th
mesure de sa production. Ann. Soc. sess., pp. 109-114.
Linn. Lyon, new ser., vol. 46, pp. MURCHISON, C. (ED.)
125-127. 1935. A handbook of social psychology.
EIDMANN, H. Worcester, Clark University Press.
1927. Die Sprache der Ameisen. Rev. Zool. PARR, A. E.
Russe, vol. 7, pp. 39-47. 1927. A contribution to the theoretical analy-
FABRE, JEAN-HENRI sis of the schooling behavior of fishes.
1903. Souvenirs entomologiques. Paris, Occas. Papers Bingham Oceanogr.
Delagrave, vol. 6. Coll., no. 1, pp. 1-32.
FIELDE, ADELE M. PRATT, K. C.
1902. Notes on an ant. Proc. Acad. Nat. 1925. Thermokineties of Crematogaster lineo-
Sci. Philadelphia, vol. 54, pp. 599- lata (Say). Jour. Comp. Psychol.,
625. vol. 5, pp.265-270.
26 AMERICAN MUSEUM NOVITATES [No. 1253
SANTSCHI, F. maze adjustment. Ibid., vol. 35, pp.
1911. Observations et remarques critiques 149-176.
sur le mecanisme de l'orientation chez [In press.] Studies on the army-ant behavior
les fourmis. Rev. Suisse Zool., vol. 19, pattern. Nomadism in the swarm-
pp. 303-338. raider Eciton burchelli. Proc. Amer.
1930. Nouvelles experiences sur l'orienta- Phil. Soc., vol. 89, no. 1.
tion des Tapinoma par s6cr6tions SHAPLEY, H.
dromographiques. Arch. Psychol., 1920. Thermokinetics of Liometopum apicu-
vol. 22, pp. 348-351. latum Mayr. Proc. Natl. Acad. Sci.,
SCHLAIFER, A. vol. 6, pp. 204-211.
1942. The schooling behavior of mackerel: THORPE, W. H.
a preliminary experimental analysis. 1939. Further experiments on olfactory
Zoologica, vol.27, pp. 75-80. conditioning in a parasitic insect.
SCHNEIRLA, T. C. The nature of the conditioning proc-
1929. Learning and orientation in ants. ess. Proc. Roy. Soc. London, ser. B,
Comp. Psychol. Monogr., vol. 6, pp. vol. 126, pp. 370-397.
143. 1939. Further studies on pre-imaginal olfac-
1938. A theory of army-ant behavior based tory conditioning. Ibid., ser. B, vol.
upon the analysis of activities in a 127, pp. 424-432.
representative species. Jour. Comp. TROTTER, W.
Psychol., vol.25, pp. 51-90. 1916. Instincts of the herd in peace and war.
1940. Further studies on the army-ant be- New York, Macmillan and Co.
havior pattern. Mass organization WHEELER, W. M.
in the swarm-raiders. Ibid., vol. 29, 1910. Ants, their structure, development,
pp. 401-460. and behavior. New York, Columbia
1941. Studies on the nature of ant learning. University Press.
I. The characteristics of a distinctive 1928a. The social insects, their origin and evo-
initial period of generalized learning. lution. New York, Harcourt, Brace
Ibid.,vol. 32, pp. 41-82. and Co.
1943. The nature of ant learning. II. 1928b. Foibles of insects and men. New
The intermediate stage of segmental York, Alfred A. Knopf.