Of the Ants Formica Exsecta Nyl

ACTA NEUROBIOL. EXP. 1973. 33: 597-622

ETHOLOGICAL STUDIES ON POLYCALIC COLONIES OF THE ANTS FORMICA EXSECTA NYL.

Janina DOBRZAfiSKA

Department of Neurophysiology, Nencki Institute of Experimental Biology Warszawa, Poland

Abstract. The members of a polycalic colony of Formica exsecta freely walk over its various nests, while performing social functions. It seems that the external service in the species ceases to belong strictly to one nest and forms a supra-nest attachment. The probability of this relation is more credible in the light of the results of this research which disclosed a tendency to local migrations. For, both these ethological peculiarities require an inhibition of the attachment to the re- stricted habitat, which seems to be characteristic of the species of the genus Formica. It seems that we are dealing here with a process of colony integration, and that we witness the extention of the social unit from the nest into the whole colony.

INTRODUCTION

Many species of ants, including those of the genus Formica, build multi-nest colonies, knlown in myrmecological literature as polycalic ones. Such a colony is, as a rule, developed by splitting from a parent nest (9) and, therefore, all nests in the colony are of common origin. A com- plex of families of ants, which preserve their mutual ties by common feeding areas, common paths and common breeding (of aphids i,s named a multi-nest colony. This phenomenon has been confirmed in the case of the Formica by Kiil (14), Dlussky (4, 5) and Marikovsky (16). Mari- kovsky (16) and Dlussky (3, 5) consider the inter-nest exchange of brood, which in their opinion certainly takes place in the fall, as an indispensable feature of the colony. Inside the colonies there also occurs an inter-nest exchange of adult individuals. It is continued often during the entire period of summer vegetation. This phenomenon was described by Huber (13) in F. pra- 598 J. DOBRZANSKA tensis and by Okknd (17) in F, rufa. okland was the first \to confirm that inter-nest exchange of the inhabitants must have an influence on integration ,of the colony, preserving the mutual ties between thfe families. Besides the exchange of individuals between the parallelly existing nests of the colony, the phenomenon of migration, i.e. removal to another habitat is also known. Bath the exchange and the removals are performed in the way of a mutual carrying, ,as was already described by Huber (13) and Adlerz (1). The up-toidate description [of this phenomenon is included in the monograph written by Wilson (19). The ant-workers engaged with removal carry in their mandibulae the brood, the sexual individuals and their own fellow-workers. Arnoldi (2) was the first who found that the consistent postures assumed by the ants being carried show differences characteristic of taxonomic groups. The formation of polycalic colonies is, as a matter of fact, the first example of the remwal since splitting from the nest is such a partial removal. During a period of prosperity and, consequently, an increase in population 'of the nest, a drive for establishing a new nest and oarry- ing to it some of their fellow-ants and their brood is developed in some workers. Transfe~redto a new place, the workers return, however, in great numbers to the parent nest, frequently carrying the brood. This was described by Huber (13), Fore1 (10) and Wheeler (18). The process of removal is, therefore, very long, ,sometimes continuing throughout the entire period of summer vegetation. It was already Huber, who with his great penetrating capability of observation, concluded that the ants which found a new nest and are first !to move from the old one, are the most excitable individuals, the ones which are most responsive to certain stimuli. In the case of splitting, these are probably the stimuli connected with the density of population which increases in the parent nest. The drive evoked by these stimuli has to (overcome the conservatism, that is, an attachment to the ants' old habitat which is characteristic of the species of the Formica (6, 7, 18). Precisely this conservatism compels less excitalble ants which were passively carried to a new nest to return to the parent nest. The phenomenon of complete migration is also known. It may be caused by factors which (are harmful to the community and which compel to migrate from the previous habitat. Such factors may occur iln the form of sudden cataclysms as, for instance, a flood, an abtack by strmger neighbors, etc. These situations compel the ants to perform an instant escape as they pose a mortal danger to the lingering members of the community. However, there also happen factors which force the ants to a complete migration such as unfavorable meteorological conditions, a displacement from the feeding area by strmger neighbors, an invasion POLYCALIC COLONIES OF ANTS 590 of parasites, or other adverse conditions, which increase gradually. If such is the cause, in various individuals reactions may be released which depend on their individufal excitability much the same as on the exces- sive density of population. Then, the migration process resembles the first type of removal caused by an overcrowded nest, when there occurs a conflict of contradictory tendencies: some individuals are busy with the removal and some others attempt to return to the old habitat. A removal, revealing the existence of external phenomena harmful to the entire community and which cannot be overcome m the spot, may be recognized by, among other things, the o~bservation of the final effect: regardless of the duration of the removal it leads to the abandonment of the old nest and to the migration of the entire community to a new habitat. The described cases of a complete migratioln among ants of the genus Formica are rather rare.

I. INTER-NEST RELATIONS IN A COLONY

Material and method Mutual relations between particular families of a polydic colony of Formica exsecta were studied in 1967 in a 13-nest colony, henceforth called C13, the plan of which is given in Fig. 1. The colony covered about 300 m2 of a grass-covered clearing in the Solska Forest. The pre- liminary olbservations revealed that four nests (i.e. NV, NIX, NX and NXII) were uninhabited by ants at the beginning of the experiments. To study the movement of ants in the area of the colony, all individuals, regardless of their functions, were first marked with dyes identic-

NVl NIX N1@ NII. NY NX NU[@ NN* 25

Fig. 1. Sketch of colony C13. Dimensions of circles indicating the nests are in proportion to their actual size. ally on one of the nests (NVZ) as the were walking over it. Even pre- liminary results have revealed the necessity d an additional marking of the ants involved in definite social functiolns in particular nests (specific- ally in nests ZZZ and XIZZ). The building function has been chosen as most appriopriate fw this purpose for the following reasons: (i) it does not leave any doubts as to which of the nests is performed; (ii) the building activity of workers is confined to the dome of the nest where they are distinctly visible and, consequently, the possibility of committing errors in determining the number of marked wo~kersis reduced to a ba~eminimum; (iii) the errors concenning quality are also ruled ~utin practice, that is, the building function cannot be ascribed to an ant which does any other work since building material transported by a worker is on the whole easily recognizable; in rare cases when the observer is in dmbt, it is quite enough to watch the use made by the ant of the object it is transporting to make sure that it is a building material. If necessary and practicable, the shape of the marks of some ants was recorded individually. In the present paper a nest oln which a definite group of ants has been marked is called an AMN (ants-marking nest) and the remaining ones - other nests (in relation to the former group). The observations discussed here were started withiln 2 days after ant-marking had been completed. Results According to the experimental method described above, 254 individuals were marked, regardless of their functions, on nest AMNVZ, and 224 building ants on nest AMNZZI. In the next stage of experiments, all nests of the colony were observed and the presence of marked individuals on them was recorded. Within a fortnight, of 203 recordings olb- tained from both groups 199 (98O/o!) came from other nests (in relation to each group). On om AMN's there were only four recordings of marked anlts. In view of the fact that all individuals met on AMNVI were marked, we can ascertain that those were ants performing various functions of external service. The numbers cited above give ample evidence that the ants marked on the two different nests, regardless of their functions, moved in large numbers to other nests of the colony. Figure 2 presents a comparison 01 the mobility of ants from the two AMN's: on the left - movement from AMNVI and on the right - from AMNIZZ. The situation of each nest in this Figure depends on its distance (see Fig. 1) fmthe AMN's (proportions have not been exactly pre- POLYCALIC COLONIES OF ANTS

, , , , , , , , , , , , , I"""' F3l - 44 40 36 32 28 24 20 16 12 8 4 0 0 4 8 12 16 20 24 28 32 36 40 44 48 52 56 Observaflons Observof,brrs Fig. 2. The distribution of newcomers from the AMN's; those from AMNVI on the left, from AMNIII on the right. Horizontal axis, total number of recordings from all observations; hatched areas indicating the number of the individual engaged in building activities. sented). As revealed by the comparison of both sides of this drawing, the character of the movements to other nests is different in the case of the newcomers from particular AMN's. Each of these groups was moving in its own manner, preferring different nests and in different numerical proportions. Thus, for instance, 82'010 of all recorded arrivals from AMNIII were found on NI and only 14010 of 'those from AMNVI. On the other hand, AMNIII, which was left by all the ants there marked, received a large group of 17 ants from AMNVI, to which in turn only one individual came from AMNIII. The distance was nolt, therefore, a decisive factor in the preference !of particular nests by the arrivals from both AMN's, which clearly follows from the topographical conditions (see Fig. 1 and 2). The size of nests also did not play any role: the ants were walking from AMNVI to the smaller nests NVIII and NVII; while the marked workers from AMNIII preferred NI, considerably larger than AMNIII. As was mentioned above, the nests NV, NIX, NX and NXII were uninhabited when these experiments had started, and nest NXI was being abandoned during the observations. Thus Fig. 2 indicates that the interchange of ant-workers took place among all inhabited nests. A remarkable fact was that the newcomers took part in building work at a new nest. As mentioned above, the building functions has for methodical reasons only been selected from other functions for the purposes of checking whether or not newcomers participated in the social activity in a new nest. The results obtained give an unequivocal affirmative answer to this question. The fact that a larger number of building ants marked on AMNIII undertook this function rather then the ants marked on AMNVI (see Fig. 2), stems from the fact that on the last-named nest all of the individuals met there were marked, and, therefore, these ants could perfmomed each of the functions belonging to the external service. Consequently, the number of building ants among them was considerably smaller. Zt was necessary to check the general character of the phenomenon of the same individuals undertaking work on various nests of the colony. For this purpose still anlother gmup of workers was marked. This group participated in work which was done on a new, just constructed nest, AMNXIII. Due to the small dimensions of that nest I succeeded in marking only 31 indivilduals including 23 building-workers and 8 ants which participated in removal from NXI to AMNXIII. All ants of this group, 20 in number (including 12 building workers and 8 ,busy with transfer) observed afterwards, stayed and were active only on AMNXIII. None of them was found on other nests. In addition, 1 newcomer from AMNIII and 3 from AMNVI were met on NXIII. All 4 performed a building function on their new nest. Particularly noteworthy is the fact that the most intensive engagement in work was recorded on AMNXIII, since 75Oio (that is, 15 out of 20) of the individuals observed were performing building function, including 5 out of 8 ants which had formerly been carriers during the removal.

Fig. 3. The distribution of the ants, marked (on the 1st and 2nd day of observation) on nest NVZ of colony C13. Vertical axis, the successive days of observation; horizontal axis, number of recordings on particular nests of the colony on each day. POLYCALIC COLONIES OF ANTS 603

As a result of this removal nest NXI was definitively abandoned by ants. This increased the number of uninhabited nests tlo five. Figure 3 shows the course of the described processes exemplified within the group of ants marked on the basis of their presence on the AMNVI. The chaotic character of the movement of the ants was striking. They simultaneously disbanded into several nests and after some time often returned to a previously frequented nest. It could be supposed that the maked group of ants had been secondarily divided into subgroups, some on which moved to a different nest in the colony, and another stayed in the previous nest. However it is not consistent with Table I, which shows an example of the behavior of a few individually

Examples of repeated records of the same individual in colony C13. Roman numerals denote serial numbers of nests; b denotes the engagement in nest construction

Place Ant I Place of recording i of marking

I AMN VI N VI, N VI, N VZI (b) AMN VI N VII, N VII, N VIII I AMN VI N VI, N III, N III AMN VI I N Ill (b) N III (b),

I AMN III (b) N 1(b), N I (b) AMN 111 (b) h' I (b), N I, AMN VI N VII, N VII, N VII, N VIII (b) Aphids, N XIII (b) Aphids, N I identified workers. It indicates that the same individuals appeared suc- cessively in the different nests and some of them changed a nest twice a fortnight. Their presence in these nests cannot be considered as an accidental stay connected wiih their Iorage in a territory, since they become engaged there in social works. One can also observe that the aphid-milk worker can supply honey- dew which they collect to various nests of the colony.

The firs't inferences It has thus been found, that the members of a colony of Formica exsecta walked freely over various nests (of this colmony, performing social functions on each of them. Thus the inter-nest connections here are tighter than those previously observed among colonial ants. It seems that the external service in this species ceases to belong strictly to one nest and becomes a supra-nest attachment. Thus, there an advanced form of integration of colony talkes place. The supra-nest attachment of part of the inhabitant of a colony must therefore be a counteraction against the process of isolation of the separate families in each nest and has to reinforce the mutual ties between the families. We can imagine, that we witness the extension of the social unit - with all aspects of social life - out from the nest into the whole colony.

11. MIGRATIONS IN F. EXSECTA

The statical character of the studies described above was their fundamental shortcoming. It has been found that a certain number of ants recorded on one of the nests of a colony was observed after sme time on another nest of this colony. No data were, however, obtained ccm- cerning the ci~cumstancesof this movement and no processes involved in it were revealed. The observation of the ants' behavior during the movement between the nests was precluded by the densely overgrown terrain of colony C13 (except for a small area around NXIII). Thus, it was necessary to undertake similar studies on another colony whose terrain, different in character, would permit a close and permanent observation. Such studies were undertaken in 1971 in a forest at Bialoleka Dworska near Warszawa.

Material and Method

A colony, originally consisting of three nests and, therefore, called by the working name d C3, became the object of these studies. It was situated on a small clearing of a young planted pine forest. The clearing was overgrown with a sparse heath and short sun-dried grass which provided convenient conditions for observation. It should be mentioned that the summer of 1971 was exceptionally dry in Poland and, consequently, unfavorable for the existence of ants. This circumstance is of a certain importance to a search of the factors which stimulate the phenomena described. Workers which were presented rn the nest N1, were marked and then their presence and the functions they performed on and between other nests were recorded. In colony C3 it turned out that the appearance of these ants on another nest was, on the whole, a sign of their migration to this nest. Thus, the ants taking part in each successive removal were differently marked depending on the function performed. All wodcer ants, marked as they performed the same function, are called a functional group, with the name of a function added. Thus each ant of the POLYCALIC COLONIES OF ANTS 606 same group had performed a given function at least once during some of the removals (and was marked on this time). The following functional groups have been obtained by this method: Group 1, building a new nest; Group 2, carrying fellow-workers during removal; Group 3, workers carried during removal; Group 4, carrying bmd (larvae and pupae) during removal. The last three funmctions are performed only during the process of removal. On the other hand the first one can be performed before the removal has begun. As it was defined each functional group includes the ants marked during various removals. A pupconsists, therefore, of a few subgroups. The name of each subgroup includes, in addition to the name of function, the specification of a given removal. (For instance: "the subgroup carrying workers to N3"). As follows fmall the considerations presented above, the groups differed from each other in respect of number and were marked during different periods and under various conditions. To compare the data obtained from observation, it was necessary to reduoe to a bare minimum all methodical errors unavoidable under such circumstances. In doing so, the following sources of errors have been taken into account: 1. Particular functions were performed by unequal numbers of individuals and the ants performing various function could not be equally easily marked. In addition, functional groups were marked at various moments and marking even a particular subgroup might have taken a few days. For this reason on each day of observation I was occupied with varying numbers of individuals in a given group (or subgroup) and with varying quantitative pmportions between groups. To make observation data comparable, the numbers obtained for each group (or subgroup) and for each day were converted in relation to the number of individuals marked on a given day. Relative numbers obtained in this way were used for the graphical elaboration of the results. Consequently, the diagrams express a daily mobility of each of the functional groups, ir- respective of an actual number of ants performing this function and the number of individuals currently rnariked in a given group. The diagrams are not, therefore, an expression of numerical ratios but they merely illustrate further (statistical) activities of each functional group or subgroup within the colony. 2. Observations were conducted daily on each nest displaying activity at the moment. For technical reasons it was impossible to observe all nests at the same time. On the other hand, the activity of Formica exsecta depends very much on the time of day as well as on only fluctua- tion however small, of the weather. To reduce the errors resulting from the differences in time to a bare minimum, the number of individuals observed on each nest was recorded each quarter of an hour during the entire period of activity on the nest on a given day. The sum of these all-day data demonstrates the daily frequency of the ants activity on a given nest. 3. Each functional group was observed for a long time, that is, for a few weeks to 2 mionths (depending on the time of its marking; and, therefore, the possibility of obliterating the dye used for marking ants had to be taken into raccount. From the viewpoint of accuracy, this is unfortunately the greatest drawback in determining the results. For the durability of a dye 'depends on many factors elusive in practice. These could be for instance, the place on the ant's body on which the mark is painted; the thickness of dye used which steadily changes (exposed to the sun the whole day, the dye thickens and should be thinned down now and then); maybe also on the ,speed of drying, that is, on the am- bient temperature at the moment of marking; and finally, on it, how often a given ant cleans itself or is cleaned by its fellow ants both of which may be accidental and depend on considerable individual vari- ability in the 'behavior of an'ts. My o'bservations in this field have shown that the first symptoms of the abrasive wear of the dye used in these studies ("Wilbra", a dye for leather) occurred after 3 weeks and after 2 months barely visible traces of dye were sporadically observed. The rate of the obliteration of marks is, of course, not uniform. It increased day after day after the appearance of the first symptoms of the erosion of the dye. However, there was no practical possibility of calculating the abrasive wear of marks on each day and even if it were, it would require a certain approximation. For these reasons, )on the basis of my observations I a'dopt the following criteria: I assume that 3 weeks after marking, the coefficient (of practically visible marks amounts to '/a (that is, I assume that 'I8 of the ma& is worn), after 4 weeks - a/,, after 5 weeks - 4/8, after 6 wee'ks - 11, and after 7 weeks I assume the marks as non-existent 1. This principle is not, of course, quite accurate and causes certain jumps in numerical data. However, in combination with the above-mentioned system of marking for several days running and with the method of elaborating the results in relative numbers, the errors decrease considerably, whereas the data I take into account in drawing conclusions seem to be so distinct that even a considerable methodical error may ,be dis- regarded. Obviously, this does not entitle one to use this same color of mark for another functional group; under such circumstances even a single mark preserved may cause a fundamental error. POLYCALIC COLONIES OF ANTS

Results The history 'of the colony C3 As mentioned above, at the beginning of the second part of these experiments, the colony was composed of three nests. Three others were subsequently built during various periods. The order of their conskuc- tion is shown on the plan (Fig. 4). The original three nests (Nl, N2, N3) were followed in succession by N4, N5 and, finally, N6.

7m 4 4 494- @ - birches 4 -pine grove

Fig. 4. Removals in \colony C3. Dimensions of circles indicating the nests are in proportion to their actual size; their shading darkens with the time of the existence of nest: the younger the nest, the lighter is the color of a circle; black circles denote nests which existed when the experiments were undertaken. Con- tinuous arrows indicate the direction of removals, dashed ones indicate those of abortive attempts to found or inhabit new nests. Dates over the arrows indicate the duration of migration activity from the first attempts to make an old nest habitable or to found a new one.

Without detailed observations, it might seem that during the summer the colony increased considerably, since instead of the original three- nest colony we ultimately had a six-nest colony. As a matter of fact the true situation was somewhat different. The appearance of the three new nests was by no means a symptom of the development of the colony. It had already been revealed in the initial observations that the entire population of the colony currently inhabited only one of the then existing nests, namely N1; while the remaining two were empty. On July 21 and 23 an attempt was undertaken to make N3 habitable but it resulted in a failure after 3 days and no acts of removal t~ this nest

5 - Acta Neurobiologiae Experimentalis 608 J. DOBRZANSKA took place. On August 3, a removal was started from N1 but the ants did not move to either of the two which already were in existence. Instead, before the removal, an intensive building activity was undertaken on August 1 and 2 at an entirely new place 8 m from N1. This resulted in the construction of N4. The removal of the colony to the last-named nest twk place on August 3, 4 and 5, the latter being a day on which the rest of the ants remaining in N1 were carried to the new nest and N1 was finally abandoned in that summer. On August 6, the day after this removal was completed a migration was already started towaxds N3, a direction opposite to the previous one. This removal took 5 days and was mpleted on August 11 2. During that same period between August 9 and 11, a new attempt was undertaken to found another nest, N5, in a place situated still farther east sf N4, i.e. in the opposite direction to N3. This attempt was, however, overcome, by the tendency to move to N3. As a result both N4 and N5 were finally abandoned on August 11. However, after the completion of the removal, the colony also in N3 stayed for only 1 day (Aug. 12). On August 13, while the building wdk was still in progress on this nest, construction was started in an entirely new place (N6) and was so vigorously conducted that as early as the next (day the brood could be carried to the new nest. Late on that day N6 was equal in size to the old nest, N3. I have never before obse~vedsuch a rapid and intensive rate of construction. Later on, two other attempts at founding a new nest, were observed near N6 viz. in N7 (on Aug. 16 and Sept. 1-3). Both attempts failed so that up to the end of the season the colony stayed in N6 and hibernated there. Figure 5 was composed for colony C3 by the same method as Fig. 3 for colony C13. It shows the movements of the ant group marked on the basis of their presence on the nest N1. The picture which we observe in Fig. 5 is, however, highly )different from that in Fig. 3 because the removals in colony C3 described above were always completed rapidly, i.e., they led to the abandonment of the previously inhabited nest. The futile attempts to migrate were always short-lived and only a few individuals todk part in them. Except for the abortive attempt at founding simultaneously with N4 also nest N5, the ants istayed in no more than two nests simultaneously. Their presence was always connected with the actual process of removal. During the period of stabilization

? The day of August 8 has been left out of the observations and recording, since on this day the ants were prevented from any activity by unfavorable meteorological conditions. POLYCALIC COLONIES OF ANTS

Fig. 5. The distribution of the ants, marked (between Aug. 25 and Sept. 2) on nest N1 of colony C3. Vertical axis, successive days of observation; horizontal axis, number of recordings on particular nests of the colony on each day. The figures, which correspond with days of a removal to a given nest are shaded. the whole mlony inhabited only lone nest. It indicates that, despite of appearamces, C3 was a monocalic (one-nest) colony. If me examines the plan of the colony in Fig. 4, he may notice that the nests N1, N3 and N6 were situated in the clearing described above; and, therefore, their microclimatic conditions were fairly similar. Quite different were the conditions and character of N4 and N5. They were situated outside the clearing and screened and shaded by a young pine forest. They differed in structure from the remaining nests in a very important way, namely, both were constructed without earth work. A mound in the form a roof was erected above an intact surface of moss. Pupae were laid on this moss.

The process of migration in colony C3 Now, let us analpse in detail the behavior of ants during removals. A sum of the ,data on all the ants marked (except flor the broad-carrying grwup) between August 5 and 23, that is, during the entire period of migrations and the first days of stabilization (in nest N6), is shown in Fig. 6. The Figure does not, therefore, depict the behavior of particular individuals or their groups, but that of the mlony as a whole. I would like to call the reader's attention to the fact that Fig. 6 differs from Fig. 2 in its arrangement. The figures in its center do not designate the serial numbers of particular nests as in Fig. 2, but suc- Fig. 6. The participation of all marked ants in removals. Horizontal axis, relative number of recordings; vertical axis, observation days. Arrows on the numbers of observation days indicate an occurrence of removal on a given day and the direction of movement. The length of horizontal bars indicates a relative number of the individuals marked as recorded on a given day. Brackets on the margin give information as to which nest corresponds to an actual activity of the ants marked by a given bar. The type of this activity is shown by the tinge of a bar: the hatched parts indicate participation in building, the shaded parts indicate carrying the workers, the checquered parts indicate carrying the brood. cessive days of observation. Thus each horizontal bar of figure indicates a relative number !of ant recordings on a given day. Brackets put on the margins of the drawing indicate the nests to which the activity of individuals shown in particular bars should be referred. Arrows drawn on the numbers of the observation days point the 'direction of a current removal. (For the sake of clarity, the drawing does not contain the futile attempts described in the historical part which did not lead to the migration KX€ the colony.) The la& of arrows on the numbers re- presenting days is an indication that the colony actually stayed in a state of stabilization. Starting the analysis of the drawing, we primarily find that during the period of rem'ovals (days with arrows) the general activity of the colony expressed in an increased mobility of its members is ccmsiderably POLYCALIC COLONIES OF ANTS 611 more intensive than during the period of stabilization (days without arrows). It also turned out that the frequency of workers on the nest about to be abandoned could occasionally be as intensive as that on the new one. This is quite obvious since the active workers engaged in the removal of the colony walk to and fro, carrying all their belongings, and this, in turn, is bound to stimulate the ants remaining on the old nest. However, disregarding the arrows, it is sufficient to look at the bars depicting the ants behavior on each day in order to distinguish the direction of migration. The 'building activity gradually decreases oln the nest which is about to be abandoned and increases on the new nest. In addition, the carrying of the belongings also takes places in the direction of the new nest 3. Thus, the appearance of ants on the old nest is merely an expression of a general excitation caused by the migration. Although a considerable number of ants stays on the old nest, activity of the entire society is transferred to a new nest. The following case contradicts the rule described above. On the 2nd and 5-th days a small number of ants carried into the abandoned N4 was recorded. It would not be the result of the experimenter's inter- ference 4 since the ,action of building of the 8bandoned nest N4 was resumed on the 7th day. No case has ever beep recorded d such carrying of the belongings during other removals (from N1 and from N3) to an abandoned nest although the experimental conditions and consequently the chances of committing errors were the same. This shows that we actually have to cb with real facts of a movement opposite to the direction of migration. An interpretation of these facts are discussed in detail later. Now, let us analyse in turn the behavior of particular functional groups during migraticms. The behavior of subgroups marked during the removal to N3 which was displayed by them in two consecutive removals, that is to the same N3 (line A) and to N6 (line B), is compared in Fig. 7. This comparison is very interesting since it illustrates the behavior of ants when they abandon a nest (N3), for which they had previously shown an preference by making a removal to it. Analysing the drawing we primarily observe that each of the functional subgroups under shdy once again todk an active part in the second removal (Fig. 7B). Next,

3 The ants leaving the old nest and carrying their belongings towards the new one are, of course, recorded on the drawings on the score of the new nest. 4 It is very easy to make a mistake in this type of observations because the carrying workers become particularly timid. Even a slight, careless movement of an observer may cause these ants to hide, to change their direction and turn back. At such moments, it is very easy to commit an error in recording the direction of their movement. Fig. 7. Participation of particular functional groups in two removals in succession with consideration of actually performed operations. Functional subgroups, marked at the beginning of a removal to N3 (see Method, Part 11): Subgroup 1, building a new nest; Subgroup 2, carrying workers during removal; Subgroup 3, workers carried during removal; Subgroup 4, carrying brood during removal. The areas of circles are in proportion to the number of the individuals recorded (an absolute number of recordings is indicated by numbers in the middle of circles). Part- icular sectors indicate a relative actual (during observation) participation of the individuals recorded in particular functions: hatched sectors, participation in functions of building; shaded sectors, participation in carrying workers; checkered sectors, participation in carrying brood. Numbers within sectors indicate the percentage of ants performing a given function as compared with the total of the individuals observed (or part of the entire sector which expresses 10O0Io). Row A (a removal to N3) illustrates further behavior of the individuals marked during the same removal, which was possible due to a full 5 days taken to effect this removal. Circles in row B (removal to N6) are smaller than those in row A, since, during field studies when only a limited number of the individuals (rarked) are used, the number of recordings decreases with time. we find that subgroups marked during active functions (that is, Sub- groups 1, 2 and 4), mainly performed the same functions during the next removal but also were partly concerned with other functions. Except for building workers of Subgroup 1, which are not employed at 'brood- carrying, other subgroups may perform each function, connected with the migration. Especially characteristic is the fact that the Subgroup 3 of carried workers promptly began to participate actively in this same removal. The comparison of this Subgroup with the worker-carrying Subgroup 2 reveals an astounding similarity in their behavior, that is, sharing of different functions by the two subgmups during the second removal was quite similar to each other (Fig. 7B, Subgroups 2 and 3). POLYCALIC COLONIES OF ANTS 613

In the first removal (Fig. 7A, Subgroups 2 and 3) a certain increase in brood-carrying ants at the cost of worker-carrying ants was observed in Subgroup 3, which is quite understandable in the individuals which so far remained inside the nest. Except for this detail, the distribution of functions was also similar. Finally, all the four functional subgroups displayed during the second removal an increase in their involvement in the building function. This might be expected because N6 was newly built and N3 existed previously. Also an abortive attempt at making N3 habitable had taken place not long ago, and had been manifested by the performance of certain building operations. Thus, this nest needed considerably less building activity than N6. It was remaukable that the increased building activity during the second removal did not take place at the cost of other functions indicated in the drawing (that is, connected with the migration). The reason for this is that there was an increase in general activity, expressed in the ants mobility anld functions. The white sectors, which indicate marked individuals apparently taking no active part in the removal, are reduced here in comparison to the first removal. This is confirmed by Fig. 6 and 8 which reveal that the colony's activity increased suddenly dluring the migration and afterwards dropped once again. During the period of stabilization in nest life (after Aug. 17), the activity (keeps at a low level, not dropping to the zero point except for a few days with particularly unfavorable meteorological conditions. The question arises: what part does each functional group play during such normal, stabilized activity of the colony? Figure 8 has been drawn to answer this question. It differs from Fig. 6 in showing separately each functional group and in presenting only the stabilization period between August 17 and September 4. It reveals that the first two groups, that is, the builders and the carriers, remain on the surface; while only a decrease in mobility is recorded for the ants of these two groups. Hence, we may condude that these groups belong to the external service and that each migration intensifies their mobility and concentrates their activity in only one direction. The individuals of Group 3, which were passively carried at the beginning of the removal were also noted on the surface during the period of stabilization. However, this gmp was (in contrast to the two previous ones) very small, namely it consisted of 24 individuals. Con- sidering that the numbers in the drawing represent the relative values, each bar in Column 3 represents from 1 to 3 individuals. Almost all (20 to 24) of the ants marked in Group 3, were identified and, consequently, we know that the observations recorded in Column 3 of Fig. 8 Group 1 Group 2 Group 3 Group 4

Fig. 8. The activity of particular functional groups in nest N6 during the period of stabilization. Vertical axis, days of observation; horizontal axis, relative number of recordings. Each column corresponds to a functional group. Horizontal bars on the left or each column express the relative number of the individuals marked (in a given group), recorded on the day indicated. Hatched areas indicate these individuals' part in building. Checkered bars on the right of the columns indicate participation in abortive attempts to found a nest N7. concern the same 6 individuals which repeatedly appeared during that period on the surface of the nest. On the other hand, 19 identified individuals from this group were noted in the removal to N6. In the light of these observation, we may consider Column 3 as an expression of the fact that of this group remained after the removal in the external service. The majority of the group of carried workers during the normal colony life after removal resumed their role in internal service. In view POLYCALIC COLONIES OF ANTS 615 of this observation it is significant that the ants of this group immediately after being carried to a new nest, took an active part in that removal as well as in the next removal which took place after most of them had returned inside the nest (see Fig. 7). The workers who carried the brood (Subgroup 4), disappeared entirely from the surface shortly after the completim of the removal. They re- appeared during the next removal and took part in it in great numlbers, mostly carrying the brood. However, for 2 or 3 days after the removal they might still remain on the surface and even take part in constructing the dome of the nest, otherwise a function of the external service. Previously I happened to observe a participation of the nurse workers of Formica exsecta in building work. It was in the summer of 1966 which was rainy and cold. I observed simultaneously two colonies: a monocalic (Cl) and a two-nest (C2) one. At each opportulnity during n sunny weather, pupae were carried out to the surface of the nest for warming and drying which is a frequent phenomenon in the Formica. They were not, however, simply laid down on the surface of the mound but were placed in specially constructed cells or shelters opening outwards or depressions on the sunny side of the surface of the mound. I suppose that they were thus screened from the chilling wind as the air temperature was not very high. It is interesting that these shelters or cells were constructed by nurse workers as well as others. Aftenva~dsonly a small number of these ants remained in the external service. Most of them, together with the pupae, remained inside the nest. However on subsequent days, when a similar opportunity occurred, they once again carried the pupae to the surface of nest and took part in the construction of shelters for them. In the next year the colony C2 became secondarily mmocalic; the whole colony was located only in one nest while the other nest was abandoned. The last one was occupied by ants of the species Formica fusca, so the possibility of displacement of the previous hwts by agres- sion can be excluded, because Formica exsecta is a much more agressive species than Formica fusca.

DISCUSSION

Marikovsky (16) noticed many cases of splitting from the nest in Formica exsecta, but he did not mention the abandoning of inhabited nests. On the other hand B. Pisarski (personal information) has observed this phenomenon; but in his opinion it concerns only young families which change to a more sedentary way of life as they age. I have an important reason to maintain that migration is a constant element d the behavior of this species. The following, mostly ethu- logical facts, support this statement: 1. As unequivocally follows from the diagrams discussed above, all workers belonging to the external service take active part in F. exsecta's migrations. These are primarily all the individuals shown in Fig. 5, which illustrates the behavior of the workers rnariked on the surface of nest N1 prior to the beginning of migrations. That is, during the period of stabilization. It follows from Fig. 5 that they take an active part in each successive removal. The external service undoubtedly includes all ants which were marked in the 1st day of the construction of a new nest and prior to the beginning of removal (see Fig. 7, Subgmp 1, and Fig. 8, Group 1). Finally, here also belongs a considerable part of Group 2 at least, shown in Fig. 8 as they continue the external activity during the period of stabilization. Depending on the need but regardless of the functions they performed previously, all these ants either build the nest or carry their fellow ants or the brood during migration. 2. The most characteristic and distinguishing feature of migration in F. exsecta is the active participation of ants belonging to the internal service in the removals. This service includes the larger part of the carried workers (Group 3) who, after being carried to a new nest, take immediately an active part in this removal. However after its completion (see Fig. 8) they reenter inside the nest. To the permanent internal service did not belong only the part (a quarter) of Group 3 which remained on the surface after the removal. This indicates that these were young workers which stayed far inside the nest as the result of their reluctance to the external environment, caused by their immaturity (11, 12, 14, 15). Once overcome, this reluctance disappears (6). Similarity of behavior was o~bserved (Fig. 7) in the process of removals, of the ants carried (Group 3) and of the ants carrying (Group 2). That proves that once the direction of removal had been inldicated by the workers belonging to the external service, the division of labor into external and internal services disappeared in the course of the same removal; that is, both take an equally active part in i-t. This is an un- precedented phenomenon in other Formica, in which the carried workers frequently return to their old nest and sometimes even take back the brood which was previously carried by other ants to the new nest. The internal service undoubtedly includes the brd-carrying workers. The fact that the same group of workers carried almost the entire brood (see Fig. 7) during each removal and that it completely disappeared from the surface when the removal work was completed (Fig. 8) gives irre- futable evidence that these were nurse workers. This is by no means POLYCALIC COLONIES OF ANTS GI; contradicted by their transitory participation in the building work. The regular building workers never disappeared (as a group) inside the nest, but remain in the external service, taking part in various wolik. Even with a very low activity of the nest a certain number of such ants may be found on the surface at any time. The short-lived participation in building preceding a complete disappearance of the group of ants under study from the surface may in my opinion indicate only that this (building was directly related to the care of bmd. It is also easily imagin- able that a freshly founded nest is not yet sufficiently habitable to give sufficient protection to the brood. The previously descr+bed part of nurse workers in the construction of cells or shelters for pupae in colonies C1 and C2 confirms that, for immediate needs of the brood, nurse workers may temporarily assume the role of builders. Thus, all these indicate that functional Group 4 is composed of nurse workers and the conclusion may be drawn that the nurse workers of F. exsecta take an active part in transportation during the removals. 3. Each attempt of F. exsecta at migration, which was not undertaken by the entire community (see items 1 and 2), failed at the longest within 3 days and did not lead to acts of removal. No conflict of contradictory tendencies mentioned in the introduction and which is characteristic of other species of the Formica occurs here. The cases of carrying the workers to the abandoned nest (Fig. 6, the 2nd, 5th, 6th and 7th days), may in these definite cases be logically interpreted as follows: The 2nd day of observation (Aug. 6) was the 1st day after removal of the colony from N1 to N4 had been completed. The ants carried on this clay to N4 mld not come from N1, since \by that time the last- named nest was already quite empty. One should be aware that the movement from N4 to N3 was bound to take place (at least in the initial part of the way) by the same route which had been used on the previous (1st) day (Aug. 5), but in the opposite direction, for the removal from N1 t~ N4 (see Fig. 4). A considerable number of workers were still excited by the previous day's removal 5. The exitations evoked by various removals differ from each other in an intensified response of the moving individuals to a definite direction specific for a given removal. The instant beginning of the movement opposite to the direction of that removal could possibly evoke in those ants a reaction in the

5 It is clear, both in Fig. 6 and 8, that excitation caused by migration and expressed in an increased appearance and activity of workers in all cases persisted for a few days after the completion of removal. form of intercepting and carrying back to N4 their fellow ants "escaping" from this nest. Still easier to explain was a similar phenomenon recorded on the Zth, 6th and 7th aays (Aug. 9, 10 and 11). It was also during that period that an attempt at founding a new nest N5 (see Fig. 4) was observed, situated yet farther east, which is not shown in Fig. 6. In this case, there was an inevitable conflict of two, very antagonistic tendencies: to move east or west, viz. to settle in a still wetter terrain or to return to the dry clearing. Quite naturally, the surface of N4 had to become a scene of this struggle, expressed by hauling back to the nest of the ants, walking (or being carried) from it in a "wrong" direction. This explains the phenomenon, incomprehensible at a first glance, of renewal of the active building of this nest on the 7th day, that is, the day before it was completely abandoned by the colony. Carrying the belongings to a nest is a symptom of accepting it as home and, therefore, a number of the inhabitants can be stimulated to perform functions connected with a removal to such a nest. The building of the nest is among them the most noticeable function for an observer. Due to these antagonistic tendencies the removal to N3 was the most long-lasting one of the removals of F. exsecta observed by me. 4. All these indicate that in F. exsecta - in contrast to other species of Formica - the process of removal is not a sporadical phenomenon. For, this process is here not at all, or almost at all, inhibited by the attachment to the habitat. For this attachment is undoubtedly a factor which, in other Formica, causes only the most excitable wolikers to mn- duct the removal; while the remaining ones resist it, returning to the previous nest. In F. exsecta the whole society is seized within a few hours by the excitation connected with the drive to removal. Outwardly it is manifested firstly by sudden intense performance by a great mass of ants in all the operations involved in a removal up to the completion of migration. Secondly it is manifested by the fact that only the carrying wor'kers return for a new burden. This is the reason why the brood is carried, as a rule, within 1 or 2 days, and the carrying of workers may last a day or two longer. Thus a normal unhindered removal takes 3 days, which is a rate never observed in other species (except for those cases of cataclysms described in the introduction). If the nest which is being settled is constructed anew, the removal is preceded by a period of building in which the internal service does not yet participate. New nests are built within 1 or 2 days; this was confirmed by PisarSki (personal information). 5. The statement that migration is a permanent element of the be- POLYCALIC COLONIES OF ANTS 619 havior of F. exsecta explains the facts previously discovered that uninhabited nests may be found in each polycalic colony of this species. The cases of migration to clearly temporary and insufficiently founded shelters described above (as N4 and N5) also give evidence for a stpang drive to migrate in the species under study.

Does the fact that the colony under study turned out to be a monocalic one and that it was probably a young family preclude, in the light of Pisarski's opinion, generalization of the results obtained to include polycalic colonies also? I think it does not. First of all, only Pisarski's statement about the tendency of young ant families to migrate does include the assumption that F. exsecta are marked by precisely this tendency and may loose it only with the development of the family. On the other hand, the fact that polycalic colonies of this species continuously split is unanimously acknowledged (16, B. Pisarski, unpublished data, J. Dobrzanska, unpublished observations). As follows from this fact, new nests repeatedly appear in the colony; and, therefore, the tendency to migrate has to be constantly displayed on the colony-wide basis. The results, described in Part I, are evidence of a supra-nest attachment of ants to the external service in polycalic colonies. As follows from this principle, each worker of this service may be found in each nest of the colony, participating in its social life and, in the case of a new nest, in removals. Thus, the fact that the workers of the external service may preserve their tendency toward migrations is not contradictory to Pisarski's statement. This service does not "age" together with the one nest since it is constantly rejuvenated as the result of its supra-nest variable make-up. On the other hand, an internal service of ants which, excepting their removals never leave the nest, "ages" along with the entire nest. It is possible, therefore, that a hereditary attachment to the habitat inhibited in young ants is reinforced in the members of this service with a pro- longed stay in a single nest. This obviously'takes place in turn by an inhibition of the opposite tendency to migrate. The mutual relations of the force of these two tendencies would depend on the age of the nest. If this reasoning and Pisarski's observations are correct, we may expect that the process of possible removals from old nests may resemble that recorded in other Formica. The characteristic struggle between two opposite tendencies should increase with the age of the nest. All these rather confirm the existence of innate tendencies to migrate and explain the considerable frequency of removals in F. exsecta. In summation: the frequent migrations, abandoning of habitable nests, rapid rate of constructing new ones, harmonious participation of all functional groups in removals, rate and efficiency of all removal operations, one-way transportation traffic concentrated in time, are all symptoms of innate behavior features, fixed by evolution. After a pro- longed habitation of a single nest these features may occur less distinctly since an opposing tendency of the attachment to the habitat may be developed by the internal service under these conditions. Thus, I conclude that the tendency to migrate is an innate element of the behavior of F. exsecta. Since the observed cases of migration always take place within the boundaries of the feeding area of the colony, this is a tendency to local migrations.

There still remain three problems which I failed to solve on the basis of the results of my studies: 1. What are the stimuli which cause migration in F. exsecta and is migration always conducted with such frequency? To answer this question we must to know the causes of particular removals. In another paper I shall try to analyse the removals discussed here against the background of meteorological changes recorded in that summer. If it turns out that these changes were of fundamental importance, then we may expect that the removals do not always take place with the same frequency. Under very favorable atmospheric conditions, the mode of life of this species may be less mobile. Nevertheless, the atmospheric conditions certainly are not the only stimuli causing the migrations of F. exsecta, since in the colony under study not all removals may be justified by microclimatic changes. As mentioned above, nests N1, N3 and N6 may ldiffer only slightly from each other in this respect. On the other hand a removal from NXI to NXIII in colony C13 and the abandonment by this colony of at least two more nests (out of a total of the five nests abandoned) may be easily explained by microclimatic reasons. Thus, there are surely still other causes of this species' drive to migrate the knowledge of which requires subsequent studies. 2. The studies undertaken in Part I1 did not attain, in principle, their original aim, that is, a more profound knowledge of the character of the phenomena recorded in C13. For no supra-nest attachment sf external service found in C13 could be observed in colony C3. However, despite the lack of these phenomena in C3 or perhaps rather because of this lack, our knowledge of these phenomena did increase. For in contrast to polycalic C13 it turned out that C3 was a monocalic colony (which enabled the olbservation of a clear picture of migration unolbscured by other processes). This being so, the lack of these same phenomena in C3 does not contradict by any means the hypothesis of the supra-nest at- POLYCALIC COLONIES OF ANTS 621 tachment of the external service in F. exsecta, which could not be manifested in a one-nest colony. On the other hand, finding a tendency to local migration in the species under stuldy makes this hypothesis more probable. For the supra- nest attachment of part of the society clashes with the attachment, to a place, primarily to a habitat, which is characteristic of all species of Formica. A migration consisting in frequent changes of habitat is bound to overcome this attachment, to facilitate severing ties with the previous nest, to recognize each nest inhabited by the members of the colony as their own and, consequently, to undertake in such a nest all indispensable social work. From this stage, it is only one step to a free movement among the inhabited nests of the colony and to performance in them of functions typical of particular individuals. 3. In connection with learning of the building function by F. exsecta, found by Ddbrzanski (8), the performance of this function by nurse ants poses additional pro'blems. It should be determined whether the young nurse ants learn this function or if the elements of this wofi are con- tained within limits of innate activities. If the latter is the case, this would give a new source of knowledge of the innate elements of building behavior in this species.

SUMMARY

1. The mode of life of F. exsecta is marked by a considerable tendency to migrate. It consists in an easy abandonment of the hitherto inhabited nest and a migration to a new one which was rapidly constructed or to an old nest which had been previously abandoned by this colony. The migrations observed took place within the boundaries of the feeding area of the colony and, therefore, I call this feature of behavior "tendency to local migrations". 2. The innate character of this behavior, is proven by: (i) an active and concerted participation of the workers of both the external and internal service in removals; (ii) a prompt beginning of them to participate in a removal directly after their being passively carried to a new habitat; (iii) rapid rate of completion of the entire operation, particularly of carrying the brood; (iv) prompt and final abandonment of the previous nest by all of its inhabitants. 3. The results do not directly confirm the hypothesis on the supra- nest attachment of the exterwl service in F. exsecta colonies, set forth in Part I of the paper. However, in the light of the tendency of this species to local migrations, this hypothesis becomes more credible. This would be a further stage of extending the social unit outside the boundaries 04 one nest as a new social aspect of the ants' life on a colony scale. 622 J. DOBRZANSKA

REFERENCES

1. ADLERZ, G. 1887. Myrmecologiska Notiser. Ent. Tidskr. 8: 41-50. 2. ARNOLDI, K. V. 1932. Biologische Beobachtungen an der neuen Palaarktischen Sklavenhalterameise Rossomyrmex proformicariurn K. Arn., nebst einigen Bemerkungen iiber die Beforderungsweise der Ameisen. Z. Morph. Okol. Tiere 24: 319-326. 3. DLUSSKY, G. M. 1964. The ants of the subgenus Coptoformica genus Formica (Hymenoptera, Formicidae) USSR (in Russian, English summary). Zool. Zh. 43: 1026-1040. 4. DLUSSKY, G. M. 1965. Protected territory of ants (Hymenoptera, Formicidae) (in Russian, English summary). Zh. Biol. 26: 479-489. 5. DLUSSKY, G. M. 1965. Methods of quantitative census taking of soil dwelling ants (in Russian, English summary). Zool. Zh. 44: 716-727. 6. DOBRZANSKA J. 1958. Partition of foraging grounds and modes of conveyings information among ants. Acta Biol. Exp. 18: 55-67. 7. DOBRZAfiSKI J. 1956. Investigations on time sense in ants (in Polish, English summary). Folia Biol. 4: 385-397. 8. DOBRZARSKI, J. 1971. Manipulatory learning in ants. Acta Neurobiol. E,xp. 31: 111-140. 9. ESCHERICH, K. 1917. Die Ameise. Schilderung ihrer Lebensweise. Braun- schweig. 348 p. 10. FOREL, A. 1921-19U. Le monde social des fourmis du globe. Compare a celui de l'homme. Librairie Kundig, Gengve. 5 Vol. 11. GOETSCH, W. 1953. Vergleichende Biologie des Insekten-Staaten. Akad. Ver- lagsgesellschaft Geest and Portig, Leipzig. 482 p. 12. HEYDE, K. 1924. Die Entwicklung der psychischen Fahigkeiten bei Ameisen und ihr Verhalten bei abgeanderten biologischen Bedingungen. Biol. Zbl. 44: 623-654. 13. HUBER, P. 1810. Recherches sur les moeurs des fourmis indiggnes. J. J. Pa- schoud, Paris. 328 p. 14. KIIL, V. 1934. Untersuchungen uber Arbeitsteilung bei Ameisen. Biol. Zbl. 54: 114-146. 15. LUBBOCK, J. 1929. Ants, bees and wasps. A record of observations on the habits of social Hymenoptera. Paul Kegan. London. 430 p. 16. MARIKOVSKY, P. I. 1962. On intraspecific relations of Formica rufa L. (Hymenoptera, Formicidae). Ent. Obozr. 41: 83-93. 17. OKLAND, F. 1931. Studien uber die Arbeitsteilung und die Teilung des Arbeits- gebietes bei der roten Waldemeise (Formica rufa L.). Z. Morph. Ukol. Tiere 20: 63-131. 18. WHEELER, W. M. 1960. Ants, their structure, development and behavior. Columbia Univ. Press, New York. 663 p. 19. WILSON, E. 0. 1971. The insect societies. Harvard Univ. Press, Cambridge. 548 p.

Received 25 October 1972 '+

Janina DOBRZANSKA, Department of Neurophysiology, Nencki Institute of Experimental Biology, Pasteura 3, 00-973 Warszawa, Poland.