Carlos Roberto F. Brandão.
Division of Labor within the Worker Caste of Formica peripilosa Wheeler (Hymenoptera: Formicidae).
Psyche 85:229-238, 1978.

Full text (searchable PDF, 572K)
Durable link: http://psyche.entclub.org/85/85-229.html

The following unprocessed text is extracted from the PDF file, and is likely to be both incomplete and full of errors. Please consult the PDF file for the complete article.

DIVISION OF LABOR WITHIN THE WORKER CASTE OF FORMICA PERPILOSA WHEELER
(HYMENOPTERA: FORMICIDAE)*
Museu de Zoologia, Universidade de S5o Paulo Siio Paulo, Brasil
Polymorphism, in the study of social insects, is defined as the existence within an individual colony of two or more phases or castes belonging to the same sex, without particular regard to their genetic or environmental origin (Wilson, 1953). The adaptive result of the development of female polymorphism is the division of labor within the colony. In most ants this division is clearly seen between reproductive and non-reproductive caste, but less evident within the worker caste (Oster & Wilson, 1978). The present article utilizes the Fagen & Goldman (1977) method for estimating the total repertory size of behavioral categories of each worker subcaste listed on an ethogram or behavioral catalog.
Ethograms are the essential first step of the comparative study of behavior (Wilson, 1974). A behavioral catalog of Formicaperpilosa Wheeler, a weakly polymorphic species of the neogagates group (Buren, 19681, was constructed in order to investigate behavioral differences between the major size groups, defined here arbitrarily as three worker subcastes. Formicaperpilosa is a common ant in the southern United States and northern Mexico (Gregg, 1963). It feeds mainly on plant exudates and tends membracids of the genus Pubilia (LaBerge, 1952). Its physiology has been relatively well studied by Schumacher & Whitford (1974), Kay & Whitford (1975), Whitford, Kay & Schumacher (1975) and Schumacher & Whitford (1976). The genus Formica is of unusual interest because its species are either monomorphic or weakly polymorphic and thus span the early stage of caste evolution. Yet close studies of the polymorphic species have not been undertaken.
*Manuscript rerewed b~l the edztor October 10, 1978 229
Psdw $5:229.137 (1978). httpllp%ycb enlcIubw@5B5-219 html



================================================================================

230 Psyche [June-September
MATERIAL AND METHODS
Founding queens were collected near Portal, Arizona, in May, 1975, by B. Holldobler and individually placed in test tubes, 14.8 cm long by 23 mm inner diameter, kept moist by compact cotton plugs that trap water at the bottom of the tubes. As the selected colony grew the tube was moved to a plexiglass box 28 cm X 45 cm and 15 cm deep, the sides of which were coated with Fluon GP-1 (Northeast Chemicals Co., Woonsockett, R.I.) to prevent escape, and four similar tubes added. The colony has since been maintained on an artificial diet for ants (Bhatkar & Whitcomb, 1970) and honey water three times a week and fed freshly killed cockroaches (Nauphoeta cinerea) once a week.
Before the first set of observations the colony was moved to a glass nest made of two square plates 12.5 cm on a side, held apart by small pieces of non-toxic plasticin and taped on three sides. This simple nest permits close observation of the behavior of the entire colony. The assemblage was put on the floor of the original container next to the water tubes. The remainder of the floor served as foraging space and was kept clear for observation. The container was small enough to be placed under a swinging arm dissecting microscope. During a period of 4 weeks, a total of 18 hours were dedicated to cataloging behavior; 2809 separate be- havioral acts were recorded. The observation hours ranged ran- domly from 9100 A.M. to 1l:OO P.M.; no differences in level or pattern of activity were noted, related to time of day. This species is polymorphic in the sense of Wilson (1935); for the purpose of this investigation the workers were classified in 3 groups, minors, medias and majors. Samples of 15 specimens of each group were selected later for head width measurements in order to check the adequacy of the classification. Specimens that could not be readily placed in one of the size classes were not included in the ethogram.
The behavioral catalog of Formica perpilosa is presented on Table 1. Twenty-eight behavioral categories were observed in the minor category, 34 in the rnedia, and 1 I in the major. By fitting the frequency data to a lognorrnal Poisson distribution (Fagen & Goldrnan, 19771, the total numbers of categories, including those



================================================================================

19781 BrandGo - Worker Caste of Formica perpilosa 23 1 Table 1.
Relative frequencies of behavioral acts by the three worker categories of a single colony of Formica perpilosa. (N, total number of behavioral acts recorded for each caste). Approximate population of the nest: one nest queen, 150 minor workers, 270 media workers, 30 major workers, 50 eggs, 7 larvae and 5 pupae. Behavioral Acts
A-Grooming
1-Autogrooming head, 1st pair of legs
2-Autogrooming 2nd pair of legs
3-Autogrooming abdomen, 3rd pair
of legs
4-Allogrooming minor workers
5-Allogrooming media workers
6-Allogrooming major workers
7-Allogrooming nest queen
B-Brood Care
8-Standing at the brood pile
9-Carry egg (or eggs in succession)
10-Lick egg (or eggs in succession)
1 1 -Carry larva (or larvae in succession) 12-Feed larva
13-Lick larva (or larva in succession)
14-Carry pupa
15-Lick pupa
16-Assist eclosion to adult
C-Regurgitation Behavior
17-Regurgitation with minors
18-Regurgitation with medias
19-Regurgitation with majors
20-Donating to the Queen
21 -Lay trophic egg
22-Feed queen trophic egg
23-Feed larva trophic egg
24-Carry infrabucal pellet
25-Feed on infrabucal pellet
D- Working
26-Foraging outside the nest
27-Feed on diet
28-Feed on honey
29-Feed on cockroaches
30-Carry live nestmate
3 1 -Carry dead nestmate
32-Drag the nest Queen
33-Handle nest material
34-Carry debris
35-Lick nest wall
36-Excavating
E-Other Behaviors
37-Antenna1 tipping
38-Jittering
Total
no. behavioral categories observed
per caste
minor
workers
N = 996
.I67
.060
.023
.054
,043
.007
.o 1 1
.066
.033
-0 16
.009
0
.Ol 1
0
0
-002
-070
.040
.007
0
0
0
0
.003
.007
.082
.022
.029
.020
0
0
0
,018
-033
.059
.092
.oo 1
.O 16
1 .o
28
media
workers
N = 1679
.13 1
-062
.022
.O 19
.095
.o 12
.049
.077
.055
.053
.024
.004
.03 1
0
0
-024
.024
.I65
-02 1
,009
0
,001
.oo 1
.009
.o 1 1
.02 1
.004
,025
.007
.oo 1
.003
.oo I
.005
.003
.034
-0 12
.002
0
1 .o
3 4
major
workers
N = 134
.224
-052
0
0
,052
.030
.022
.134
0
0
0
0
0
0
0
0
.052
.29 1
.045
,008
0
0
0
.090
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1 .o
11




================================================================================

23 2 Psyche [June-September
not seen? were estimated to be respectively 29? 36 and I I, with 95 per cent confidence intervals of (25? 33) (25? 47) and (5? 19). Some behavioral categories deserve special mention: I.
Carry pupa or lick pupa: Care of pupae consisted exclusively of assisting eclosion to adult. The pupae remained in the brood chamber and only when the colony was disturbed (which did not happen during the drawing of the catalog) media workers carried them to the new nest.
2. Lav rrophic egg: The actual laying of trophic eggs was not seen, but 3 times media workers were seen offering the larvae and the queen small, round, shiny objects, looking like eggs, but different from the normal ones laid by the nest queen. It is not impossible that the laying of trophic eggs is a real but rare event. 3.
Recruiiment behavior: In order to determine whether the workers utilize any kind of food recruitment, the colony was deprived of food for one week and the nest connected to an arena by a bridge of round sticks. Honey water was presented on the arena floor. Two periods of one hour observation were recorded. Initially all ants that reached the food source were collected and not allowed to return by the bridge. Between the two experiments the honey source, the bridge and the arena floor were changed to avoid recognition by the ants. In the first period 35 ants reached the food source, in the second 262. This experiment clearly shows that this species uses food recruitrnent. Returning ants were observed to rub the tips of the abdomen on the bridge sticks (probably laying a scent trail) and to display no nestrnates. During the construction of the catalog, however, no recruitrnent behavior was noted, probably because the colony was kept fed to saturation. 4. Jirrering and antenna1 ripping: These behavioral categories were described by Wilson (1976), but nothing is known about their meaning. 5. Defensille behavior: Our perpilosa colony was not stressed to ellicit defensive behavior. However, a stray individual of the ant Novomessor iockerelli was found on the nest floor being attacked by media and minor workers of Formica perpilosa; not even in this situation the majors were observed outside the nest. This of course does not mean that majors cannot play a role in defending the nest against predators or raids by other ants, but it is apparent defense is not their characteristic behavior. A more detailed account of the defensive behavior of this species should be useful. After the behavioral catalog was drawn? 50 specimens had their head widths and lengths measured and plotted. The linear ana- rnorphosis afforded a linear regression of head length on head width, where the coefficient of deterrnination is r2 = -964. When logarithrnically plotted? in order to compare with Wilson's rnodel



================================================================================

19781 Brand50 - Worker Caste of Formica perpilosa 233 (1953), the log-log anarnorphosis afforded a linear regression of head length on head width, The equation is log L = -2795 - .8602 log W
or, alternatively,
L = 1.9033 w * ~ ~ ~ ~
The coefficient of determination, r2 is also -964, showing of course an excellent fit.
I have chosen the following criteria for the definition of the categories of workers: rninors, head width 1 rnrn or less; medias, 1 to 1.4 rnrn; majors, larger than 1.4 rnrn.
The rnajors represent only 6.6% of the nest population and their behavioral catalog comprises merely 11 behavioral categories. Their behavior, as seen in Table 2, is xnostly directed toward grooming and regurgitation with nestrnates.
DISCUSSION AND CONCLUSION
The behavioral catalog and the head shape curve of our Formica perpilosa colony show that the worker caste is strongly polyethic, in spite of being only slightly polymorphic. The rnedia category is responsible for most of the behavioral categories. The majors are specialized in regurgitation with nestrnates and may act as a trophic subcaste as in Camponotus (Colobopsis) fraxinicola (Wilson, 1974). F. perpilosa rnajors also possess a relatively large abdornen and Table 2. Relative frequencies of groups of behavioral acts listed on Table 1 by the three worker subcastes in a single colony of' Formica perpilosa. (N. total number of' behavioral acts recorded for each subcaste.) Groups of'
behavioral acts
(from Table 1)
minor media major
workers workers workers
N=996
N=I679 N=l34
A-Grooming .365 .390 -380
B-Brood Care ,137 ,248 .I34
C-Regurgitation
.I27 ,241 .486
D-"Working" .355 ,116 0
E-Jittering & Antenna1 tipping .O 17 -002 0



================================================================================

234 Psyche [June-September
were never observed perforrning any defensive or "working" be- havior. They were recorded as using alrnost half of their tirne doing trophallaxis or offering infrabucal pellets to rninors and medias. King and Walters (1950) have found a very similar situation in "Formica rufa rnelanotica" (=F. obscuripes) in natural nests. They were able to prove a correlation between polyrnorphisrn and polyethisrn, showing that minor workers were specialized in attend- ing aphids. The major workers again remained inside the nest rnost of the tirne.
On Table 2 one can notice that all the categories of F. perpilosa spend the same arnount of tirne perforrning grooming behavior. The rninors are alrnost exclusively responsible for several tasks grouped as "working" and the rnedias spend a more uniform arnount of tirne perforrning each group of behavioral acts. Wilson (1953) proposed a model for the origin and evolution of polyrnorphisrn in ants. In 197 1 he reviewed the ant caste systern and said that five steps can be recognized in the evolution within the worker caste: rnonornorphisrn, rnonophasic allornetry, diphasic allornetry, triphasic allornetry and complete dimorphism. Age polyethisrn is the responsible phenomenon for caste structure in inonornorphic species, and the polyethic classes are sometimes referred to as physiological classes. Traniello (1978) reported an apparent lack of temporal division of labor in the primitive ponerine ant Arnblyopone pallipes, which appears to have the rnost primitive caste systern yet documented in ants.
Monophasic allornetry is the commonest manifestation of non- isometric growth. Studies on the division of labor in weakly polyrnorphic species showed a "by preponderance" division of labor, i.e., any given worker should be capable of perforrning any given normal task.
In Formica polyctena size variation is weakly correlated with division of labor. The behavioral variation observed by Otto (1958) consisted mainly of age polyethisrn and individual peculiarities. Although the F. perpilosa subcastes were not divided using age parameters, I believe that age does not account for a large fraction of the total behavioral variation.
Majors of advanced polyrnorphic ant species, especially corn- pletely dimorphic species, where intermediates no longer exist and the two remaining classes are remarkably different in morphology,



================================================================================

19781 Brand50 - Worker Caste of Formica perpilosa 235 usually act as soldiers. This is not true in the present case, where another kind of specialization was revealed. In the complete dimorphic species there is no behavioral overlap between the castes. Each category can be easily distinguished from the others rnorphologically and behaviorally. Majors of Zacryptocerus varians (members of one of the nine genera in which complete dimorphism is easily recognized) function primarily in defense (Wilson, 1976) and deserve the title of soldiers. Yet they sometimes wash and manipulate larvae and pupae. But during a high intensity attack, majors block the nest entrance with their saucer-shaped heads and are more persistent and effective than minors in forcing the enemy back to the entrance and finally out of the nest.
Dr. Edward 0. Wilson, reviewing the manuscript, was kind enough to suggest that F. perpilosa majors represent an early stage in the evolution toward repletes, as seen in Proformica and Myr- mecocystus (Wilson, 1971). The specialization shown byperpilosa majors (see Table 2) agrees with this view. F. perpilosa seems to occupy, from the viewpoint of evolution of division of labor, an intermediary position between rnonornorphic and completely dimorphic species.
I am indebted to the "Fundaq5o de Arnparo a Pesquisa do Estado de S5o Paulo" for a scholarship (Biologicas 771 1208). I wish to express my appreciation to my parents for supporting my visit to Harvard University; and to Prof. and Mrs. B. Patterson, Dr. and Mrs. W. L. Brown, Jr., and Dr. E. E. Williams for their kind hospitality.
I am specially grateful to Dr. E. 0. Wilson for the opportunity to learn ant rearing techniques, for the suggestion of the problem, assistance and uses of facilities during the research and critically reviewing the manuscript. I thank Dr. and Mrs. B. Holldobler, Gary Albert, Hiltrud Engels and specially Katherine Horton and James Traniello for advice and help with the ethograrn method, and also the latter for showing me his unpublished data on Amplyopone behavior; Mr. 0. Schrnidt for conducting the repertory estimations; Dr. Francisca C. do Val for critically reading the manuscript.



================================================================================

236 Psyche [June-September
Finally, I am grateful to Dr. P. E. Vanzolini for the helpful suggestions during the elaboration of this paper. BHATKAR, A., AND W. H. WHITCOMB.
1970. Artificial diet for rearing various species of ants. Florida Entomol. 53: 229-232.
BUREN, W.F.
1968. Some fundamental taxonornic problems in Formica (Hyrnenoptera: Forrnicidae). J. Georgia Entornol. Soc., 3(2): 25-40. FAGEN, R. M., AND R. N. GOLDMAN.
1977. Behavioral Catalogue Analysis Method. Anirn. Behav., 25(2): 261-274. GREGG, R. E.
1963.
The ants of Colorado. University of Colorado Press, Boulder, Colorado. xvi + 792 pp.
KAY, C. A., AND W. G. WHITFORD.
1975. Influences of temperature and humidity on oxygen consumption of five Chihuahuan desert ants. Cornp. Biochern. Psysiol. A., 52: 281-286. KING, R. L., AND F. WALTERS.
1950. Population of a colony of Formica rufa melanotica Emery. Proc. Iowa Acad. Sci., 57: 469-473.
LABERGE, W. E.
1952. Locality records of two ants found in Kansas. J. Kansas Entornol. Soc. 25: 59.
OSTER, G. E., AND E. 0. WILSON.
1978. Cast and Ecology in the Social Insects. Princeton University Press, Princeton, New Jersey (in press).
OTTO, D.
1958. ~ber die Arbeitsteilung irn Staate von Formica rufapratensis minor Gosswald und ihre. Verhaltensphysiologischen Grundlagen: Ein Beitrag zur Biologie der Roten Waldarneise. Wiss. Abh. Deutsche Akad. Landw. - Wiss Berlin 30: 1-169.
SCHUMACHER, A. M., AND W. G. WHITFORD.
1974. The foraging ecology of two species of Chihuahuan desert ants: Formica perpilosa and Trachymymex neomexicanus (Hyrnenoptera: Forrnici- dae). Ins. Soc. 21: 317-330.
1976. Spatial and temporal variation in Chihuahuan desert ant faunas. South- west Natur. 21: 1-8.
TRANIELLO, J. F. A.
1978.
Caste in a primitive ant: Absence of age polyethisrn in Amblyopone. Unpublished manuscript.