The Slave Raids of Harpagoxenus americanus.
Psyche 34(1):11-29, 1927.
This article at Hindawi Publishing: https://doi.org/10.1155/1927/15970
CEC's scan of this article: http://psyche.entclub.org/pdf/34/34-011.pdf, 2024K
This landing page: http://psyche.entclub.org/34/34-011.html
The following unprocessed text is extracted automatically from the PDF file, and is likely to be both incomplete and full of errors. Please consult the PDF file for the complete article.
19271 The Slave Raids of Harpagoxenus americanus 11 THE SLAVE-RAIDS OF HARPAGOXENUS AMERICANUS. BY W. S. CREIGHTON,
Bussey Institution, Harvard University.
The observations here recorded were made at the American Museum of Natural History's Station for the Study of Insects at Tuxedo, N. Y. Through the courtesy of Mr. Frank Johnson of New York City the author was enabled to spend the summer of 1926 at the Station. To Mr. Johnson and to Dr. Frank E. Lutz, whose encouragement and support greatly facilitated this work, the writer wishes to express his sincere thanks. After the manuscript had been prepared it was learned that Dr.Sturtevant had made a number of observations about the same time on Harpagoxenus colonies found in New Jersey. I have had the opportunity of reading Dr. Sturtevant's manuscript and his results and mine amplify each other with very little repetition. In several cases observations have confirmed hypotheses in a remarkable manner, all the more striking in that neither of us knew of the work of the other. Dr. Wheeler has, therefore, suggested the simultaneous publication of the articles as they were originally prepared.
In 1893 Pergande found, near Washington, D; C., a mixed nest composed of Leptothorax curvispinosus (Fig. 1. B), and an unknown ant to which Mayr subsequently gave the name Tomo- gnathus americanus (Fig. 1, A). Later the generic name was changed by Forel to Harpagoxenus, the specific name remaining unaltered. Besides the type specimens of H. americanus Mayr had a few others accidentally taken by Schmitt while collecting beetles at Beatty, Pa. During the next twelve years there are no further records of this ant until, in 1905, Dr. Wheeler dis- covered, near Bronxville, N. Y., three small mixed colonies of H. americanus and L. curvispinosus nesting in hollow elder twigs. The next observation on this rare insect came twenty years later when Sturtevant, during the summer of 1925, found on Naushon Island (Woods Hole) a mixed Harpagoxenus-L. curvispinosus colony inhabiting an oak gall.
Pu&e 34:) 1-29 (1927). hup Ytpsychu einclub orgt34/%Ol l.html
Fig. 1. A, worker of Harfmgoxenus americanus; B, worker of Leptothorax cvrvispinosiis, after Wheeler.
During the past summer I found a number of nests of H. americanus near Tuxedo, IS. Y., and was fortunate enough to see these ants engaged in slave raids. In all but one nest the slave ant was L. longispinosus. The remaining colony contained slaves of both L. longispinosus and L. cilrvispinosus, a condition parallel to that described for the European H. sublcevis, which enslaves both L. aceruorum and muscorum. Dr. Wheeler pos- tulated that such colonies should occur, since he found a queen- less lon,gispinosus-curvispinosus colony in the vicinity of the Harpagoxenus nests. However, since he did not think H. americanus a dulotic ant, he considered the colony as one which needed only the advent of a dealated Harpagoxenus queen to form a mixed colony containing two species of slaves. For reasons to be subsequently given I would consider Dr. Wheeler's mixed Leptothorax colony as a remnant, originally formed by dulosis, from which the Harpagoxenus had migrated or been killed off while raiding.
Before taking up the discussion of H. americanus it seems advisable to briefly outline the ethology of the European species. , H. suUcevis has been known since 18'48, and since that time a number of observations have been made on its habits. It appears to have a wider distribution than its American cogener, for it has
19271 The Slave Raids of Harpagoxenus americanus 13 been taken in Finland (Nylander), Denmark (Meinert), Sweden (Stolpe and Adlerz), Saxony, West Prussia, Bohemia (Viehmeyer) Karnten, Austria (Wolf), and the Swiss Engadine (Emelius). In this last location it occurs at an altitude of 1600-1700 meters. The pioneer observer of sublcevis was Adlerz, who in 1896 con- ducted numerous experiments on these insects in artificial nests, supplemented by observations in the field. He found that in the mixed colonies the slave ant was usually L. acervorum, oc- casionally L. muscorum and rarely L. tuberum. Normal deflated Harpagoxenus queens were never present, their place in the colony being taken by the ergatoid female, a wingless worker- like insect possessing ocelli and a receptaculum semiw's. Alderz further showed that the sublcevis workers, if forced to do so through the absence of slaves, are able to feed themselves but do not care for the brood. This is normally tended by the slaves who feed the larvae with regurgitated material and fragments of insects. This last observation was later confirmed by Vieh- meyer, who amplified the work of Adlerz by discovering, near Dresden in 1906, the true winged queens of sublcevis which Adlerz had supposed did not exist. Viehmeyer's observations 08) show that the actions of a sublcevis female in founding a colony closely parallel those of sanguinea queens. After the entrance of the parasitic female into the Leptothorax nest the workers attack her, but are eventually killed or driven away by the larger insect who then takes possession of the nest and brood. Adlerz had previously observed similar actions of the ergatoid females and had concluded that sublcevis customarily obtains its slaves in this manner. Such dispossession was thought to be fundamentally different from the raids of the amazons and sanguinary ants, who, after the pillage of the strange nest, return to their own with the plundered brood. Both Adlerz and Viehmeyer saw the "raids" of Harpagoxenus, if by this term we may translate the ergringstag of the former writer and the Raubzug of the latter. Unfortunately both these investigators interrupted the proceedings by disturbing the nests, with the result that it is impossible to tell from their observations whether they witnessed a true raid or the dispossession of Lep- tothorax nests by the Harpagoxenus. Indeed, Adlerz considered
14 Psyche [February
that this last was what he had seen, but Viehmeyer seems to have interpreted the two forays which he discovered as true raids. Moreover each found mixed colonies which contained slaves of both L. acervorum and L. nmscorum. Since the pos- sibility that such colonies could have been formed by dispossess- ion is very remote, these investigators concluded that H. sub- lceuis must make dulotic raids. It is gratifying to be able to supplement their observations with others on the dulosis of H. americunus.
The slave raids of Harpagoxenus show none of the organ- ization and spirit so characteristic of those of Polyergus. They fall far short even of the lesser degree of cooperation exhibited by F. sanguinea. Nevertheless they are of great interest, since they appear to be the last manifestation of a vanishing character. They show the decay of dulosis and foreshadow a state of abject parasitism. During the progress of a raid the Harpagoxenus workers and their slaves cluster about the entrance of their nest in a manner suggestive of Polyergus, but with this the similarity ends. There is no rapidly moving phalanx of raiders, no con- centration about the entrance of the raided nest, no frantic activity to enlarge the entrance. The Harpagoxenus leave their nest singly and amble awkwardly and uncertainly to the nest of their victims. Only once did I see more than one Harpagox- enus leave the nest at the same time. On one occasion a column of three departed for the Leptothorax nest. However, this column broke up almost at once and was, I believe, purely fortuitous. On arriving at the Leptothorax next the Harpago- xenus wastes no time in preliminaries but enters at once. Having secured a larva or pupa it emerges as quietly as it entered and returns with its burden to its own nest. Quite often they lose their way and I have seen a number of them captured by small spiders while raiding. The action of Harpagoxenus after it enters the raided nest could not, of course, be followed in the field. However, observation of ants in artificial nests show that the brood is obtained by force and not by stealth, although indeed, the action of the Leptothorax during a raid furnishes ample evidence that this is the case.
The nest which yielded most of the observations on the
The Slave Raids of Harpagoxenus americanus slave-raids was ideally situated for such work. It was in a
crevice in one of the many out-jutting ledges of rock which stud- ded a slope above a small pond. Jammed into the crevice was a small flake of stone which formed a little shelf on the perpen- dicular face of the ledge, and between this flake and the solid rock the nest was constructed. The entrance had been enclosed by a carton of lichens, perforated in the center by an almost circular aperture. Eleven feet away under the bark of a small log which rested on the ledge was the nest which the raiders visited. Between the two nests the ledge was bare except for large black lichens and occasional leaves and fragments of bark. The progress of the raiders could therefore be followed with the greatest ease, an important consideration when dealing with small, slowly moving insects.
The first raid was observed on August 4th. Arriving at the nest at 3:00 P. M. I noticed a crowd of Leptothorax and Har- pagoxenus about the entrance. At 3:02 a Harpagoxenus arrived with a pupa. Another at 3:04. In both cases the longispinosus workers crowded about the returned raider apparently attempt- ing to relieve it of its burden. A third pupa was brought in at 3:08 and carried directly into the next by the Harpagoxenus, which turned and backed into the entrance pulling the pupa in after it.
At 3:25 a Harpagoxenus arrived with a callow. Another with a pupa at 3:26. In each case the raider entered the nest with its burden in the manner just described. During transpor- tation the pup= are held in such a way that they project forward and upward in front of the head of tlic raider. The log in which the raided nest was located was small and badly decayed. The bark had loosened with the decomposition of the underlying wood and, as was later determined, just under the bark in a roughly oval excavation was the raided nest. To this the Harpagoxenus gained access through a rupture in the bark several inches away. The actions of the raiders on ar- riving at the log were somewhat disconcerting. They entered as nonchalantly, if I may be pardoned an anthropomorphic term, as though it had been their own nest. However any doubts as to the nature of their visit were quickly dispelled by the action of the Leptothorax. About ten inches from the point
16 Psyche [February
where the raiders entered a group of longispinosus, with part of the brood, had collected on the surface of the bark. They were huddled above the brood or moving uneasily nearby. In the group were several males. The importance of this observation cannot be minimized. It is proof positive that the actions of the Harpagoxenus were dulotic and not migrational. It removes the possibility that the nest in the log was a remnant of the mixed nest in process of transportation to a new site. It shows that the Leptothorax were aware of the presence of the raiders and, being unable to prevent the pillage of their nest, had resorted to flight to save the brood. Coupled with the actions of t3he raiders it establishes their dulosis in a manner which seems incon- trovertible.
Additional evidence of a less conclusive nature was obtained by an examination of the raided nest. This was made on August 9th, after a heavy thunderstorm had put an end to the raids. The Leptothorax colony was in no respect a remnant. "No Har- pagoxenus were present. There were about forty longispinosus workers, a brood consisting largely of larvae, and a dozen virgin queens.
The old queen I could not find.
It is entirely likely
that she had perished during the raid, but in any case the presence of the virgin queens is a guarantee that the colony had been queenless for only a short time, if at all. Alderz's observa- tion that sexual forms of the slaves are sometimes present in mixed colonies of H. sdcevis invalidates what might be other- wise considered additional conclusive evidence of the individuality of the raided nest. However the character of the Leptothorax colony is at leash not contradictory t-o the conclusion reached above.
I shall cite one further field observation before taking up those made on ants in artificial nets. It may perhaps serve to impress the reader with the degenerate nature of the dulosis of Harpagoxenus. On August 5th I arrived at the mixed colony at 2:00 P. M. The entire ledge was in the shade and no ants were outside the nest, although I could see the head of one in the en- trance. During the next three quarters of an hour a few Lep- tothorax left t,he nest and wandered about on the ledge nearby. During t,his time also the shadow had shifted, so that when at
19271 The Slave Raids of Harpagoxenus americanus 17 2:45 the first Harpagoxenus emerged the ledge was in the bright sunlight. After a short period of aimless wandering the Harpa- goxenus started in the direction of the raided nest. Its progress was slow and uncertain, and on several occasions it became lost, to judge from its haphazard movements. It entered the Lep- tothorax nest at 3:14, having required twenty-nine minutes to cover 3.35 meters. Five minutes later it emerged with a callow and began to retrace its steps. In so doing it passed close to a longispinosus 'worker. The later drew back at its approach and then stood perfectly quiet until the raider had passed, an action which I subsequently observed many times in the artificial nests. Despite the added burden the Harpagoxenus completed the return trip in twenty-two minutes, entering its nest at 3:42. Neglecting the preliminary movements after its emergence from its nest, and deducting the five minutes spent in the raided nest, the entire trip of about 6.7 meters had been completed by the Harpagoxenus in fifty minutes. This gives it an average speed of 13.5 em. per minute. Compare this with Dr. Wheeler's observations ('10) on the speed of Polyergus when raiding, viz: 1.3 meters per minute. It is true that Harpagoxenus is a small ant, scarcely a third the length of Polyergus, yet its smaller size will hardly account for its discrepancy in speed. I recently had the opportunity to study the rate of movement of Iridomyrmex humilis, a smaller ant than Harpagoxenus, and found that under optimum conditions it moved at the rate of approximately 2.75 meters per minute. Size, then, has little to do with the activity of an ant. If, as seems altogether likely, we can assume the correctness of Dr. Wheeler's postulate, ('07) that the ancestors of dulotic ants were active, predatory insects; if from this con- dition dulosis arose as a refinement of active predatism; then we must regard an ant which, though a raider, is slow and clumsy of movement as one in which the dulotic instincts are nearing extinction.
We now take up a series of observations upon ants in arti- ficial nests, which are of interest in that they suggest what takes place when the raiders first enter the nest of their victims. On August 20th) while collecting in a damp wood, Mr. Brandt Steel called my attention to what I at first took to be a mixed colony
18 Psyche [~ebruar~
of Harpagoxenus and L. curvispinosus.The nest was situated in a hollow twig that rested on a moss-covered rock. The open end of the twig had been plugged with a beautiful lichen carton with the customary circular aperture in its center. On the outside of the twig a group of the yellow curvispinosus clustered about a Harpagoxenus worker, while on the rock immediately below was a~ similar group. The entire colony was easily collected, placed in a handkerchief, and subsequently transferred to a plaster observation nest. On splitting the twig two more Harpagoxenus workers were found. Once the nest was installed in the obser- tion nest it became apparent that I had misjudged the situation. The only queen present was that of the Leptothorax. Moreover under the binocular I could see the curvispinosus workers clinging with locked mandibles to the legs and antennae of the Harpago- xenus. Here was no mixed colony but a number of raiders, fiercely resisted by the curvispinosus. That the latter had suf- fered in spite of their superior numbers was evident from their missing legs and antennae. While I was watching, one of the curvispinosus maneuvering for a grip on the antenna of the Harpagoxenus had both its own antennae bitten off when it came within reach of the intruder's jaws. Three of the four Harpa- goxenus workers were thus surrounded. The fourth had found a shelter under a piece of leaf mould. Here it remained coming out only to drive away any curvispinosus that approached its refuge. By next day the curvispinosus had established themselves in a small cavity in the plaster at one end of tlhe nest. The brood had been carried inside and the entrance partially blocked by vege- table detritus. During the night two of the Harpagoxenus wor- kers had succombed to the attacks of the curvispinusus and a third seemed badly crippled. The remaining Harpagoxenus, in all probability the one who had stayed under the leaf mould, was now surrounded by three curvispinosus workers which had fastened their mandibles on its legs and antenna. It made no efforts to extricate itself during the entire day but must have succeeded in doing so during the following night, for on the morning of the 2nd) it was wandering about the nest apparently uninjured. At this time there were nine dead curvispinosus workers in the nest.
The Slave Raids of Harpagoxenus americanus At 1:37 P. M. the Harpagoxenus entered the little cavity in which the Leptothorax had placed their brood. Thirty seconds later the curvispinosus queen came hastily out and ran to the opposite end of the nest. After another thirty seconds a cur- vispinosus worker emerged with a pupa. This worker was fol- lowed by the Harpagoxenus for a short distance, but the latter quickly returned to the cavity. One minute later it appeared in the entrance to the cavity with a Leptothorax worker in its man- dibles. This it dropped outside the cavity and immediately reentered. A general distrubance now pervaded the Leptothorax. An exodus of those in the cavity, many of them carryingpupse, was followed by a spread of the excitement to the workers in the nest, who ran about in a confused and terrified manner. A few bold individuals remained at the entrance of the cavity and two ventured inside long enough for each to recover a pupa. A third worker who attempted to enter was apparently seized by the Harpagoxenus, for it disappeared into the cavity with startling suddenness and did not come out again.
At 2:00 P. M. the Harpagoxenus began a series of maneuvers which can best be described as "reconnoitering.". At first it thrust only its head out of the entrance of the cavity and waved its antennae about. At such times any curvispinosus near the cavity rushed away as though in terror. After about forty minutes of this behavior, it began making short sorties from the cavity. At first it hardly more than emerged, returning at once, but after twenty minutes it began to extend its reconnoiters. Twice during these sorties the Harpagoxenus attacked and injured a curvispinosus worker. In both cases this was done in a delibe- rate and spiritless manner. Most of the curvispinosus workers who chanced to encounter the intruder ran from it and were not attacked. When it left the cavity at 3:40 a change seemed to come over the Harpagoxenus. It no longer lumbered about the nest in its customary manner. Its movements were quick and active. During the next thirty-two minutes the Harpagoxenus attacked and injured eight curvispinosus workers with a ruth- lessness and a "mordlust" that would have done credit to PoZyerqus. The victims were bitten in the thorax and abdomen and left in a dying condition on the floor of the nest. Twice
20 Psyche [February
during the absence of the Harpagoxenus from the cavity a cur- vispinosus entered and secured a part of the brood, and once the raider returned in time to drive away two others who were about to enter. During the foregoing period most of the Leptothorax had collected in two groups on the floor of the nest, the larger of which contained the queen and the majority of the brood carried out of the cavity. The Harpagoxenns now began to pass close to the cum'spinostts. At 4:19 it approached the smaller group and attacking the workers, drove them away from the brood. This it examined with great care, touching the larvse and pup= with its antennae. Two cuwispinosus workers which attacked it in a half-hearted manner were quickly driven away. To my sur- prise it left the brood unmolested and turned to the larger group of Leptothorax which was now in great confusion. Some of them '
the Harpagoxenus drove away, attacking so vigorously two workers burdened with pupae, that the pups were abandoned and their bearers put to flight. It then attacked a curvispinosus callow and soon afterwards bit off the antennae of a worker. At 4:28 it returned to the cavity having taken no further notice of the brood. At this point the observations were interrupted for an hour and a half. When I returned the Harpagoxenus had left the cavity and had been seized by three curvispinosus workers. All of its former spirit and activity was gone, nor did it ever again manifest these characteristics. It lay supine and unresisting on the floor of the nest, its antagonists attached to its legs with locked mandibles.
This was the condition at 10:00 P. M.
ever by next day ifc had freed itself and, to judge from the dim- inished brood of the Leptothwax, had carried a number of larvie and pup into the cavity. At 11 :17 A. M. it emerged, secured a pupa and returned with it to the cavity. The disorganization of the Leptotfwrax was complete. Twenty dead or dying workers were scattered about the nest. The queen lay on one side ap- parently dead.
The remaining workers moved aimlessly about, a few carrying pupa or larve. They no longer resisted or ran from the Harpagoxenits but stood perfectly quiet when it ap- woached, and even allowed it to touch them with its antennae. by the following day the Harpagoxenus had obtained practically all of the curssspznosu8 brood. Only five larvae and pup remain-
19271 The Slave Raids of Harpagoxenus americanus 21 ed outside the cavity. During the next night the curvispinosus recovered a part of their brood but this was again taken by the Harpagoxenus. This see-saw possession of the brood continued until September 3rd., when it was ended by the death of the Harpagoxenus. At this time there were eight curvispinosus wor- kers still alive. They immediately occupied the cavity and resumed the care of the brood.
Any interpretation of the facts just described may be ' criticized on the grounds that the ants were captive insects placed in artificial surroundings.
I am convinced that their
actions may be explained on just such a basis, viz: that they were captives, unable to leave the artificial nest. There can be little doubt that the actions of the Harpagoxenus constitute a "dis- possession" of the Leptothorax of their nest and brood. Never-
t(he1ess this cannot be regarded as the founding of a mixed colony. With this in mind I examined the four Harpagoxenus individuals with the greatest care and can positively state that not one of them was an ergatoid female, if one may judge by the absence of ocelli or a female type of gaster.
Under such circumstances the
most plausible explanation is that of an interrupted raid. A condition making the return of the raiders to their own nest impossible, in this case confinement in the artificial nest, is, I believe, the reason for such occurrences. I have already given reasons for believing that the vestigial dulosis of Harpagoxenus makes its raids, even under the most favorable circumstances, very precarious. The surroundings of the nest from which the raids described above took place were as nearly ideal as could have been wished. The distance between the two nests was compararatively short. The smooth rock surface with its scat- tered lichens offered a minimum of obstructions and a substrate which facilitated the progress of the raiders. Yet even here the Harpagoxenus had difficulty in finding their way to and from the raided nest and frequently lost their way. Under the less favor- able conditions of terrain which usually surrounds the colonies the chances for a successful raid would be greatly lessened. The initial entry into the raided nest is, as the above observations show, a difficult matter requiring much time. During the inter- val, perhaps several days, from the advent of the first intruder
22 Psyche [February
to the successful defeat of the defenders of the nest, there is ample t'ime for altered external conditions. A heavy thunder shower might completely change the character of the ground between the two nests, efface any scent trails present, and render the return of the raider to its own nest extremely difficult or impossible. Under such circumstances there would result a mixed nest containing no queen but Harpagoxenus and Lepto- thorax workers. I found one such nest during the past summer and two of the nests which Dr. Wheeler discovered near Bronx- ville were of this character. It seems likely that the development of the ergatoid females in the European sublcevis may have come about as a response to frequently interrupted raids. An insect of this type, able to participate in a raid and equally able to function as a queen, should a return to its own nest become im- possible, is at least a logical outcome of such conditions. An enumeration of all the mixed colonies found during the past summer will be given here, together wit'h such observations as have not been already recorded.
No. 1, found Aug. 4th, is the colony from which the raids took place. Since it was never distrubed an accurate count of the number of ants present cannot be given. It was, however, a comparatively large colony and I should estimate that it con- tained at least a dozen Harpagoxenus workers and ten or twelve times that number of L. longispinosus slaves. No. 2, discovered on Aug. 6th, was the largest colony of any that were seen. It was under a stone on one of the ledges about thirty-five meters to the west of No. 1. Since the nest was almost
entirely surrounded by an area of perfectly bare rock it was an easy matter to collect the whole colony. Stragglers that escaped
the initial transfer were picked up with a camels-hair brush and placed in the artificial nest. I am certain that not more than half a dozen ants escaped.
The colony contained a normal
Harpagoxenus queen with vestiges of wings, 17 Harpgagoxenus workers, 30 L. curvispinosus workers, 206 L. longispinosus workers 2 longispinosus males and considerable brood, most of which subsequently proved to be Leptothorax. The Harpagoxenus workers were very inactive, rousing themselves only to solicit
19271 The Slave Raids of Harpagoxenus americanus 23 food from their slaves.
The Leptothorax workers spent much
time diligently licking their lethargic masters. No. 3 can hardly be considered a mixed colony, for as such I designate the L. curvispinosus nest attacked by the four Harpa- goxenus raiders. It was found Aug. 20th) and is included here for the sake of clarity.
No. 4 was taken on Aug. 24th. It was located under a small flake of stone on the top of a boulder about eight meters from the spot where No. 3 was discovered. It contained a normal, dealated Harpagoxenus queen, 11 Harpagoxenus workers, 113 L. longispinosus workers and a fair amount of brood. I feel sure that the raiders found in No. 3 came from this nest, but ex- periments to prove this, while not invalidating the opinion, gave inconclusive results. Six Harpagoxenus males were produced in this colony after its transfer to an artificial nest. These are so different from the insect figured by Alderz ('96) as the male of H. sublcevis that, were it not for his statement that he observed Harpagoxenus males in copulation with ergatoid females, it would seem likely that he had described aberrant males of Leptothorax. A figure of a Harpagoxenus male, together with a description is given at the end of this paper.
No. 5 was found on Aug. 30th.) in dry, fairly open woods. The colony was under a flake of stone on the top of a boulder. Not wishing to disturb it I removed the covering stone only long enough to make certain of the presence of the Harpagoxenus queen, and to count the Harpagoxenus workers, of which there were twelve. I did not atetmpt to count the longispinosus slaves but would estimate that there were about a hundred. No. 6, a small queenless colony was also found on Aug. 30th., about fifteen meters from No. 5. It consisted of two Harpago- xenus workers, ten L. longispinosus workers and a small brood of a dozen or more larvae. When the stone covering the nest was removed both the Harpagoxenus and the longispinosus seized the larvae and attempted to move them out of danger. I am of the opinion that this nest represents a L. longispinosus colony "dis- possessed" by raiders from No. 5, who for some reason were unable to return to their own nest.
No. 7 was found on Sept. 3rd., about seventy-five meters
24 Psyche [February
to the west of No. 5.
It was plainly an incipient colony con-
taining a normal dealated Harpagoxenus queen, one Harpago- xenus worker and thirty-five longispinosus slaves and a large brood composed entirely of larvae. The queen was subsequently found to lack the last pair of legs, only the stumps of these re- maining.
This mutilated condition, was also true of a dealated Harpa- goxenus queen which was taken on Aug. 3rd. This insect, which
was crawling over a stone when discovered, lacked the left antenna and one leg on the left side. It seems likely that it had been repulsed in an attempt to enter a nest of Leptothorax, since there were two fairly large longispinosus colonies about three meters from the stone where the Harpaqoxenus queen was found. A most careful search failed to reveal any mixed colony in that vicinity.
As may be seen from the foregoing descriptions H. americanus presents one striking contrast to the European species. Normal winged queens are the general rule with americanus instead of the ergatoid females which usually take their place in sublcevis. The ergatoid form seems to be as rare in the colonies of amer- icanus as are the winged queens in those of sublcevis. I have never seen an ergatoid female in any of the colonies thus far observed.
While much remains to be done it is now possible to sketch the ethology of H. americanus; a hasty sketch to be sure with many missing features hypothetically supplied, but at least a beginning. In this picture we see the fertilized Harpagoxenus queen entering a Leptothorax nest by force. Having driven away or killed the original owners of the brood she appropriates this, tends it and is in turn tended by the resulting Leptothorax workers, until in time there arises a mixed colony. When the colony is well established the dulotic instinct manifests itself in the Harpagoxenus workers. These gain entrance to some Lep- tothorax nest after a long struggle in which their greater hardiness and superior size finally enables them to kill or intimidate the Leptothorax workers.
If the external conditions are favorable the raid is carried to a successful conclusion and the brood of the pillaged nest is carried back to the mixed colony. If the
19271 The aave Raids of Harpagoxenus americanus 25 return of the raiders is rendered impossible then a fragmentary mixed Harpagoxenus-Leptothorax colony results. The phylogeny of Harpagoxenus has been a matter of con- siderable speculation. The development of its raiding habits, a point formerly much disputed, is now agreed as paralleling that of Polyerqus. Viehmeyer ('2 1) eventually concluded that his hypothesis of the lestobiotic derivation of the raids of this ant was untenable and accepted the above explanation. On the other hand, there has never been any doubt of the close morpho- logical affinities of Harpaqoxenus to Leptothorax. L. acervorum (subgenus Mychothorax) is considered as closely related to, or identical with the ancestral form. It is generally believed that the parasitic ants (intraspecific parasitoids) have been devired from those species which they parasitize. (Wheeler '19). Oc- casionally when the parasitic relation has been of long standing the original host may have become rare or extinct and have been replaced by another closely related to it. The relation of Har- pagoxenus to its slaves clearly illustrates both these points. The resemblance between the European H. sublcevis and L. acervorum, its usual host, is close. On the other hand H. amer- icanus parasitizes either L. longispinosus or curvispinosus, neither of which it particularly resembles. This is exactly what we should expect from the distribution of the host species. L. acervorum, a dominant ant in the European palseartic fauna, is rare in North America. The dominant neartic forms in this country are L. longispinosus and L. curvispimsus. It would be expected that the American Harpagoxenus would have adopted these forms in place of the rare acervorum. In the study of phylogenetic relationships the males play an important part because of their morphological stability. Since the males of H. americanus have been hitherto unknown it seems advisable to devote a brief survey to the phylogeny of this ant. Through the kindness of Dr. Wheeler I have been enabled to examine material in his collection, and on this the following comparisons were made. We should expect morphological sim- ilarities to be most manifest in the males of the two genera and least apparent in the workers. Nevertheless even the latter exhibit a number of significant features in common. The work-
tennffi. The importance of this lies in the fact that the sub- nus MycMhorax, to which acervor~m belongs, is the only one of e holsrctic eubgenera of T~ptotkwax which exhibits this charac- ristic. "Furthermore, in both Sfychotkorax and Harpagoxenus here is an impressed mesoepinotal suture. Since the cornbina- ion of these two characters is the criterion by which Myckothoraa s separated from all the other subgenera in the group, their currence in Huqmgoxenus leaves little room for doubt as to the me relation of the two genera.
he similarity between the males of H. subICEvis and L. urn was close enough to confuse even such a careful ob- rver as Adierz. (96). "Unfortunately it has not been possible mine the male of subIcesis but a comparison of Adlerz's with the male of acervorum showed differences which are surprising in the light of that author's statement. In H. cunus there are even greater differences. Nevertheless insects show a sufficient number of similar characters to the view that Harpagoxenus has been derived from My- oihoras,.
The males of both genera have twelve-jointed an- nm and in H, fiub!ċ´oem, the shape of the joints is almost identical th that of L. acervorum. Curiously enough except for the her of joints, the antenna, of H. amricanus resembles that '
, longispinosus rather than acervorm. The mandibles of 11. ifwis are similar to those of L. acervwum in that they are thless and so short that they do not meet. The mandibles of ameriwmt~ are longer, toothed and muwonate at the tip. regard to petiolar structure H. amen'mnus more closely ap oaEhes L. acervorum than does sublosvis. Neither of the former ess the ventral tooth which arms the second node of the tiole in H, sublmis. It is interesting to note that the second tiolar node in L. ammorum is somewhat expanded laterally, ince the squamiform second node is a marked characteristic of goxenus. Allowing for the minor variations which occur the speciea, the anterior wing venation in the males of the hind wings of H. am-
two genera is practically identical. The iconus are veinless, buti those of sublasil
19271 The Slave Raids of Harpagoxenus americanus 27 From the foregoing consideration it is clear that H. sublcevis shows closer affinities to L. acervorurn than does H. americanus. This may be correlated with the habits of the two slave-makers. Sublcevis has apparently always lived with the host from which it was derived and has therefore changed only slightly. On the
other hand, arnericanus, adopting new hosts, has diverged to a greater extent from the ancestral type.
In conclusion is given a description of the male of H. amer- icanus (Fig. 2.)
A, male of Harpagoxenits americanus; B, head of same.
Length: 2.7 mm.
Color : head, thorax and abdomen brownish black; antenna and legs very pale, almost transparent in fresh specimens. Head rugulose with numerous erect hairs. Thorax feebly rugulose, somewhat glabrous, and with fewer hairs. Abdomen glabrous with sparse erect hairs. Antennee 12-jointed. Funi- culus without a distinct club but the joints gradually increasing in diameter towards the tip. First funicular joint pyriform, much broader than those immediately succeeding it. Second and third funicular joints cylindrical and distinctly shorter than the adjacent joints. The following joints sub-oval and grad- ually increasing in length towards the tip. Scape one quarter the length of the funiculus. Scape and funiculus clothed with a short erect pubescence. Antenna1 scrobes much shallower than in female and worker, but distinct. Mandibles long and narrow, feebly toothed and sharply mucronate at the tip. Neck long and flattened dorso-ventrally.
Anterior face of mesonotum abruptly projecting above pronotum. Mayrian furrows strongly impressed at the pro- mesonotal suture, but becoming feeble at their point of con- fluence.
Fore wings with a short open radial cell. Hind wings veinless except for faint impressions at the base of the wing. Epinotum unarmed. Second node of petiole without ventral tooth, squamiform, broader and less constricted behind than in worker. Petiolar hairs sparse.
1896. Myrmccologiska Studier. 111. Tomognathus sub- lcevzs. Bihang Till K. Svenska Vet. Akad. Handlingar. Band 21, Aft. IV No. 4, pp. 3-76.
Sturtevant, A. H.
1925. Notes on the Ant Fauna of Oak Galls in the Woods Hole. Psyche vol. 32, pp. 313-314.
1906. Beitrage zur Ameisenfauna des Konigreiches Sachsen. Abhandl. naturwiss. Gesell. Isis, Dresden, pp. 55-69.
19271 The Slave Raids of Harpagoxenus americanus 29 1908. Zur Koloniegrundung der parasitischen Ameisen. Biol. Centralbl., vol. 8, pp. 18-32.
1921. Die mitteleuropiiischen Beobachtungen von Harpa- goxenus. Biol. Centralbl., vol. 41, pp. 269-278. Wheeler, W. M.
1907. The Origin of Slavery among Ants. Pop. Sci. Month., vol. 71, pp. 550-559.
Ants, Their Structure, Development and Behavior. Columbia University Press, 1910.
1919. The Parasitic Aculeata, a Study in Evolution. Proc. American Phil. Soc., vol. 58, pp. 1-40.
Volume 34 table of contents