Z v
i % ^ '% f
w *> \ mm .z \
% "mms
iUfy,
1 \ #
warn
!► — \ _
NV '
I, ^
\M0/‘ \ w
4.- ~~
\ V>
6> ” : ^ m^Mk, 'o.
^ %
3fm
-: M-rf
"V V
f- ->„ Iffi
va whO^S
PSYCHE
A Journal of Entomology
Volume 72 1965
Editorial Board
Frank M. Carpenter, Editor P. J. Darlington, Jr.
W. L. Brown, Jr. H. W. Levi
E. O. Wilson H. E. Evans
Published Quarterly by the Cambridge Entomological Club Editorial Office: Biological Laboratories 1 6 Divinity Ave.
Cambridge, Mass., U. S. A.
The numbers of psyche issued during the past year were mailed on the following dates:
Vol. 71, no. 4, Dec., 1964: March 6, 1965
Vol. 72, no. 1, March, 1965: June 25, 1965
Vol. 72, no. 2, June, 1965: September 25, 1965
Vol. 72, no. 3, Sept., 1965: January 18, 1966
PSYCHE
A JOURNAL OF ENTOMOLOGY
Vol. 72 March, 1965 No, 1
W. M. Wheeler Memorial Issue
CONTENTS
CONTENTS
The William Morton Wheeler Memorial Issue 1
Trail Sharing in Ants. Edward O. Wilson 2
Simultaneous Care of More Than One Nest by A mmophila azteca
Cameron ( Hymenoptera, Sphecidae). Howard E. Evans 8
The Ant Larvae of the Subfamily Leptanillinae (Hymenoptera,
Formicidae). George C. Wheeler and Jeanette Wheeler 24
The Australian Ants of the Genus Pristomyrmex, with a Case of
Apparent Character Displacement. Robert W. Taylor 35
A New Species of Megalomyrmex from the Chilean Andes
(Formicidae, Hymenoptera). George Etter shank 55
The Habits and Distribution of Cryptocerus rohweri Wheeler (Hymen- optera: Formicidae). William S. Creighton and William L. Nutting 59 Contributions to a Reclassification of the Formicidae. IV. Tribe
Typhlomyrmecini (Hymenoptera). William L. Brown, Jr 65
A Revision of the Ant Tribe Cardiocondylini (Hymenoptera,
Formicidae). Jonathan Reiskind 79
“Queenlessness,” Worker Sibship, and Colony Versus Population Struc- ture in the Formicid Genus Rhytidoponera. Caryl P. Haskins and
Roy M. Whelden 87
Observations on the Nesting Site and Biology of the Arizona Damp- wood Termite, Zootermopsis laticeps (Banks) (Hodotermitidae).
W . L. Nutting 113
CAMBRIDGE ENTOMOLOGICAL CLUB Officers for 1964-65
President...................... . R. W. Taylor, Harvard University
Vice-President J. Reiskind, Harvard University
Secretary .........H. Reichardt, Harvard University
Treasurer ...F. M. Carpenter, Harvard University
Executive Committee.. N. S. Bailey , Bradford, Mass.
E. G. Macleod, Harvard University
EDITORIAL BOARD OF PSYCHE
F. M. Carpenter (Editor), Professor of Entomology , and Alexander Agassiz Professor of Zoology, Harvard University
P. J. Darlington, Jr., Alexander Agassiz Professor of Zoology, Harvard University
W. L. Brown, Jr., Associate Professor of Entomology, Cornell University; Associate in Entomology, Museum of Comparative Zoology
E. 0. Wilson, Professor of Zoology, Harvard University
H. W. Levi, Associate Curator of Arachnology, Museum of Com- parative Zoology
H. E. Evans, Curator of Insects, Museum of Comparative Zoology
PSYCHE is published quarterly by the Cambridge Entomological Club, the issues appearing in March. June, September and December. Subscription price, per year, payable in advance: $4.50 to Club members, $5.00 to all other subscribers. Single copies, $1.25,
Checks and remittances should be addressed to Treasurer, Cambridge Ento- mological Club, 16 Divinity Avenue, Cambridge, Mass.
Orders for back volumes, missing numbers, notices of change of address, etc., should be sent to the Editorial Office of Psyche, Biological Laboratories, Har- vard University, Cambridge, Mass.
IMPORTANT NOTICE TO CONTRIBUTORS
Manuscripts intended for publication should be addressed to Professor F. M. Carpenter, Biological Laboratories, Harvard University, Cambridge, Mass.
Authors contributing articles over 4 printed pages in length may be required to bear a part of the extra expense, for additional pages. This expense will be that of typesetting only, which is about $10.00 per page. The actual cost of preparing cuts for all illustrations must be borne by contributors: the cost for full page plates from line drawings is ordinarily $12.00 each, and the full page half-tones, $18.00 each; smaller sizes in proportion.
AUTHORS SEPARATES
Reprints of articles may be secured by authors, if they are ordered at the time proofs are received for corrections. A statement of their cost will be furnished by the Editor on application.
The December, 1964 Psyche (Vol. 71, no. 4) was mailed March 6,
1965.
The Lexington Press. Inc., Lexington, Massachusetts
Psyche, 1965
Vol. 72, Plate 1
William Morton Wheeler
march 19, 1865 — APRIL 19, 1937 PROFESSOR OF ENTOMOLOGY, HARVARD UNIVERSITY
1908-1937
Vol. 72
No. 1
PSYCHE
March, 1965
WILLIAM MORTON WHEELER MEMORIAL ISSUE
March 19, 1965, is the centennial of the birth of William Morton Wheeler, Professor of Entomology at Harvard University from 1908-1937 and an active member of the Cambridge Entomological Club for that period. The Editorial Board of Psyche has designated the present number of the journal as the William Morton Wheeler Memorial Issue and has arranged to include in it articles on ants and other social or subsocial insects. No attempt has been made to obtain contributions by all of Professor Wheeler’s former students; most of the papers published here were already in the editorial office before plans for the memorial issue were definitely made. Included are contributions by three generations of students whose interest in social insects can clearly be traced to Professor Wheeler. We are indebted to Professor W. L. Brown, Jr., of Cornell University for the use of the accompanying photograph, which was taken about five years after Professor Wheeler was appointed at Harvard University.
A biographical account of Professor Wheeler, with a complete list of his publications, was published in Psyche, Volume 44, No. 3., 1937.
F. M. Carpenter, Editor
I
JWTHSONIA* jlil |
WSTlTUt^
TRAIL SHARING IN ANTS
By Edward O. Wilson Biological Laboratories, Harvard University
introduction: the kinds of trail sharing
A very few cases have been recorded of ant workers regularly utilizing the trails of other ant species. Forel (1898) designated as “parabiosis” the following complex behavior that includes trail sharing. Colonies of the Neotropical rain forest species Crematog aster limata parabiotica Forel and Monacis debilis (Emery) [ —Doli - choderus debilis var. parabiotica Forel] commonly nest in close association, with the nest chambers kept separate but interconnected by passable openings; while the workers forage along common odor trails. Wheeler (1921) confirmed the phenomenon and showed that, in the one instance where he observed food gathering, the two species were attending membracids together. Wheeler also discovered a similar association between Crematogaster parabiotica and Campo- notus femoratus (Fabricius). Both species were observed utilizing common trails and gathering honeydew from jassids and membracids on the same plants, as well as nectar from the same extrafloral nectaries of Inga. Not only were the Crematogaster and Camponotus workers tolerant of each other in this potentially competitive situation, they were on quite intimate terms. They “greeted” each other with calm antennation on the trails, and on three occasions Wheeler observed Camponotus actually regurgitating to Crematogaster.
It has not been established whether parabiosis is mutualistic or parasitic in nature. The distinction must be a subtle one in such a complicated relationship. The form <( parabiotica” of Crematogaster limata is evidently always associated with other ants. If future taxonomic studies prove it to be a species distinct from limata , it is a likely parasite. It would then be shown to be dependent on its associ- ates, while the latter species often nest and forage by themselves. But the prima facie case for mutualism seems even stronger. The broods are never mixed, and as Weber (1943) points out on the basis of his own studies, all of the parabiotic species participate vigorously to- gether in nest defense. There is no evidence that the presence of the Crematogaster harms the other species, except possibly by competition for the same food resources. On the contrary, Camponotus femoratus
Manuscript received by the editor January 5, 1965.
2
1965]
Wilson — Trail Sharing
3
maintains flourishing populations in localities where virtually every colony lives in parabiosis with Crematogaster.
While the Neotropical parabionts are doubtfully mutualistic, the relationship of the European Camponotus lateralis (Olivier) and Crematogaster scutellaris (Olivier) can be classified as weakly parasitic. Goetsch (1953) and Kaudewitz (1955) have described instances in which Camponotus workers followed the Crematogaster trails in large numbers to the Crematogaster feeding grounds and exploited the same food resources during the same time of day. The Crematogaster were hostile to the Camponotus, which assumed a crouching, conciliatory “Wartestellung” on meeting the host workers. Unlike the Neotropical parabionts, the two species nest separately. Moreover, the relationship is not obligatory on the Camponotus lateralis , since the colonies of that species are often found far removed from Crematogaster colonies.
I will now describe a third example of trail sharing which I recently discovered between the dolichoderine Azteca chartifex Forel and formicine Camponotus heehei Wheeler. This case is of additional interest in that it seems to illustrate a close approach to the third or neutral class of symbiosis, namely commensalism.
AZTECA CHARTIFEX AND CAMPONOTUS BEEBEI
During a trip to Trinidad, West Indies, in 1961, my attention was drawn to Camponotus heehei, a formicine ant previously known from only several specimens collected in Trinidad and British Guiana. On each of three occasions on which the species was encountered, twice at Spring Hill, Arima Valley, and once near Cumuto Village on the Aripo Savanna, workers were found running over tree trunks along the odor trails of the much more abundant and aggressive dolichode- rine Azteca chartifex. The Camponotus were never found away from the Azteca trails. Extended observations at Spring Hill revealed that the Camponotus always followed the Azteca trails for long distances with fidelity equal to that maintained by the Azteca them- selves. That this was true trail symbiosis was further evidenced by the fact that no other alien species remotely approximated such behavior. Workers of several other arboreal species occasionally blundered into the same Azteca files but ran abruptly away without tracing the main route of the files.
One of the Spring Hill Camponotus nests was located. It was in a dead, hard branch of a mango tree that had fallen and lodged in the crown of a three-meter-tall grapefruit tree in a citrus plantation. The Camponotus workers were seen to emerge from their nest holes, run
[March
Psyche
Minor worker of Camponotus beebei Wheeler from Spring Hill, Trinidad.
1965]
Wilson — Trail Sharing
5
down the mango branch to the branches of the grapefruit tree, which held an Azteca colony, and follow the Azteca trails to the ground. The Azteca workers seldom ventured up to the Camponotus nest. The Camponotus occupied scattered flat galleries in the mango branch. When cut apart the nest yielded 2 winged queens, 16 males, 6 major workers, 36 minor workers, and several larvae and pupae in various stages of development. The mango tree, from which the Camponotus colony fragment had evidently recently fallen, was also occupied by Azteca chartifex. In a second locality at Spring Hill, Camponotus workers were tracked up into the foliage of a tonka bean tree ( Dipteryx sp.) beyond a large Azteca nest, but the Camponotus nest was not found. Nevertheless, it was evidently separate from the Azteca nest.
Both the Azteca and Camponotus followed the Azteca trails to the bases of the nest trees. Presumably both foraged extensively on the herbaceous ground vegetation, but their diets were not determined. Regardless of the nature of the diets, competition between the two species was reduced by the existence of opposite diel schedules. The Camponotus foraged apparently exclusively during the day, at the time the Azteca files were at their lowest ebb. In the early evening the number of Azteca workers on the trails were seen to increase by as much as a hundred-fold, but not a single Camponotus worker was found through several (hours of searching during this time.
The Camponotus workers, then, “borrow” the Azteca trails when the owners put them to minimal use. The Azteca workers on the Spring Hill trails were hostile to the Camponotus workers and attacked them on the rare occasions when the latter slowed in their running, but the Camponotus were larger and faster and usually easily avoided their hosts without causing any visible disturbance. The Camponotus were never observed to interfere with the Azteca in any other way.
On the basis of the first observations it could still be legitimately asked whether the Camponotus were merely using the same visual or tactile “landmarks” on the tree trunks as the Azteca, rather than following their odor trails. This possibility was eliminated by the following experimental result. A freshly killed insect was pinned to the trunk of a tree one meter beneath the trail along which both species were running but within the range of occasional Azteca scouts. Within ten minutes, two Azteca workers had found the insect and laid odor trails from it back to the main trail. In the next five minutes over 100 Azteca workers moved back and forth along the new trail to the insect. In the same interval three Camponotus
6
Psyche
[March
workers, a major and two minors, approached along the main trail and, on reaching the junctures of the new trails, departed down them for various distances. The major went all the way to the insect and prowled around it for several minutes before returning to the main trail. In two subsequent replications of the experiment, two of thirteen and one of five Camponotus workers passing along the main trail were deflected onto the Azteca side trails during the period of peak Azteca response to the baits. Such deviations from the main trail were never observed except at this time. It was concluded that the Camponotus respond to the Azteca communication.
The following observation led to the further conclusion that the Camponotus were tracking the Azteca olfactorially rather than visually. Occasionally around midday the Azteca were unusually scarce on the main trail, while the Camponotus remained moderately common. Stretches of 30 to 50 cm. of the trail were often bare of Azteca, but many individual Camponotus followed the established track just as well. On close examination I found no alterations in the surface structure of the main trail, other than the postulated chemical one, that could have supplied the Camponotus with a clue.
Although the Camponotus beebei utilize Azteca trails extensively, the following observation shows that they have maintained their own, private trail system. On a single occasion in February a line of seven Camponotus were seen moving along the main Azteca trail. Four of the workers ran in a tight group directly behind the leader, frequently advancing enough to touch the abdomen of the ant ahead. When the leader was touched, it dashed forward at a faster pace over a short distance. This part of the behavior was typical of communication by “tandem running”, which I have described earlier in a paper on the genera Cardicondyla and Camponotus (Wilson, 1959). The re- maining two workers followed at a greater distance, tracing each twist and turn taken by the leader. During the next 15 minutes several other Camponotus workers passed the same way, again tracing parts of the route of the leader with close fidelity. After that time, new Camponotus workers continued to run on the Azteca trail but ignored the Camponotus trail. There could be no doubt that the lead ant had secreted an odor trail of the recruitment type (see Wilson, 1963). It was laid on top of the Azteca trunk trail, which for most of its length was about 10 centimeters wide. Equally interesting was the fact that only the Camponotus responded to it. The Azteca workers continued to pass along their own trail during the episode but failed to orient to the inner track followed so closely
1965]
Wilson — Trail Sharing
7
by the Camponotns. Thus the Camponotus workers appear to respond to two odor trails, while the host Azteca respond only to one.
Acknowledgements
This study was supported by a grant from the National Science Foundation. The figure was prepared by Mrs. H. C. Lyman.
ABSTRACT
Trail sharing is a rare event in ants. Of two previously described cases, one is interpreted as part of a relationship that is either mutu- alistic or weakly parasitic, probably the former, and the other as part of a weakly parasitic relationship.
A third, new case has been discovered which appears to be com- mensalistic. On Trinidad, West Indies, workers of the rather scarce formicine Camponotus beebei utilize the arboreal odor trails of the abundant dolichoderine, Azteca chartifex. The Camponotus “borrow” the latter’s trails during the day, when Azteca foraging is at a low ebb. The Camponotus workers are treated hostilely by the Azteca workers but are too swift and agile to be caught; their presence does not disturb the Azteca seriously. On a single occasion Camponotus workers were observed to lay their own private recruitment odor trail on top of the Azteca trails. The Camponotus trail lasted for about fifteen minutes and had no visible effect on the Azteca .
Literature Cited
Forel, A.
1898. La Parabiose chez les Fourmis. Bull. Soc. Vaud. Sci. Nat., 34: 380-384.
Goetsch, W.
1953. Vergleichende Biologie der Insekt-Staaten. Geest und Portig K.-G., Leipzig. 482 pp.
Kaudewitz, F.
1955. Zum Gastverhaltnis zwischen Crematogaster scutellaris 01. mit Crematogaster lateralis bicolor 01. Biol. Zentralbl., 74: 69-87.
Weber, N. A.
1943. Parabiosis in Neotropical “ant gardens.” Ecology, 24: 400-404.
Wheeler, W. M.
1921. A new case of parabiosis and the “ant gardens” of British Guiana. Ecology, 2:89-103.
Wilson, E. O.
1959. Communication by tandem running in the ant genus Cardio- condyla. Psyche, 66: 29-34.
1963. The social biology of ants. Ann. Rev. Entomol, 8:345-368.
SIMULTANEOUS CARE OF MORE THAN ONE NEST BY AMMOPHILA AZTEC A CAMERON (HYMENOPTERA, SPHECIDAE)1
By Howard E. Evans Museum of Comparative Zoology
In attempting to trace the origin of social behavior among wasps, in his Social Life Among the Insects (1923), William, Morton Wheeler selected Arnmophila as “a paradigm of the whole group of Sphecoids and solitary Vespoids”. Were he alive today, and able to utilize all the considerable knowledge of this genus gained in the last three decades, it seems likely that he would embrace Ammophila even more enthusiastically as a paradigm not only of the solitary wasps but of several preliminary stages in the origin of sociality.
It has been shown by Evans (1958, 1959) and by Powell (1964) that the North American species of this genus which have been studied can be arranged in series as follows: (1) strictly solitary
species which utilize a single large caterpillar per nest, (2) species which mass-provision with two to several small caterpillars, (3) species in which provisioning is commonly “delayed” such that the last prey is brought in after the egg has hatched, and (4) species employing progressive provisioning regularly. Several other aspects of behavior are roughly correlated with this progression : for example, species using smaller caterpillars usually carry the prey in flight, and these same species generally carry the soil of excavation away from the nest in flight. Also, most records of gregarious nesting pertain to species employing progressive provisioning.
It is apparent that the European species can be arranged in a very similar series (Adriaanse, 1947; Teschner, 1959). One European species, A. puhescens Curtis, illustrates still a fifth stage in this ethocline: the female maintains two or three nests at one time,
remembering the location of each of them accurately and behaving in accordance with the status of the egg or larva in each nest as deter- mined during frequent inspections (Baerends, 1941). Simultaneous care of more than one nest is otherwise virtually unknown among digger wasps, although a few species of Bembicini which make more than one cell per nest are reputed to begin provisioning a second cell
Supported by a grant from the National Science Foundation, no. G17497. Most of the studies reported here were conducted at the Jackson Hole Biological Research Station, Moran, Wyoming.
Manuscript received by the editor January 5, 1965
8
1965]
Evans — Armnophila azteca
9
before the previous cell is fully stocked (Janvier, 1928; Tsuneki, 1956).
Powell (1964) suggested that studies with marked wasps might well reveal that some of our North American A?n?nophila maintain more than one nest simultaneously. By coincidence, during the same week that Powell’s paper appeared, I was able to establish that this is, in fact, the case. Working along the Snake River in Jackson Hole, Wyoming, I found that Ammophila azteca Cameron behaves in a manner strikingly like pubescens in almost every detail relating to provisioning. My studies were relatively brief and leave a number of questions unanswered, but they seem worth putting on record at this time with the thought that others may be able to extend our knowledge of this widely distributed species before I am able to do so.
Ammophila azteca has been the subject of three published notes, all of them brief. Hicks (1935) observed several wasps digging and closing nests near Boulder, Colorado (identified as aculeatus Fernald, a synonym ) . Evans ( 1 963 ) reported on a single nest found in Yellowstone Park, Wyoming, and Powell (1964) presented prey records from California and Baja California. The last two authors both identified the species as pilosa Fernald, a name now regarded as a junior synonym of azteca Cameron.2 I also found one female of this species nesting at Great Sand Dunes National Monument, Colorado (elevation 7800 feet), in August 1964, and have included this record below. This is chiefly a montane species ; Powell’s records are from 6000 feet elevation in Baja California and over 10000 feet in California; the Jackson Hole and Yellowstone localities where I have worked are both at about 6800 feet. However, Menke (in lift.) reports the species from near sea level in California and various places in Canada.
General aspects of ecology and behavior. — Most of my studies were conducted in a small area of flat, bare soil along the Snake River at the Cattlemen’s Bridge, about one kilometer east of the Jackson Hole Biological Research Station, Moran, Wyoming. The first observations were made on July 18, 1964, the last observations on August 14. My impression is that this species became active only a few days before my initial observations and that it had nearly completed its nesting season by August 14. In this area, where the active season for most wasps is no more than 4-6 weeks, progressive
2This is a new synonymy, and should be credited to Arnold Menke, who is currently revising the genus Ammophila and who has studied Cameron’s type. I am indebted to Dr. Menke for identifying the specimens collected in the course of these studies and for critical reading of the manuscript.
IO
Psyche
[March
provisioning would seem uneconomical unless the wasp is able to maintain two or more nests simultaneously. However, two progressive provisioned which maintain only one nest at a time nested in some abundance in the same area; these were Bembix spinolae and Steniolia obliqua. The only other species of Ammophila found here was a single female A. macra Cresson which nested on the edge of the bare soil on August 14; I have presented a few notes on this wasp below, as it presented a striking contrast to azteca.
In Jackson Hole, A. azteca appeared to be strictly confined to patches of bare, moderately firm, sandy loam along the river; none were seen or taken in general collecting in other habitats. The major nesting area was about 20 meters long and varied from 5 to 8 meters in width, paralleling the river and separated from it only by a narrow, oblique bank which was not used for nesting. Parts of this area were covered with grass and herbs, chiefly around several trees, but these places were not utilized by the Ammophila. Also, the Ammophila did not nest in a small plot of very loose sand which was occupied by a colony of Bembix spinolae. The area was surrounded on the three sides away from the river by open woodland in which the dominant tree was lodgepole pine (Pinus contorta latifolia) ; near the river there were also narrow-leaved cottonwoods (Populus angustifolia) and willows ( Salix spp.). I estimated that there were about 50 females scattered about the nesting area. My notes cover 20 females, but only a few of these were marked, and only four of the marked individuals were followed over a period of days. The most complete observations pertain to no. 2030, which was marked with a red spot on the second day of study ( July 19) and followed until August first, including one complete day of observation (July 31) and observations for several hours on five other days.
That it was impossible to make continuous records of more than a very few females was a consequence of the fact that females spaced themselves widely, each maintaining a small nesting arena, no more than 30 cm. in diameter, where all her nests were prepared (Figs. 6, 7). These arenas were separated, for the most part, by half a meter or more. On one occasion an unmarked female shared a circumscribed area with a marked female (no. 2028C), the nests of the two being more or less intermingled, but this was exceptional. I observed no aggression among females provisioning their nests, but females digging in proximity would sometimes attack one another. It seems possible that there is a measure of territoriality among females of this species, but since the wasps spend only a small portion of their time in the nesting arenas, contacts between neighboring females are few.
1965] Evans — Ammophila azteca II
Both males and females were seen frequently on the flowers of Solidago and Erigeron growing in and near the nesting area, and it is probable that they obtain nourishment throughout the season from this source. Early in the season males were observed in considerable numbers flying close to the ground in the nesting area. On one occasion three males were seen digging intermittently and trying to enter a vertical hole, presumably in anticipation of the emergence of a female. I did not observe mating, chiefly because most of my studies were made a week or more after nesting had begun, when most mating had probably ceased and the males were declining in numbers. Mating in several species of Ammophila has been well described (e.g., Baerends, 1941; Olberg, 1959).
As compared to other wasps in this area, Ammophila azteca made its appearance rather early each morning (0830-0900) and was active until late in the afternoon (1630). Active wasps were often noted in partial shade or during periods of partial cloudiness, but they invariably disappeared when the sun was heavily obscured for more than a few minutes. I did not discover where they went during these inactive periods or at night. Hicks (1935) found a female of this species “sleeping” on a stem in late afternoon, presumably solitarily. The wasp held her body at nearly a right angle to the stem, grasping the support with her mandibles and some of her legs. A number of species of Ammophila are known to sleep in somewhat this manner, often more or less gregariously (e.g., Evans and Linsley, i960).
Digging the nest. — Nests appear to be started at any time of day, but more especially in the mid-morning hours or late in the afternoon. Digging females periodically produce a “chirping” sound as they break through the soil with their mandibles. One female starting a new nest dug 1 cm in 13 minutes, the second cm in another 7 minutes; she completed the burrow and cell in a total of 58 minutes. This female had previously been digging at another point 10 cm away, but she abandoned this nest after apparently striking a stone; the aban- doned burrow was left open.
When the wasp has loosened a small lump of earth, she holds it between her mouthparts and front legs, backs from the burrow, rises into the air a short distance, and with a slight jerk flings it onto the ground. She may fly in any direction from the entrance, but most wasps make the majority of flights on the same side, such that the earth comes to form a diffuse but discernible mound in the form of a small arc. The mound is rarely more than 0.5 cm deep, and varies from 2 to 4 cm wide by at least twice that long, although its borders are ill-defined. In some nests no mound can be seen, as the female
Psyche
[March
Figs. 1-3. Typical nests of Ammophila azteca (all from Jackson Hole, Wyo.) showing the three types of closures. Fig. 1. No. 2018, an empty nest closed with a single pebble and a small amount of sand. Fig. 2. No. 1993, a nest with a single caterpillar bearing the egg, showing a typical temporary closure. Fig. 3. No. 1996, a fully provisioned nest which had received the final closure.
scatters the soil rather widely. In any case it weathers away in a few days, and some of the soil may be used in closures.
There appears to be considerable variation in the distance the female flies with the soil, and consequently the distance from the nest entrance the mound accumulates. In Jackson Hole, most mounds were only a few cm from the entrance (2-6, rarely up to 20). Hicks (1935) observed very similar digging behavior in Colorado, but he found that the soil was carried a considerable distance from the nest, in one case four feet (1.2 m). Hicks timed several females and found that they carried from five to seven loads per minute. “One load”, he reports, “was so large and heavy that the wasp was not able to fly with it after a typical and normal fashion and only reached the refuse spot by short flights and hops.” Great variation in the distance the soil is carried was noted by Hicks (1932) in another species, A. aherti.
Burrow diameter in A. azteca is about 5 mm; the terminal cell is horizontal and has a diameter of about 10 mm, a length of 20-25 mm. All burrows at Jackson Hole were vertical or very nearly so; they varied in length from 3 to 6 cm, nearly all of them measuring between 4 and 5 cm (Figs. 1-4). The single nest excavated at the Great Sand Dunes, Colorado, was very similar, having a vertical burrow 3.2 cm long reaching a horizontal cell 2.2 cm long. The nest from Yellow- stone was very similar in size but was dug into the sloping of a foot- path and had an oblique burrow (Fig. 5 in Evans, 1963).
Closure of the nest. — As in many species of this genus, three types of closure' can be distinguished : initial closure of the empty nest, temporary closures during provisioning, and final closure of the fully provisioned nest. The initial closure consists of a single pebble or hard lump of earth which is placed in the burrow entrance with the
1965]
Evans — A mmophila azteca
13
mandibles, often after “trying for size” several such lumps and rejecting each of them until one is found which fits the mouth of the burrow (Fig. 1). A small amount of sand is often scraped over this closure, but no pounding occurs. The wasp normally returns to this nest within a few hours or early the next day with the first prey, on which the egg is laid. The single nest studied at the Great Sand Dunes, Colorado, was like the many observed at Jackson Hole: the female removed a single small stone from the entrance before intro- ducing the first caterpillar and laying an egg upon it.
Following oviposition, and after supplying each additional cater- pillar, a very different type of closure is made (Fig. 2). Again, various lumps and pebbles are “tried for size”, but this time the initial lump is placed well down in the burrow, at least half way down and often near the bottom; the object selected must lodge perfectly in the burrow so that material may be packed above it (this is the hauptverschlussklumpchen of Baerends and other European workers) . When the initial lump is in place, the wasp scrapes sand into the nest with her front legs, facing of course away from the hole. She may utilize some of the soil of the mound if this is close by ; or if loose soil is not available she may loosen some by biting the ground around the entrance. Periodically small pebbles or bits or earth are picked up and placed in the burrow with the mandibles. From time to time the wasp remains with her head down in the burrow and moves her body up and down, packing the soil in place with blows of the antero-ventral surface of the head in the usual manner of wasps of this genus. When the burrow is full, additional packing occurs, the wasp sometimes retaining her grasp on a small pebble while so doing (“using a tool”) and then leaving the pebble in place in the fill (Fig. 8). Finally, sand is scraped in various directions over the top, resulting in thorough concealment of the entrance. The closure is prepared very rapidly and requires only a minute or two; one wasp completed a typical temporary closure in only thirty seconds. It is probable that the same pebbles and lumps of earth serve in successive closures, so one would expect a reduction in the length of time re- quired after the first temporary closure.
Final closure may be indistinguishable from temporary closure unless one follows the provisioning of the nest to determine that this is, in fact, the final closure. My limited data suggest that in the final closure ( 1 ) the initial lump is always placed at or near the bottom of the burrow (Fig. 3), and (2) packing with the head is relatively prolonged and some packing while holding a pebble in the mandibles always occurs toward the end of the closure (Fig. 8).
14
Psyche
[March
Further studies will be needed to quantify these factors and to deter- mine if there is a real difference between temporary and final closure. The closure observed by Hicks (1935) in Colorado was presumably temporary, since no “tool using” was noted. The wasp studied by Hicks used “an old hackberry seed to close the upper end of the shaft. Sand was scraped in over this, and some score or more of objects were brought to further cover and conceal the nest site”.
Provisioning the nest. — Females were observed on several oc- casions flying closely about low branches of willows and cottonwoods, apparently searching for prey. On one occasion a wasp seized a caterpillar on a willow branch, but the latter thrashed violently and the wasp left without stinging it; this was a considerably larger caterpillar than was ever found in the nests of the wasp. The 44 prey taken from various nests were all of roughly the same size (slender, 12-18 mm in length) ; all were “naked” larvae, and all were green in color except for a few gray or reddish geometers. Specific identifi- cation of the prey was not possible, but many (perhaps all) belonged to groups which feed upon broadleafed trees rather than upon conifers. The following were recovered from the various nests dug out: Hymenoptera [det. B.D. Burks]
Tenthredinidae : Neimatinae: Nematus (Pteronidia)
sp. and Amauronematus sp. 28
Lepidoptera [det. D.M. Weisman]
Geometridae (four spp.) 7
Gelechiidae (apparently all one sp.) 8
Sphingidae: Smerinthus sp. (early instar) 1
The use of sawfly larvae in considerable numbers is of interest, since Adriaanse (1947) found that A. pubescens restricts itself to caterpillars, the closely related A. campestris to sawflies. In the population of A. azteca studied, several nests contained nothing but sawfly larvae, a few nothing but moth larvae, and a very few (e.g., no. 2012) both sawfly and moth larvae. Individual wasps tended to stock successive nests wholly with sawfly larvae (rarely wholly with caterpillars), but there were numerous exceptions. I noted no shift in type of prey concordant with the progress of the season.
Prey records from other parts of the range of this species indicate use of lepidopterous larvae of several different groups, including relatively hairy forms, but there are no further records of use of sawfly larvae. The single female studied at the Great Sand Dunes brought in a gelechiid larva very similar to those used in Jackson Hole. The nest studied at Yellowstone contained one noctuid, one geometer, and five larvae of lycaenid butterflies (Evans, 1963).
1965]
Evans — A?nmophila azteca
15
Powell (1964) found a lycaenid larva to be used in Baja California, while in Mono County, California, a larva of the moth family Pterophoridae was employed. As Powell points out, lycaenid larvae are covered with short, secondary setae, while pterophorid larvae have tufts of elongate setae. The caterpillars taken by Powell measured 14 and 15.5 mm in length, while those taken from the nest in Yellowstone measured from 8 to 15 mm in length. Thus it appears that size of the prey is of critical importance, but not its vestiture or taxonomic affinities.
Powell reported that the lycaenid larva was carried in flight, about a meter above the ground. I observed prey carriage in flight once in Colorado and many times at Jackson Hole; in no case did I observe prey carriage over the ground. The female straddles the caterpillar, grasping it with her mandibles a short distance behind the head and during flight embracing the body of the caterpillar with her legs, the prey being venter-up. She lands a short distance from the nest and walks to the entrance, usually without circling or hesitation, then drops the caterpillar at the threshold while she removes the closure. At Jackson Hole, I experienced much difficulty approaching females closely for photography at this time; they would very readily take flight with their prey, often disappearing for several minutes.
Removal of the closure occurs very rapidly, taking only a few seconds if it is an initial closure, generally less than 30 seconds in any case. Soil particles and objects removed from the burrow are deposited within a few cm of the entrance and may be reused when the closure is restored. When the burrow is clear, the wasp grasps the anterior end of the prey with her mandibles and backs into the nest with it. The entire process (arriving with prey — clearing entrance — draw- ing in prey — reclosing entrance) occurs with remarkable speed, some individuals requiring less than a minute. When it is considered that only from one to four prey are introduced per day, it will be appreci- ated that even prolonged periods of observation afford one only fleeting opportunities to study this behavior.
In addition to entry with prey, each nest is entered periodically when the wasp is without prey. There are presumably “inspections” which function to impress upon the wasp the size and food require- ments of the larva, as Baerends (1941) found to be true in A. pubescens. Inspections may occur at any time of day, but most characteristically occur early in the morning (0830-1030). In fact, the first act of each Ammophila arriving in the morning consists in entering the nest and closing it again. I did not observe any one female inspect more than one nest at this time; apparently when there
Psyche
[March
16
4
NEST DEPTH
T3 A I A2 A3 A4 A5 SEGMENT OF PREY
Fig. 4. Number of nests of Ammophila azteca (all Jackson Hole, Wyo.) exhibiting various depths (in centimeters). In each case the measurement was rounded off to the nearest half cm in the field. Fig. 5. Number of examples of oviposition on various segments of the prey (all Jackson Hole, Wyo.). T3 — third thoracic segment; Al . . . = various abdominal seg- ments.
is more than one active nest, she inspects the one containing a larger larva (see detailed data on no. 2030, below). However, other in- spections may occur later in the day (see also below). The entry and reclosure resemble closely the behavior occurring when the wasp is bringing prey.
The maximum number of prey found in any nest at Jackson Hole was six (2030D), but the nest studied in Yellowstone contained seven (Evans, 1963). It appears that about two days are required for hatching of the egg, about five days for larval feeding. Apparently females bring in the last prey and make the final closure when the larva is in the last instar and at two-thirds its full size; thus final closure usually occurs six or seven days after the nest is dug.
Oviposition. — The egg is laid on the first prey placed in the cell. It is glued firmly by the anterior end to the side of the caterpillar, in a more or less vertical position (Fig. 9, lower figure). As in many species of A?nmophila, there is much variation in the placement of the egg. At Jackson Hole, one egg was found on the third thoracic segment, others on each of the first five abdominal segments, but more on the second abdominal segment than elsewhere (Fig. 5). The single nest studied in Colorado contained a caterpillar bearing an egg on the fourth abdominal segment.
Successive nests of individual females. — As mentioned earlier, it became apparent in the course of studies at Jackson Hole that each female prepares all of her nests in a small nesting arena. It also became apparent that each female prepares an unusual number of nests. For example, when I dug out the nest of no. 2013, an un-
1965]
Evans — A mmophila azteca
7
marked female studied early in the season, I found not one nest but six, all in an area measuring 8X14 cm. This was on July 27, and I could not believe that females had been nesting for more than two weeks (I began collecting in this area July 4, and took the first A. azteca , a male, on July 7) . But if it takes six or seven days to rear a larva, this female should have started nesting at least 30 days earlier (one nest was still unprovisioned). This example and several similar ones made it apparent that some overlapping of successive nests must occur. I therefore marked several individuals and at- tempted to follow them over several days. Since my observations were not absolutely continuous, and since during the one complete day of observation I could keep only two wasps under full surveillance, because of the wide spacing mentioned earlier, my notes are not entirely convincing except in two cases. The arrangement of suc- cessive nests of these two individuals is shown in Figs. 6 and 7> and I here present, in abbreviated form, my field notes concerning one of them, no. 2030, which I painted with a red spot on July 19.
July 19. Brought prey at 1100 hrs to site A (Fig. 6).
July 20. Working on a nest at B.
July 23. Prey to B at 1530.
July 25. Prey to nest C at 1630.
July 29. Inspection followed by final closure at C.
July 31. Complete day of observation. Arrived in area at 0815. 0845 : Wasp appeared at site D and made inspection, followed
by temporary closure.
1022: Prey to D, followed by what appeared to be a final
enclosure, including “use of tool” and scraping of sand over top for several minutes. Left at 1036.
1037: Reappeared at same site, walked about, flew off.
1047: Reappeared at same site, walked over old nest and around
area until 1050.
1053 : Back again, walking over nesting arena with her abdomen
held high; off at 1055.
1 1 13: Reappeared and began to remove closure at E. This
appeared to be a rather thorough temporary closure. After a brief “inspection” she made a hasty temporary closure and left at 1120.
1237: Back to E with a sawfly larva; opened nest and drew it
in within one minute of arrival. At closure three pebbles were “tried for size” before finding one which fit the bore of the burrow properly. Left at 1240.
1 355 : Back to E with another sawfly larva. After introducing
i8
Psyche
[March
the prey made a fairly prolonged closure which I took to be a final closure but which proved not to be (see Aug. I ). During this she drove away two chrysidids (see below, under “parasites”).
1400, 1409, 1414, 1445: Each time she reappeared in the nesting arena, walked about for a few seconds to a minute, then flew off.
Started digging at G.
Stopped digging, leaving hole open, and is now walking about the nesting arena.
Digging at G1.
Closed burrow with a single pebble, scraped a small amount of sand over top, and flew off.
Landed at nest, then flew to a Solidago in blossom two meters away, remained for a few minutes, then was not seen again today.
August 1. 1020: Landed at G1, remained two minutes and flew off.
This is a cool, partly cloudy morning.
1040: Now cloudy and windy (began to rain at 1100). I dug
out the nesting arena, eventually finding 12 nests and one incompleted burrow (G) (see Fig. 6). Eight of these had cocoons, one was an empty nest (G1), one had a single caterpillar with an egg (F), and two had wasp larvae (D and E, closed yesterday, E apparently tempo- rarily, since the wasp larva was small and there were only 3 sawfly larvae in the cell).
Apparently nest F had been initiated on July 30 and was not visited on July 31 (the wasp presumably remembering that it con- tained an egg and did not require additional prey). Thus we can say that this wasp very definitely maintained two nests simultaneously (D and E ) , very probably three ( if we include F ) , and that she was associated with three nests in the course of one day (D, E, G1) while presumably retaining a fourth in her memory. Clearly this calls for a much more detailed study, but it would appear that the models provided by Baerends (1941) for A. pubescens apply very well to this species.
Some unexplained features of behavior. — On two different oc- casions I dug out a series of nests in a single nesting arena and found that all or most of the nests contained eggs. In one case there were six nests, five of them containing a single prey with an egg, the remaining one being empty. In another case there were four nests, all of them containing only one prey; in three cases the prey bore an egg, in the
1450:
1457:
1502 : 1600:
1632:
1965]
Evans — A mmophila azteca
19
Fig. 6. Arrangement of successive nests of no. 2030. Nests A-G1 were prepared in that order (see text) ; unlettered nests were not seen to be dug or provisioned and were presumably completed before the beginning of intensive observations (at least in part). Fig. 7. Arrangement of nests of no. 2031, which nested about 2 meters from the preceding at Jackson Hole.
fourth case a small larva. Unfortunately in both instances the wasp associated with the nesting arena was unmarked, and I did not observe the site for a long enough time to be certain that only one female was responsible for each set of nests. As mentioned earlier, I did find one case in which an unmarked female shared a nesting arena with a marked female. If the above two cases did involve a single female each, then that female must have prepared several nests on the same day, or over no more than two or three days, and laid several eggs in rather rapid succession. When the larvae in these nests approached maturity, the female would have to obtain many prey in order to provision all of these nests. Whether truly synchronous nesting (as opposed to the overlapping of successive nests) occurs in this species remains to be determined.
On one occasion (2007D) I dug out a nesting arena and found three nests, one of which contained one caterpillar, the other two two sawflies each : but none of the nests contained an egg. This suggests that these might have been storage burrows, the prey to be exhumed later and used for oviposition or for feeding a larva. If in fact this species does at times maintain several nests synchronously, then the maintenance of storage nests would seem of great adaptive value, since it would spread out the requisite hunting over a greater number of days. It is possible that the maintenance of storage burrows is not a rare phenomenon in A?nmophila. Hartman (1905) found two
20
Psyche
[March
Fig. 8. Female Ammophila azteca at final closure of her nest. A small pebble is being held in the mandibles and used for packing soil in the filled burrow.
nests of A. procera in Texas which did not contain eggs, Tilden (1953) had the same experience with this species in California, and Criddle (1924) observed a female of this same species in Manitoba place a caterpillar in a nest one day and then dig it out the next day, return it to the nest, and lay an egg upon it This is an aspect of behavior requiring much further study.
Parasites. — Chrysidid wasps (Ceratochrysis perpulchra Cresson) were observed commonly in the nesting area of A. azteca at Jackson Hole, but I am unable to state definitely that they were parasitizing the Ammophila. The chrysidids would often land on the soil near Ammophila nests, and on several occasions they were seen to enter nests either partially dug or partially closed, in which case they were normally driven away by the Ammophila. On one occasion a chrysidid returned after the Ammophila had completed her closure and spent several minutes around the nest, but I did not see her dig into it.
1965]
Evans — Ammophila azteca
21
Notes on Ammophila macra. — As noted earlier, a single female of the large species A. macra Cresson was found nesting on the periphery of the azteca nesting area at Jackson Hole. She was seen digging her nest at noo on August 14. The soil was taken from the burrow in lumps much as in azteca , but it was carried over the ground and deposited in a diffuse pile 6-14 cm from the hole. At 1115 the wasp selected a large stone and placed it in the burrow (which was fully 1 cm in diameter). She then dropped in several small pebbles and proceeded to scrape in sand, from time to time turning around and packing the sand in place with her head while making a loud buzzing sound. When the burrow was nearly full, she added several more stones, pressing each into place. Finally, she dug soil from a small, semicircular hole 3 cm from the entrance, making a shallow quarry or “false burrow”, and scraping this soil into the nest entrance. This was packed into place by holding a stone in her mandibles and pounding, finally leaving the stone in place. The wasp then picked up a stick 4 cm long and dragged it over the top of the covered nest entrance, then added a second stick 7 cm long. She completed the closure at 1135 and was not seen again. The nest was dug out two days later and found to contain a single large sphingid larva ( Smerin - thus geminatus Say) bearing an egg on the side of the fourth ab- dominal segment (Fig. 9, upper figure). The burrow was vertical, 6 cm long, the cell horizontal and 5 cm long.
Thus this species differs from azteca in several important features: the soil is carried from the nest on the ground ( and the prey doubtless carried on the ground) ; some of the soil for closure is obtained from a quarry and various objects are placed on top of the closed nest; and the nest is stocked with a single large caterpillar, thus eliminating the possibility of progressive provisioning or of the overlappng of pro- visioning of successive nests. A. macra bears a close resemblance to the well-studied species procera Dahlbom, and the behavior is similar to that species, although procera more commonly flies with the soil when digging.
Summary. — Ammophila azteca differs from other known North American species of its genus in that several nests are maintained simultaneously; in this respect it bears a close resemblance to the European species puhescens. Data indicate that at least two and at times probably three or more nests in various stages are maintained at one time, the female remembering the precise location of each and inspecting each nest, or at least most nests, each day before provision- ing. The total duration of each nest is six or seven days.
The prey consists of caterpillars and sawfly larva of many diverse
22
Psyche
[March
Fig. 9. Above, larva of sphingid bearing the egg of Ammophila macra. Below, larva of a geometrid bearing the egg of A. azteca. These were extracted from nests in close proximity at Jackson Hole.
groups, all of them diurnal leaf-feeders of about the same size (8 to 1 8 mm, usually 12 to 16 mm). The egg is laid on the first prey in each nest and shows much variation with respect to the segment on which it is laid. Up to seven prey are provided per nest.
Soil is carried from the burrow in flight, and the prey is carried to the nest in flight.
Initial closure of the nest consists of a single object placed at the top of the burrow. Temporary closures after oviposition consist of a large object blocking the lumen of the burrow on top of which loose soil and smaller objects are packed, with or without “use of a tool”. At final closure, a “tool” is apparently always employed.
Fragmentary data suggest the possibility that some females maintain storage burrows, this perhaps being correlated with maintenance of synchronous nests, all at about the same stage, by some females. More data on these and many other points are needed.
Brief observations on one female A. macra nesting in the same area revealed several important differences between this species and azteca.
References
Adriaanse, A.
1947. Ammophila campestris Latr. und Ammophila adriaansei Wilcke: ein Beitrag zur vergleichenden Verhaltensforschung. Behaviour,
1: 1-34.
1965]
Evans — Ammophila azteca
23
Baerends, G. P.
1941. Fortpflanzungsverhalten und Orientierung der Grabwespe Ammophila campestris Jur. Tijdschr. Ent., 84: 68-275.
Criddle, N,
1924. Observations on the habits of Sphex procera in Manitoba. Canad. Field Nat., 38: 121-123.
Evans, H. E.
1958. The evolution of social life in wasps. Prox. Xth Internat. Congress Ent., 2: 449-457.
1959. Observations on the nesting behavior of digger wasps of the genus Ammophila. Amer. Midi. Nat., 62: 449-473.
1963. Notes on the prey and nesting behavior of some solitary wasps of Jackson Hole, Wyoming. Ent. News, 74: 233-239.
Evans, H. E. and E. G. Linsley
1960. Notes on a sleeping aggregation of solitary bees and wasps. Bull. So. Calif. Acad. Sci., 59: 30-37.
Hartman, C.
1905. Observations on the habits of some solitary wasps of Texas. Bull. Univ. Texas, 65: 1-73.
Hicks, C. H.
1932. Notes on Sphex aberti (Hald.). Canad. Ent., 64: 145-151.
1935. Notes on rare western sphecid wasps. Pan-Pac. Ent., 11 : 97-102. Janvier, H. (F. Claude-Joseph)
1928. Recherches biologiques sur les predateurs du Chili. Ann. Sci. Nat, Zool., (10) 11: 67-207.
Olberg, G.
1959. Das Verhalten der solitaren Wespen Mitteleuropas (Vespidae, Pompilidae, Sphecidae). Berlin, Deutscher Verlag Wissenschaf- ten, 402 pp.
Powell, J.
1964. Additions to the knowledge of the nesting behavior of North American Ammophila (Hymenoptera : Sphecidae). Jour. Kansas Ent. Soc., 37: 240-258.
Teschner, W.
1959. Starrheit und Variabilitat im Verhalten von Sandwespen. Zool. Beitrage, (n.f.)4: 411-472.
Tilde n, J. W.
1953. The digging and provisioning behavior of Ammophila saeva Smith (Hymenoptera: Sphecidae). Pan-Pac. Ent., 29: 211-218.
Tsuneki, K.
1956. Ethological studies on Bembix niponica Smith, with emphasis on the psychobiological analysis of behaviour inside the nest (Hymenoptera, Sphecidae). Mem. Fac. Lib. Arts, Fukui Univ., Ser. II, Nat. Sci., no. 6, pp. 77-172.
Wheeler, W. M.
1923. Social life among the insects. New York, Harcourt Brace & Co., 375 pp.
THE ANT LARVAE OF THE SUBFAMILY LEPTANILLINAE (HYMENOPTERA, FORMICIDAE)
By George C. Wheeler and Jeanette Wheeler University of North Dakota, Grand Forks
The Leptanillinae are a small subfamily comprising 14 species in 3 genera: Leptanilla Emery, 11 species; Leptomesites Kutter, 1 species; Phaulomyrma G. C. & E. W. Wheeler, 2 species.
This subfamily has been recorded only from the warmer parts of the Old World: Corsica, Sardinia, North Africa, India, Malaya,
Java, Queensland, Western Australia and Japan. Seemingly it is rare, but the paucity of the records may be due to minute size and hypogeic habits. As W. M. Wheeler optimistically remarked (1932, p. 54) : “We should expect careful collecting with the Berlese funnel to bring additional forms to light in South Africa, Madagascar, Asia Minor and India, or even, perhaps, in the warmer parts of the New World.’’ Brown (1954, p. 28) noted: “The habits of the species are such as to render their discovery highly fortuitous under present collecting methods.’’ In the 30 years since Wheeler wrote, only two species have been discovered (one in India and one in Japan), which would rather support another statement by Wheeler in the same article (p. 57-58) : “The Leptanillinae . . . must be very ancient, like many other components of the microgenton. . . . L. swani is particularly interesting in this connection, because the extreme south- western corner of Australia, in which it was taken, is known to possess the oldest and least disturbed fauna of any portion of the continent.’’ Antiquity is compatible with a disjunct distribution and with rarity.
The genus Leptanilla was established by Emery in 1870 and then for half a century was kicked about over the taxonomic table. Origi- nally Emery placed it in the “Dorylidae” near Typhlopone. Mayr, however, in a letter to Emery (date not given — see Emery 1904) dissented, maintaining that it belonged with the Myrmicinae. Emery was evidently convinced, for in 1875 he removed it to the “Mirmi- cidei” in the neighborhood of Stenamma and Liomyrmex. In 1877 he moved it to the vicinity of Monomoriuin and Leptothorax in the “Myrmicidei genuini” but marked it with a query to signify un- certain position. In 1881 Ern. Andre had it in the first tribe, “Myrmicidae verae,” of the “Myrmicidae” but mentioned its affini- ties with the “Dorylides.”
Forel in 1893 did not mention the genus but might have meant
24
1965]
Wheeler and Wheeler — Ant Larvae
25
to include it in the subfamily Myrmicinae, when he said that the fourth tribe (“Myrmicii”) included “les autres genres de la sous- famille des Myrmicinae.” And later (1901) he excluded it from the Dorylinae when he says, “Done, je maintiens la sous-famille Dory- linae limitee aux genres Dorylus, Aenictus, Eciton et Cheliomyrmex In Dalla Torre’s “Catalogus” (1893) it was still in the Myrmicinae but near Trigonogaster and Pheidologeton. In 1895 Emery was still of the opinion that it belonged in the subfamily “Myrmicini” in the second tribe (“Myrmicii”) near Huberia and Phacota. But nine years later (1904), after describing the female of L. revelierei, he returned it to its original subfamily (Dorylinae). In the following year in Ashmead’s skeleton it stood between “ fLiomyrmex” and Epipheidole in the tribe Stenammini, subfamily Myrmicinae, family “Myrmicidae.” In 1907 Santschi described males of 3 species, which he referred to the genus Leptanilla and claimed that their doryline affinities justified Emery’s original allocation of the genus. It should be noted, however, that males of Leptanilla have never been taken with females or workers; hence it is not certain that those described by Santschi belong to this genus. In the “Genera Insectorum” Emery (1910) established for the genus a separate tribe (Leptanillini) in the subfamily Dorylinae, where it seemed destined to abide in iso- lation: Wheeler (1910), Forel (1917 and 1923), and Wheeler (1922) did not disturb it. Wheeler (1923, p. 335) suggested that even further isolation might be necessary: “I believe that the tribe Leptanillini, which Emery includes among the Dorylinae, will have to be separated out as a distinct subfamily (Leptanillinae) . Dr. George C. Wheeler finds that the larva of Leptanilla is very aberrant, and the characters of the adult are either quite unlike those of other Dorylinae or only superficially similar and due to convergence, or similarity of subterranean habits.” By 1932 (p. 57) the separation had been effected : “Emery, as is well known, regarded the Leptanil- linae as constituting a special tribe of the Dorylinae, but Dr. G. C. Wheeler and I have raised the group to subfamily rank.” Bernard (1951) raised the taxon to family rank.
Brown and Nutting (1950, p. 124) in their study of the wing venation for the family Formicidae wrote: “The position of the leptanillines is debatable. They are usually treated as a subfamily by modern authors, and the venation is so highly reduced in the forms we have seen that little may be deduced from them concerning relationships to the other subfamilies.” Brown was apparently still puzzled in 1954 (p. 28) for he wrote: “This little subfamily has suffered such drastic anatomical reduction in most of the usually
26
Psyche
[March
valuable phylogenetic characters that it is doubtful whether we shall ever be certain of its true affinities. ... It is possible that forms as yet unknown will reveal their ancestry more clearly. Until that time, however, subfamily rank for the Leptanillinae may as well be main- tained. Present opinion seems to favor relating this group to the Dorylinae ”
The larvae of the Leptanillinae do bear certain resemblances to the known doryline larvae (Dorylus, Aenictus , Eciton, Cheliomyrmex) : the long slender body and the small feebly sclerotized mandibles. On the other hand, they differ in the constriction at the metathorax, the long hairs on the abdomen, the absence of hairs on the head and the shape of the head. But these differences become trivial and insignificant beside the three which not only differentiate them from the Dory- linae, but also from all other known formicid larvae (160 genera) :
( 1) the peculiar projection from the ventral surface of the prothorax;
(2) the reduction of the spiracles from the normal ten pairs to a single pair, which is located on abdominal somite III; and (3) the shape and stance of the mandibles.
Subfamily leptanillinae Emery
Elongate and very slender; slightly constricted at the metathorax; anterior end curved ventrally ; remainder of body straight and clavate. With a curious complex structure projecting anteroventrally from the ventral surface of the prothorax. Only one pair of spiracles, located on abdominal somite III; each spiracle opening eccentrically on a naked circular area. Body hairs simple ; the minute hairs very abundant and uniformly distributed ; a few conspicuous long hairs sparsely scattered. Head naked and elongate. Antennae small, slightly^ raised, each with two sensilla. Labrum slightly broader than long, with the ventral border semicircular; posterior surface spinulose, with the spinules in rows. Mandibles apparently turned laterally (instead of medially, as is usual with ant larvae) ; feebly sclerotized; each with a rather long slender sharp-pointed apical tooth, which curves laterally; lateral (=outer) border furnished with several long slender sharp-pointed teeth; anterior surface with rows of spinules. Labium thin, flap-like and narrowed basally; lateral surfaces sclero- tized; each palp a low ventrolateral projection bearing five sensilla.
Of all the ant larvae studied we have found the Leptanillinae the most difficult to process partly because of their small size but chiefly because of their extreme slenderness. A slender larva is more apt to collapse than a stout one. Furthermore the constriction at the metathorax aggravates the difficulty of cleaning and predisposes to breakage in that region. As a result processing has often left us only
1965]
Wheeler and Wheeler — Ant Larvae
27
fragments to mount on slides. Fragments are all right — in fact, necessary for some parts — if none is lost and if they can be correctly oriented.
References. — W. M. Wheeler (1923, p. 335) used larval characters in establishing the subfamily (quoted above and also by G. C. Wheeler, 1928, p. 88-89 and referred to by G. C. and E. Wheeler, 1930, p. 198).
G. C. Wheeler (1928, p. 89) justified W. M. Wheeler’s establish- ment of the subfamily. (Repeated G. C. and E. Wheeler, 1930, p. 199.)
Kutter 1948 p. 294: “Alle bis jetzt bekannt gewordenen Larven der Leptanillinae haben den teilweise chitinisierten, ventralen Thorakalanhang gemein, wie offenbar auch die Senkrechtstellung der Mandibeln, wahrend der Besitz des als Tympanalorgan bezeichneten Organs noch nicht als typisch fiir alle Larven der Unterfamilie bezeichnet werden darf.”
Bernard, 1951, p. 1041 : “Larves eucephales, carnivores; nourries par les ouvrieres.”
Genus Leptanilla Emery
We are unable to separate Leptanilla generically from Leptome- sites: the difference between the larvae of the two known species of the former are as great as the difference between either species and the larva of the latter. Therefore the subfamilial description will suffice for the genus.
Bernard (1951, p. 1017) described primitive larvae and mentioned the larva of Leptanilla as an example.
Kutter ( 1948, p. 292) differentiated the two genera by the absence of the naked area around the spiracle (“tympanum”) and the structure of the ventral prothoracic projection. As we show below, this dis- tinction is no longer tenable.
The two species of Leptanilla differ with respect to the following characters: size and shape of terminal boss; complexity of ventral prothoracic projection; size and arrangement of long body hairs; shape of head ; number of teeth and spinules on the mandibles; and the sclerotized band bordering the spiracular area.
Leptanilla revelierei sardoa Emery
(Fig. 1-8)
Length approximately 1.3 mm. Body elongate and slender; thorax slightly curved ventrally, slightly constricted at the metathorax; abdomen straight and clavate, the diameter increasing gradually to abdominal somite V and decreasing to the posterior end which is
Psyche, 1965
Vol. 72, Plate 2
Wheeler — Ant Larvae
1965]
Wheeler and Wheeler — Ant Larvae
29
rounded and terminates in a small naked hemispherical boss. With a complex structure projecting anteroventrally from the ventral surface of the prothorax (see Fig. 4-5). With only one pair of spiracles, located on the third abdominal somite near the posterior border; each spiracle opening eccentrically on a naked circular area, which is bordered by a narrow heavily sclerotized band. Body hairs simple. Of three types : ( 1 ) numerous, increasing in length from
0.005 mm on the thorax to 0.02 mm at the posterior end, without alveolus and articular membrane, uniformly distributed but lacking from the anterior portion of the prothorax, the circumspiracular area and the terminal boss; (2) 0.01-0.15 mm long (shortest on the prothorax), few, conspicuous, the longest attenuated and flexible distally, with alveolus and articular membrane, absent from the dorsal surface of abdominal somites III-IX; (3) about 0.3 mm long, with attenuated flexible tip, two (one dorsal and one ventral) on the posterior end near the terminal boss. Head naked; elongate; widest above the antennal level ; cranium subpyriform in anterior view. Antennae small ; slightly raised ; each with two sensilla, each of which bears a spinule. Labrum slightly broader than long; the ventral border semicircular; posterior surface spinulose, the spinules in long rows, the rows concentric with the ventral border. Mandibles with the toothed border directed laterally; feebly sclerotized; each with a rather long slender sharp-pointed apical tooth, which is curved laterally; lateral border with four long slender sharp-pointed teeth; a few rows of rather large spinules on the anterior surface. Maxillae adnate to the labium; palp a stalked hemispherical knob directed laterally and bearing five sensilla; no galea seen. Labium a thin flap, narrowed basally; lateral surfaces sclerotized; each palp a low ventrolateral projection bearing five sensilla; an isolated sensillum between each palp and the opening of the sericteries; the latter a transverse impression on the anterior surface.
Explanation of Plate 2
Figs. 1-8. Leptanilla revelierei sardoa. 1, head in anterior view, X417; 2, larva in side view, X76; 3, left mandible in anterior view, X847; 4, ventral prothoracic structure in side view, X423 ; 5, ventral prothoracic structure: left half in posterior view, right half in anterior view, X423 ; 6-8, three body hairs, X500.
Figs. 9-18. Leptanilla swani. 9, head in anterior view, X417 ; 10, larva in side view, X76; 11, left mandible in anterior view, X 1333 ; 12, left maxilla in anterior view, X847; 13 and 14, two body hairs, X500; 15 and 16, spiracle and circumspiracular area in side and surface phantom views (hairs omitted) X333; 17 and 18, ventral prothoracic structure in anterior view and in side view, X667.
30
Psyche
[March
Material Studied. — Three larvae labeled “Sardegna: Golfo
Aranci. I. 1909 A. Dodero.” These are the specimens studied by G. C. Wheeler (1928). We studied them first with a phase microscope; then they were dismounted, stained lightly, remounted and studied under both phase and light microscopes.
Literature. The description and figures by G. C. Wheeler (1928, p. 85-87) have been completely revised for this article. Bernard, 1951, Fig. 949 C, D after G. C. Wheeler, 1928.
Leptanilla swani Wheeler (Fig. 9-18)
Length about 1.4 mm. Body elongate and very slender; anterior end curved ventrally, slightly constricted at the metathorax ; remainder of body straight and clavate; diameter increasing gradually to abdominal somite VII and diminishing to the posterior end, which is round-pointed and terminates in a small naked boss. With a complex structure projecting anteroventrally from the ventral surface of the prothorax (see Fig. 17-18). With only one pair of spiracles, located near the posterior border of abdominal somite III; each spiracle opening eccentrically on a naked circular area. Body hairs simple. Of two types: (1) abundant and uniformly distributed (except on the terminal boss and the circumspiracular area), minute (0.005-0.024 mm long), longest near the posterior end, without alveolus and articular membrane; (2) long (0.084-0.23 mm), slender, with the apical portion fine and flexible, with alveolus and articular membrane, few, conspicuous, absent from the dorsal surface of the abdomen. Head naked; elongate; widest at the level of the antennae; cranium subpyriform in anterior view. Antennae small and slightly raised; each with two sensilla, each of which bears a spinule. Labrum slightly broader than long, with the ventral border semicircular; posterior surface spinulose, the spinules rather numerous and long, arranged in rows concentric with the ventral border. Mandibles with the toothed border directed laterally; feebly sclerotized; each with a rather long slender sharp-pointed apical tooth, which curves laterally; lateral border with six long slender sharp-pointed teeth; anterior surface spinulose, the spinules numerous, rather long and arranged in rows; posterior surface with one row of long spinules. Maxillae adnate to the labium; palp an irregular knob projecting laterally and bearing five sensilla; no galea seen. Labium a thin flap, narrowed basally; lateral surfaces sclerotized; each palp a low ventrolateral projection bearing five sensilla; an isolated sensillum between each palp and the opening of the sericteries; the latter a transverse slit on the ventral border.
1965]
Wheeler and Wheeler — Ant Larvae
31
Material Studied. — Three larvae from Glittering, Western Australia collected by D. C. Swan. These are the specimens referred to by W. M. Wheeler in 1932 (p. 56-57). We studied them first with a phase microscope; then they were dismounted, stained lightly, remounted and studied under both phase and light microscopes.
In 1963 Rev. B. B. Lowery of St. Ignatius College (Sydney, Australia) generously sent us 40 larvae of this species, which he had collected at Cunningham’s Gap in southern Queensland. The two best specimens have been kept in alcohol. Many others were processed according to our standard technique (i960) and studied under both phase and light microscopes.
Literature. W. M. Wheeler (1932, p. 56-57) compared the larva of L. swani with that of L. sardoa by quoting from a letter from G. C. Wheeler to whom he had sent the larvae for study.
Genus Leptomesites Kutter
As explained above under the genus Leptanilla we are unable to separate Leptomesites generically from Leptanilla. Therefore our subfamilial description will have to suffice for this genus.
Kutter (1948, p. 287) has characterized the genus thus: “Korper sehr langgestreckt-zylindrisch, gegen das Ende keulenformig aufge- trieben. Labrum ohne gezahnte, laterale Lappchen. Mandibeln abwarts gerichtet, gezahnt. Auf der Ventralseite des Prothorax mit breitem, in der Mittelpartie mit chitinisiertem, quergerilltem, lappigem Anhang. Beiderseits dieses einzigartigen Anhanges je eine sehr lange, sichelformig ventralwarts gebogene Borste. (Ein Tympanalorgan, wie es von der Leptanilla-Larve gemeldet wird, konnte nicht mit Sicherheit nachgewiesen werden.)”
Leptomesites escheri Kutter (Fig. 19-28)
Length about 1.5 mm. Body elongate and slender, apparently constricted at the metathorax; anterior end slightly curved ventrally; abdomen clavate, with the posterior end narrowly rounded. Project- ing ventrally from the anterior portion of the prothorax is a curious flap-like structure furnished with 14 ridges on the posterior surface. Only one pair of spiracles present, on abdominal somite III (or IV?). Body hairs simple. Of two types: (1) minute (0.009-0.026 mm
long), longest near the spiracle, numerous, uniformly distributed, without alveolus and articular membrane; (2) short to very long (0.042-0.096 mm), a few on each somite, the apical portion fine and flexible, with alveolus and articular membrane. Head naked; elongate; widest above the antennal level ; cranium subovoidal. Antennae small
32
Psyche
[March
Figures 19-28. Leptomesites escheri. 19, head in side view, X278; 20, head in anterior view, X278; 21, left mandible in anterior view, X423; 22, left maxilla in anterior view, X415; 23, restoration of larva in side view (minute hairs omitted), X33; 24-26, three body hairs, X139; 27 and 28, ventral prothoracic structure in side view and in anterior view, X272.
and slightly raised ; each with two sensilla, each of which bears a spinule. Labrum slightly broader than long; the ventral border semicircular; posterior surface spinulose, the spinules in long rows, the rows concentric with the ventral border. Mandibles with the toothed border directed laterally; feebly sclerotized; each with a rather long slender sharp-pointed apical tooth, which is curved laterally; lateral border with nine long slender sharp-pointed teeth; anterior surface furnished with numerous short rows of long spinules, the rows so close together that the spinules overlap. Maxillae conoidal,
1965]
Wheeler and Wheeler — Ant Larvae
33
pointing laterally; palp apical and consisting of five sensilla; galea represented by two sensilla, each of which bears a digitiform spinule. Labium feebly bilobed, flap-like, narrowed basally; palp a low ventro- lateral projection bearing five sensilla; an isolated sensillum between each palp and the opening of the sericteries; the latter a transverse slit on the ventral surface. Hypopharynx spinulose, the spinules in short transverse rows, the rows so close together that the spinules overlap.
Material Studied: — Two semipupae (?) from southern India. These are the specimens studied by Dr. Heinrich Kutter (1948, p. 290-292) of Mannedorf, Switzerland. In 1963 we asked Dr. Kutter whether he would be willing to let us examine them. His generous response was to send them to us as a gift. We studied them first with the phase microscope; then they were dismounted, stained lightly, remounted and studied under both phase and light microscopes. We were not able to inflate the wrinkled specimens; therefore our drawing of the larva in side view is a restoration.
Literature. Kutter (1948, p. 290-292) : a description of the
species; Fig. 6, larva in side view; Fig. 7, anterior end enlarged.
Literature Cited
Andre, Ern.
1881-1882. Species des hymenopteres composant le groupe des formicides d’Europe, etc.
Ashmead, W. H.
1905. A skeleton of a new arrangement of the families, subfamilies, tribes and genera of the ants, or the superfamily Formicoidea. Canad. Ent. 37: 381-384.
Bernard, F.
1951. Super-famille des Formicoidea. In: P. P. Grasse (ed.), Traite de Zoologie, 10 (2) : 997-1104. Masson et Cie, Paris.
Brown, W. L.
1954. Remarks on the internal phylogeny and subfamily classification of the family Formicidae. Insectes Sociaux 1: 21-31.
Brown, W. L., and W. L. Nutting
1950. Wing venation and the phylogeny of the Formicidae. Trans. Amer. Ent. Soc. 75: 113-132, 2 pi.
Dalla Torre, K. W. von
1893. Catalogus Hymenopterorum hucusque descriptorum systematicus et synonymicus. Vol. VII: Formicidae (Heterogyna) . W.
Engelmann, Leipzig. 289 p.
Emery, C.
1870. Studi mirmecologici. Boll. Soc. Ent. Ital. 2: 9 p, 1 pi.
1875. Le formiche ipogee con descrizioni di sp. nuove o poco note. Ann. Mus. Civ. Genoa 7: 465-474.
1877. Saggio di un ordinamento naturale dei myrmicidei e consider- azioni sulla filogenesi delle formiche. Boll. Soc. Ent. Ital. 9: 1-17, 1 pi.
34
Psyche
[March
1895. Die Gattung Dorylus Fabr. und die systematische Eintheilung der Formiciden. Zool. Jahrb. Abt. Syst. 8: 685-788.
1904. Le affinita del genere Leptanilla e i limiti delle Dorylinae. Arch. Zool. Napoli 2: 107-116.
1910. Fam. Formicidae, Subfam. Dorylinae in Wytsman’s “Genera Insectorum.” Fascicle 102: 34 p. 1 pi.
Forel, A.
1893. Sur la classification de la famille des formicides, avec remarques synonymiques. Ann. Soc. Ent. Belg. 37: 161-167.
1901. A propos de la classification des fourmis. Ann. Soc. Ent. Belg. 45: 136-141.
1917. Cadre synoptique actuel de la faune universelle des fourmis.
Bull. Soc. Vaud. Sci. Nat. 51: 229-253.
1923. Le monde social des fourmis du globe. 5 : 174 p., 3 pis. Librairie Kundig, Geneva.
Kutter, H.
1948. Beitrag zur Kenntnis der Leptanillinae. Eine neue Ameisengat- tung aus sud-Indien. Mit. Schweiz. Ent. Ges. 21: 286-295.
Santschi, F.
1907. Fourmis de Tunisie capturees en 1906. Rev. Suisse Zool. 15:
305-334.
Wheeler, G. C.
1928. The larva of Leptanilla. Psyche 35 : 85-91.
Wheeler, G. C., and Esther W. Wheeler
1930. Two new ants from Java. Psyche 37: 193-201.
Wheeler, G. C., and Jeanette Wheeler
1960. Techniques for the study of ant larvae. Psyche 67 : 87-94. Wheeler, W. M.
1910. Ants: Their structure, development and behavior. Columbia
University Press, New York. 663 p.
1922. Ants of the American Museum Congo Expedition. Bull. Amer. Mus. Nat. Hist. 45: 1-1139.
1923. Social life among the insects. Harcourt, Brace & Co., New York. 375 p.
1932. An Australian Leptanilla. Psyche 39: 53-58.
THE AUSTRALIAN ANTS OF THE GENUS PRISTOMYRMEX , WITH A CASE OF APPARENT CHARACTER DISPLACEMENT1
By Robert W. Taylor Biological Laboratories, Harvard University
Introduction
The Old World myrmicine ant genus, Pristomyrmex Mayr 1866, contains about 40 named forms, mostly from rain forests of the Ethiopian, Oriental and Australian regions. The Australian mainland fauna is small compared to that of New Guinea; it includes four species, three of which are newly described here. There are two species-groups, those of P. foveolatus n. sp., and P. quadridentatus (Andre) 1905, each containing two species.
The former group includes P. foveolatus and P. thoracicus n. sp., closely related species with strong foveolate-reticulate sculpturation, having clear affinities with components of the New Guinean fauna, and known only from rain forests in the Cairns district of North Queensland. Foveolatus seems to be restricted to the Atherton Table- land, at elevations greater than 1000 ft, while thoracicus is apparently more widespread, but is not known from the Tableland proper.
The affinities of the quadridentatus group are less clear; its species occur further south, in extreme S.E. Queensland, and N.E. New South Wales, and they may be derived from a separate, older, Mela- nesian stock. The group includes P. quadridentatus and P. wheeleri n. sp. Quadridentatus was originally placed in a separate monotypic genus, Odontomyrmex Andre 1905, which was synonymized with Pristomyrmex by Emery ( 1922). The lack of foveolate sculpturation on the head and body, and the presence of pronotal spines were thought by Andre to taxonomically characterize Odontomyrmex . But these characters occur together or separately in many Pristomyrmex species, and cannot be used to logically define a separate taxon (Brown, 1953). Nonetheless the “Odontomyrmex” habitus seems to characterize a natural phyletic group within the Australian fauna, and quadriden - tatus and wheeleri are undoubtedly cognate, or at least closely related. The former species is typically self-colored golden-brown, but it has an apparently conspecific bicolored variant, quadridentatus variety
Research supported by U.S. National Science Foundation Grant No. GB 1634 (1963-1965). Australian field studies in 1962, reported here, were supported by the Committee on Evolutionary Biology, Harvard University.
35
36
Psyche
[March
queenslandensis Forel 1915, which is synonymized below. These two forms differ only slightly, except in coloration, and distributional evidence suggests that the bicolored form is a character-displaced variant, produced wherever the range of quadridentatus overlaps that of wheeleri (see fig. 15, plate 3, and discussion on p. 53). Like their northern congeners the quadridentatus-g roup species are known only from rain forest, usually at elevations of ■+ 1000 ft.
Available biological information is given in the taxonomic section to follow. Nesting apparently occurs in rotting logs in all species except wheeleri, colonies of which are usually located in the soil, under stones. Feeding and foraging behavior have not been extensively studied, but the bicolored form of quadridentatus forages nocturnally, in the open, on logs, tree trunks and low foliage. It appears that wheeleri may restrict its foraging activity to the soil and leaf litter, since strays have not been taken elsewhere, in spite of extensive search by the author.
All Australian Pristomyrmex types are workers, and the females are known for all species except foveolatus ; I have seen no male specimens. Larvae of a North Queensland species were described by G. C. and J. Wheeler (1954). I have been unable to locate adults from the series involved ; they would conform to the foveolatus-group diagnosis according to their collector, Dr. W. L. Brown, Jr.
This study is based on material from the Museum of Comparative Zoology collection at Harvard University, from Entomology Division, CSIRO, Canberra, The National Museum of Victoria, Melbourne, and the collection of Rev. B. B. Lowery, S.J., Sydney.
MEASUREMENTS AND INDICES
The following abbreviations are used in the descriptions below. Measurements were prepared using a stereomicroscope fitted with an eyepiece scale, reading directly in units of 0.1 and 0.01 mm at 100 X magnification. The measurements cited are the maxima obtained after a series of trials, a fact of special importance when dealing with HL and WL.
HL Maximum head length, frontal view, from apex of median clypeal denticle to median occipital border.
HW Plead width, frontal view, including the eyes.
Cl Cephalic index (HW X ioo)/HL
SL Scape length, excluding articular boss and condyle ( hut in- cluding the expanded lamella at base of shaft).
SI Scape index (SL X ioo)/HW
PW Pronotum width, anterior view, immediately below the bases of the pronotal spines.
1965]
T ay lor — Pristomyrmex
37
Scutum Width (females only), dorsal view, across the widest part of the sclerite.
WL Weber s length of mesosoma, lateral view; in worker , maximum diagonal distance from junction of anterior and dorsal faces of pronotum to apex of inferior metapleural lobe; in female, the anterior limit of measurement is the point where the pro-mesono- tal suture meets the anterodorsal mesosomal profile.
Pronotal Spine L(ength), as seen in PW measuring position, true vertical distance between base of spine on its mesial side, and the level of its apex.
Propodeal Spine L(ength), lateral view, maximum distance from apex of spine to opposite edge of propodeal spiracle.
Petiole L(ength), lateral view, from tip of metapleural spine to posterodorsal edge of petiolar tergite.
Postpetiole Height, lateral view, maximum distance from apex of subpostpetiolar process to postpetiolar dorsum.
Eye Diameter, maximum diameter of eye.
Ocular Index (Eye Diameter X ioo)/HW.
Key to the Australian Pristomyrmex Species (Based on the workers)
1. Front of head, between the antennal scrobes, with numerous large shallow umbilicate foveae, their average diameter about 0.05-0.06
mm (group of P. foveolatus) 2
Area of frons between antennal scrobes entirely smooth, lacking sculpturation apart from a few minute, pilosity-bearing punctures (group of P. quad rid en tat us ) 3
2. Scapes barely reaching level of occipital border when at rest in the antennal scrobes, with the head held in full-face view
(fig. 1 ) ; Scape Index 78-83 P. foveolatus n. sp.
Scapes exceeding level of occipital border by 1-1.5 X their maximum thickness when at rest in the antennal scrobes, with the
head viewed full-face (fig. 3), Scape Index 96-101
P. thoracicus n. sp.
3. Head relatively narrow (Cephalic Index 97-103) (fig. 7), either concolorous with mesosoma or much lighter in color (plate 3) ; pronotal spines, in side view, distinctly longer and stronger than
those of propodeum (fig. 8) P. quadridentatus (Andre)
Head relatively much broader (Cephalic Index 110-116) (fig.
1 1 ) and usually much darker in color than the mesosoma ( plate 3) ; pronotal spines subequal in size to propodeal ones, or very slightly smaller (fig. 12) P. wheel eri n. so.
38
Psyche
[March
Species-level Taxonomy i. Pristomyrmex foveolatus new species (Figs, i, 2)
Holotype worker. HL 0.84 mm; HW 0.92 mm; Cl no; SL 0.75 mm; SI 81 ; PW 0.55 mm; WL 0.79 mm; Propodeal Spine L 0.35 mm; Petiole L 0.30 mm; Postpetiole Height 0.30 mm. General features as in figures 1 and 2. Mandibles with two adjacent strong acute apical teeth and a truncated basal one, separated by a broad, shallowly concave diastema. Clypeus with a median longitudinal Carina, its anterior border with 5 obtuse denticles, the median and outer lateral ones largest. Antennal scrobes shallow, their ventral borders obscure. Occipital area of cranium enclosed dorsally and laterally by a fine arched carina. Scapes barely attaining level of occipital border when lying in the antennal scrobes, with the head viewed in HL measuring position (see p. 36). Funiculus with a 3-segmented club, proportions of its segments as in Figure 1. Eyes moderately large, maximum diameter 0.12 mm, their anterior edges almost twice this distance from the anterior genal borders. Mesosomal profile as in Figure 2. Pronotum truncated anteriorly, its inferior angles edentate. Humeri each with a small acute tooth about 0.05 mm long, inclined dorsolaterally to project beyond the main lateral outline of the pronotum, when viewed from above. Posterolateral corners of pronotal dorsum forming blunt angles of about 90 degrees in dorsal view. Anterodorsal edge of mesepisternum extended anterior- ly as a strong, moderately acute process which partially overlaps the fore-coxa. Propodeal spines long, almost straight, each diverging from the dorsal midline at an angle of about 20 degrees. Metapleural spines small, moderately acute, each about 0.07 mm long, slightly wider at base. Petiolar and postpetiolar profiles as in Figure 2. Petiole slightly narrower than postpetiole when viewed from above, its crest transverse in posterior view and feebly convex.
Mandibles generally smooth, with several longitudinal ribs. Clypeus strongly shining. Head polished and reflective, with numer- ous large shallow subcircular umbilicate foveae, averaging about 0.06 mm in diameter, largest on the posterior genal surfaces, lacking on a small smooth area between eye and antennal socket on each side, generally separated by distances averaging about one-third their mean diameter. Postgenae obscurely foveolate. Antennal scrobes smooth, with an almost effaced rugulosity and several arched transverse striae anteriorly. Pronotal dorsum obscurely foveolate, with several median longitudinal costulae. Posteriorly these meet a transverse carina which apparently represents the mesonotum. Propodeal dorsum smooth,
1965]
Taylor — Pristo??iyrmex
39
framed by lateral and posterior carinae. Sides of mesosoma generally smooth, ventral parts of mesepisternum, metepisternal area, and sides of propodeum obscurely roughened. Petiole, postpetiole and gaster, antennae and legs, smooth and shining.
Pilosity moderately long and abundant on head and mesosoma, erect to sub-erect, longest on clypeus, crown of head, and pronotal dorsum; shorter erert hairs on scape, sub-erect ones on mandibles, funiculus and legs. Hairs lacking on sides of mesosoma. Petiolar and postpetiolar tergites each with three bilaterally symmetrical pairs of hairs (positions as in Figure 2), postpetiolar sternite with a single pair. Gaster very finely pilose apically, its first tergite completely bare. Pubescence lacking.
Color rich golden sienna-brown, mandibles, clypeus, antennae and legs lighter, eyes black, pilosity whitish.
Type-locality. Clump Point, west of Tully, N. Queensland.
The holotype was taken in a rotting wood fragment on the floor of very disturbed, depauperate rain forest, a few feet above sea level (June 25, 1962; R. W. Taylor acc. 1746).
Paratype workers. Three specimens collected with the holotype resemble it closely; they have the following dimensions: HL 0.82- 0.88 mm; HW 0.87-0.96 mm; Cl 106-109; SL 0.72-0.75 mm; SI 78-83; PW 0.52-0.57 mm; WL 0.77-0.79 mm; Propodeal Spine L 0.32-0.35 mm; Petiole L 0.29-0.33 mm; Postpetiole Height 0.28-0.30 mm.
An additional specimen from rain forest at Thornton Peak (Mt. Alexander), north-west of Daintree, N. Queensland (elevation not specified but probably + 1000 ft (Darlington, i960: 1 1 7 ) , December 20-23, I957i P- F. Darlington) has: HL 0.85 mm; HW 0.95 mm; Cl 1 12; SL 0.75 mm; SI 79; PW 0.53 mm; WL 0.78 mm; Pro- podeal Spine L 0.32 mm; Petiole L 0.28 mm; Postpetiole Height 0.29 mm.
This species may show slight unimodal polymorphism. In the short series available SL, and the mesosomal dimensions PW and WL are negatively allometric relative to HW, while HL is positively allo- metric.
Distribution and Material Examined. Known only from the five workers listed above. Clump Point and Thorton Peak are within the area of the main “base-of-peninsula” rain forest system of North Queensland.
Type deposition. Holotype and three paratypes in Museum of Comparative Zoology (MCZ) at Harvard University (Type No.
40
Psyche
[March
Figures 1-6: North Queensland species of Pristomyrmex foveolatus species- group — P. foveolatus n. sp., holotype worker: fig. 1, head, frontal view; fig. 2, mesosoma, petiole and postpetiole, lateral view. — P. thoracicus n. sp., holotype worker: fig. 3, head, frontal view; fig. 4, mesosoma, petiole and postpetiole, lateral view. Paratype female: fig. 5, head, frontal view; fig. 6, mesosoma, petiole and postpetiole, lateral view.
1965]
Taylor — Pristomyrmex
41
31 152); one paratype (from Clump Point) in Australian National Insect Collection (CSIRO), Canberra.
2. Pristomyrmex thoracicus new species (Figs. 3-6)
Worker. The following description is based on the holotype and 14 paratypes collected with it. Dimensions (holotype cited first) : HL 0.95, 0.89-0.96 mm; HW 0.95, 0.89-0.99 mm; Cl 100, 98-104; SL 0.93, 0.90-0.96 mm; SI 98, 96-101; PW 0.49, 0.45-0.51 mm; WL 0.91, 0.87-0.94 mm; Propodeal Spine L 0.30, 0.28-0.32 mm; Petiole L 0.34, 0.33-0.36 mm ; Postpetiole Height 0.32, 0.28-0.32 mm. General features as in Figures 3 and 4. Similar to P. foveolatus, differentiated from it by the following characters:
1. Relatively narrow head (Cl 98-105 against 106-110 in foveo- latus) .
2. Much longer scapes, which surpass the level of the occipital border by 1-1.5 X their maximum thickness when lying in the antennal scrobes, with the head viewed in the HL measuring position (SI 96-101 against 78-83).
3. Proportionately much longer mesosoma (WL 0.94-1.00 X HW, against 0.84-0.89 X in foveolatus) , which is relatively very narrow (PW 0.5 1-0.53 X HW, and 0.50-0.55 X WL against 0.56-0.62 X and 0.67-0.73 X respectively in foveolatus) .
4. Eyes relatively slightly smaller; antennal scrobes more distinctly outlined. Humeral teeth more erect, failing to project beyond lateral pronotal outline when viewed from above. Posterolateral corners of pronotal dorsum obtuse, broadly rounded in dorsal view. Propodeal spines slightly sinuate, their apices turned upwards; diverging from the dorsal midline at angles of about 20 degrees. Metapleural spines more acute. Petiolar crest strongly convex in dorsal view, its outline almost a half circle.
5. Sculpturation virtually identical. Cephalic foveae a little more spaced ; sides of mesosoma less sculptured, generally smooth and shining.
6. Pilosity similarly distributed, the hairs shorter, especially on the crown of the head, and the pronotal dorsum; less abundant on pro- podeal dorsum. Hairs of petiole, postpetiole and gaster placed exactly as in foveolatus.
Type-locality. Vision Falls, Lake Eacham National Park, N. Queensland. The type nest series was taken from a rotting log in rain forest, elevation c. 2500 ft (June 6, 1962; R. W. Taylor acc. 1399)-
42
Psyche
[March
Additional paratype workers. Three workers from Beatrice River, Crawford’s Lookout, N. Queensland (elevation probably c. 1000 ft. (Darlington, i960: 118), rain forest, February 21-22, 1958;
Darlingtons) and a single worker from Kuranda, N. Queensland (1100 ft, rain forest, July 4, 1962, R. W. Taylor acc. 1358) have the following dimensions (Crawford’s Lookout material cited first) : HL 0.86-0.89, 0.89 mm; HW 0.87-0.90, 0.93 mm; Cl 101-102, 105; SL 0.88-0.89, 0.90 mm; SI 99-101, 97; PW 0.45-0.47, 0.47 mm; WL 0.84-0.85, 0.88 mm; Propodeal Spine L 0.27-0.28, 0.31 mm; Petiole L 0.31-0.33, 0.33 mm; Postpetiole Height 0.29-0.31, 0.31 mm. Polymorphism weak, as in P. foveolatus — SL, PW and WL nega- tively allometric relative to HW, HL positively allometric.
Paratype females. Two females are present in the type series, one from the holotype nest, the other from the Crawford’s Lookout collection. Dimensions (Lake Eacham specimen cited first) : HL 0.96, 0.95 mm; HW 1.04, 1.02 mm; Cl 109, 108; SL 0.95, 0.92 mm; SI 91, 90; Scutum W 0.67, 0.65 mm; WL 1.09, 1.04 mm; Propodeal Spine L 0.32, 0.3 1 mm; Petiole L 0.38, 0.38 mm; Post- petiole Height 0.35, 0.35 mm; Eye Diameter 0.17, 0.17 mm; Ocular Index 16, 17. General features as in Figures 5 and 6. Pronotum without humeral spines; mesoscutum lacking parapsidal lines, pro- podeal spines diverging from dorsal midline at angles of 30 and 33 degrees. Mesonotal sclerites moderately coarsely longitudinally costulate, with vestiges of interstitial foveae. Pilosity as in worker; petiolar and postpetiolar hairs similarly arrayed, the petiolar peduncle with an additional pair of lateral hairs.
Color as in worker; each ocellus with an adjacent black spot; the mesosomal suture lines darkly infuscated.
Distribution and material examined. Known only from the 19 workers and two females discussed above. Lake Eacham, Crawford’s Lookout and Kuranda are all at moderately high elevations on the Atherton Tableland, within the compass of the main “base-of- peninsula” rain forest system of N. Queensland.
Type deposition. Holotype, 11 paratype workers and paratype female (from holotype nest) in MCZ (Type No. 31 153); four paratype workers and female in CSIRO; single paratypes in Queens- land Museum, Brisbane; National Museum of Victoria, Melbourne; and British Museum (Nat. Hist.), London.
3. Pristomyrmex quadridentatus (Ern. Andre)
(Figs. 7-10, Plate 3)
Odontomyrmex quadridentatus Ern. Andre, 1905, Rev. Ent. Caen., 24: 207, worker. Type-locality: Sydney.
1965]
Taylor — Pristomyrmex
43
Pristomyrmex quadridentatus : Emery, 1922, Gen. Ins., 174c: 234. Pristomyrmex (Odontomyrmex) quadridentatus v. queenslandensis Forel, 1915, Arkiv F. Zool., 9 (16): 53, worker. Type-locality: Mt. Tambou- rine (— Tamborine Mountain). NEW SYNONYMY.2 Pristomyrmex quadridentatus var. queenslandensis : Emery, 1922, Gen. In9., 174c: 234.
Syntype workers. Two quadridentatus syntypes, each labeled “Sydney/Duchaussoy/TYPE” have kindly been loaned for study by Mile. S. Kelner-Pillault of the Museum National D’Histoire Natu- relle, Paris. These specimens are here designated lectotype and paralectotype and are so labeled. The paralectotype gaster is missing.
Judging from subsequent records the Sydney type-locality may be outside the true range of this species and should be regarded with reservation as a distributional record until confirmed.
Dimensions as given in Table i (lectotype cited first). General features as in Figures 7 and 8. Mandibular dentition as in P. foveo- latus. Clypeus with a median longitudinal carina, its anterior border with three separated, moderately acute denticles (apparently repre- senting the median and outer-lateral ones of foveolatus) . Antennal scrobes open posteriorly but well defined dorsally and ventrally by distinct sharp sub-parallel carinae. Occipital carina distinct. Scapes, in frontal view, laid in the scrobes, exceeding the level of the occipital border by about 1.5 X their maximum thickness. Funicular pro- portions as in Fig. 7. Ocular Index 17-18.
Mesosomal profile as in Figure 8. Pronotum with a distinct transverse anterior carina. Pronotal spines almost vertical, consider- ably stronger than propodeal ones. Mesonotum represented by a transverse carina on mesosomal dorsum. Propodeal spines each diverging from the midline at an angle of about 12 degrees; about equal in size to metapleural spines in lectotype, somewhat smaller in paralectotype. Petiole and postpetiole as in Figure 9; petiolar crest strongly arched in posterior view ; postpetiole very slightly wider than petiole in dorsal view.
Head and body almost completely lacking sculpturation, smooth and strongly shining, except for a little fine shagreening on the pronotal collar and slight obscure roughening on the sides of the propodeum. Pilosity generally as in P. foveolatus ; petiolar and post- petiolar hairs similarly distributed, but with an additional pair on the petiolar peduncle (as in the thoracicus female). First gastric tergite completely naked. Pubescence virtually absent.
2I have not seen the queenslandensis type, but specimens from the Mt. D’Aguilar series, listed below, have been directly compared with it for me by Dr. W. L. Brown, Jr.
[March
44
Psyche
j I'Omm
Figures 7-10: Pristomyrmex quadridentatus Andre. — lectotype worker: fig. 7, head and pronotal outline in frontal view ; fig. 8, mesosoma, petiole and postpetiole in lateral view. Female (self-colored form) : fig. 9, head, frontal view; fig. 10, mesosoma, petiole and postpetiole, lateral view.
Color almost even rich light-golden sienna-brown, mesosoma a shade darker, masticatory border of mandible, and segments 2-8 of antennal funiculus darkly infuscated, less dark infuscation on scapes and the various carinae and spines of head and mesosoma. Eyes dark brown, almost black.
Additional material examined , distribution. Additional distri- butional records fall into two geographical clusters (Fig. 15) which coincide with the distribution of two morphological forms, apparently representing geographical variants of a single species. These variants are discussed in detail below and illustrated in plate 3 ; essentially there is a southern, self-colored golden-brown form (to which the lectotype and paralectotype belong), and a more northern bicolored form with dark blackish-brown mesosoma, petiole and postpetiole (equivalent to Forel's suppressed variety queenslandensis) . They are distributed as follows:
1965]
Taylor — Pristomyrmex
45
1. Self-colored form: N.E. NEW SOUTH WALES between about S. lat. 29.40 and 30.20 (not including Sydney — see above) (open circles in Fig. 15).
Specific records are: Point Lookout to Grafton , dealate female, April 1958, (P. F. Darlington). New England National Park , Platypus Creek (elevation not given, probably 2-3000 ft — see LeGay Brereton, 1963, fig. 1), in log, February 27, 1961 (Nicholls). The Dorrigo 3000 ft (W. Heron).
2. Bicolored form: Extreme S.E. QUEENSLAND, between
about S. lat. 27.20 and 28.20 (closed circles in Fig. 15).
Specific records are: Mt. D’ Aguilar Range , 2000 ft, nest in
rotting log, May 4, 1962 (RWT acc. 799). Cunningham's Gap , 2500 ft, nest in rotting log, May 1, 1962 (RWT). Stradhroke Island, dealate queen, December 3, 1912 (H. Hacker). Tamborine Mountain: no further data (A. M. Lea) ; Cedar Creek, ex rotting logs — dealate female, May 15, 1951, workers November 13, 1950 and May 12, 1951 (W. L. Brown) ; worker under log, February 27, 1956 (B. B. Lowery). Binna Burra: c. 2600 ft: nests in rotting logs, one with alate female, May 22-23, 1962 (RWT accs. 1145, 1182); several series of workers collected straying at night (see below under “ecology”) , May 21-23, 1962 (RWT). Lamington National Park, Coomera Gorge: leafmold berlesate, October 29, 1955 (T. E. Woodward) ; 2200 ft, nest in rotting log, May 23, 1962 (RWT acc. 1168).
All the above records are specifically cited as rain forest collections, except the dealate female from Stradbroke Island, which has no ecological data.
IV orker variation. Measurements and indices of workers from the above series are summarized in Table 1. The similarity of ranges and means in the two color variants is obvious and reflects their consider- able resemblance in all characters except color.
The coloration of the types (Plate 3, lower figure) is typical of the self-colored form. The bicolored form has the mesosoma, petiole and postpetiole very dark blackish-brown, contrasting sharply with the golden-brown head, gaster and appendage's (Plate 3, middle figure), and the antennal infuscation is much less intense. The pronotum and postpetiole may be medium to dark reddish-brown, usually in the larger specimens of a given series. Such individuals are probably incompletely colored ; since these areas are the last to become fully pigmented in callow specimens.
The mesosomal spines vary in thickness, length and inclination. The pronotal pair are typically longer and thicker than the propodeal ones,
46
Psyche
[March
|
SELF-COLORED FORM |
BICOLORED FORM |
|||||||
|
Series No. |
Types |
1 |
2 |
Total |
3 |
4 |
5 |
Total |
|
Number |
2 |
4 |
11 |
17 |
10 |
14 |
31 |
55 |
|
HL |
0.95, 0.99 |
1.03-1.07 (1M) |
1.00-1.06 (1403) |
0.95-1.07 (1.00) |
0.88-1.06 (0.97) |
0.92-1.11 (1.03) |
0.97-1.11 (i.oi) |
0.88-1.11 (1.00) |
|
HW |
0.95, 1.00 |
1.01-1.06 (1.04) |
0.97-1.0 6 (1.03) |
0.95-1.06 am |
0.88-1.03 (0.97) |
0.92-1.12 (1.03) |
0.89-1.14 (0.98) |
0.88-1.14 (0.99) |
|
Cl |
100. 100 |
98-100 (99) |
100-103 (101) |
98-103 (100) |
97-103 (100) |
98-101 (100) |
98-103 (101) |
97-103 (100) |
|
SL |
0.90, 0.93 |
1.01-1.06 (1.04) |
1.00-1.03 (1.01) |
0.90-1.0 6 (0.97) |
0.88-1.04 (0.97) |
0.87-1.09 (1.02) |
0.86-1.15 (1.01) |
0.86-1.15 (1.00) |
|
SI |
95, 93 |
100 |
97-103 (99) |
93-103 (97) |
97-103 (99) |
95-105 (99) |
95-103 (98) |
95-105 (99) |
|
PW |
0.46, 0.50 |
0.48-0.53 (0.51) |
0.48-0.52 (0.50) |
0.46-0.53 (0.49) |
0.43-0.50 (0.48) |
0.44-0.53 (0.50) |
0.45-0.56 (0.50) |
0.43-0.56 (0.49) |
|
WL |
0.95, 0.99 |
1.03-1.06 (1.04) |
1.01-1.06 (1.03) |
0.95-1.06 (1.00) |
0.92-1.03 (0.99) |
0.92-1.05 (1.00) |
0.92-1.15 (1.03) |
0.92-1.15 (1.01) |
|
WL X 100 HW |
100, 99 |
100 |
100-105 (101) |
99-105 (100) |
100-104 (101) |
100-101 (100) |
98-103 (101) |
98-104 (101) |
|
PW x 100 HW |
48, 50 |
48-50 (49) |
46-50 (48) |
46-50 (49) |
47-51 (48) |
47-50 (48) |
47-51 (49) |
47-51 (48) |
|
PW x ioo WL |
48, 50 |
47-50 (48) |
46-48 (48) |
46-50 (48) |
47-50 (48) |
47-50 (48) |
47-51 (48) |
47-51 (48) |
|
Pronotal Spine L. |
0.15, 0.15 |
0.17-0.19 |
0.15-0.20 |
0.15-0.20 |
0.13-0.18 |
0.13-0.17 |
0.13-0.17 |
0.13-0.18 |
|
Propodeal Spine L. |
0.23, 0.22 |
0.18-0.20 |
0.20-0.24 |
0.18-0.24 |
0.15-0.20 |
0.16-0.19 |
0.16-0.20 |
0.15-0.20 |
|
Petiole L. |
0.33, 0.33 |
0.34-0.37 |
0.33-0.37 |
0.33-0.37 |
0.32-0.37 |
0.30-0.37 |
0.30-0.38 |
0.30-0.38 |
|
Postpet. H. |
0.32, 0.33 |
0.34-0.35 |
0.33-0.37 |
0.32-0.37 |
0.32-0.35 |
0.31-0.36 |
0.28-0.35 |
0.28-0.36 |
Table 1. Pristomyrmex quadridentatus Andre. Comparison of standard measurements (in mm) and indices, yielded by specimens from all known series. Mean values italicized. The types are ostensibly from Sydney ( hut see text). Numbers in the “series” column refer to the following localities: 1. New England National Park; 2. The Dorrigo ; 3. Mt. D’Aguilar; 4. Tamborine Mt. ; 5. Binna Burra.
which are usually about sub-equal to the metapleurals. The bicolored form tends to have thicker pronotal and smaller propodeal spines than the self-colored form; but the overall variation is such that the avail- able quadridentatus material cannot be divided into two sets, co- extensive with the color-variants, on the basis of spinational characters.
1965]
Taylor — Pristomyrmex
47
A weak negative allometric relationship between spine development and body size is evidenced ; this is best marked in the bicolored form, probably because of its better representation in the study series.
Female (previously undescribed). Six female specimens are avail- able for study; one self-colored, five bicolored.
1. Self-colored form. A single dealate labeled Point Lookout — Grafton, N.S.W., has the following dimensions: HL 1.05 mm; HW 1.07 mm; Cl 102; SL 0.98 mm; SI 88; Scutum W 0.68 mm; WL 1.20 mm; Propodeal Spine L 0.22 mm; Petiole L 0.35 mm; Post- petiole Height 0.39 mm; Eye Diameter 0.22 mm; Ocular Index 21. General features as in Figures 9 and 10. Pronotum with a partly broken fine, arched transverse carina, its humeri each with a low obtuse tubercle. Parapsidal lines of mesoscutum very obscure, es- sentially vestigial. Bases of propodeal spines connected by a transverse carina. Metapleural spines slightly longer and more acute than propodeal ones.
Color and pilosity as in worker, the ocelli each with an adjacent small dark brown spot, the mesosomal suture lines and axillary sclerites infuscated dark brown.
2. Bicolored for?n. Two dealate specimens from Mt. D’Aguilar Range, one each from Stradbroke Island and Tamborine Mt., and an alate from Binna Burra, have the following dimensions: HL 0.95-1.22 mm; HW 0.99-1.30 mm; Cl 102-106; SL 0.94-1.18 mm; SI 91-95; Scutum W 0.62-0.85 mm; WL 1.10-1.42 mm; Petiole L 0.32-0.46 mm; Postpetiole Height 0.32-0.43 mm; Eye Diameter 0.19-0.26 mm; Ocular Index 19-20. The smallest specimen is from Mt. D’Aguilar Range, the largest is the Binna Burra alate (which has a forewing length of 1.9 mm, hindwing length of 1.3 mm). Distribution of indices in the series implies negative allometry of scape length and head length, relative to head width. Pronotal spine development variable, apparently negatively allometric, smallest specimen with a low obtuse tubercle, largest one with an almost imperceptible slightly raised convexity. Propodeal spines similarly variable, best developed in smallest individuals (about comparable to self-colored specimen), represented only by slightly raised propodeal corners in largest specimen. Characters otherwise as in self-colored form, with the scapes proportionately slightly longer.
Coloration as in worker, ocellar area darkly infuscated; pronotum and scutum reddish brown in the Binna Burra alate, which may be incompletely colored.
Ecology. All detailed records of quadridentatus show that nests are usually located in rotting logs, a fact confirmed in the field by
48
Psyche
[March
the author, and by the experienced Australian collector B. B. Lowery, S. J., (pers. comm.).
On May 21-23, 1962, I established that the bicolored form is probably an exclusively nocturnal forager. During 4 days of intensive collecting in rain forest near Binna Burra, workers were never encountered abroad from nests during the day. At night, however, 22 workers in seven samples were taken as strays, during about four hours total collecting. These specimens were all collected between two and six hours after sunset (the limits of the observation period), and all were in the open on the surfaces of widely separated rotting logs. Several further individuals were taken near the ground on tree trunks, and one was collected from foliage in a sweep-net. In several cases these individuals appeared to be following trails, but they were always well separated from each other, at distances of 14-20 cm. The ants moved slowly, their rate of progress being estimated at 10-15 cm per minute. The nature of the food was not determined.
Relationships. See discussion on p. 35.
4. Pristomyrmex wheeleri new species (Figs. 11-15, Plate 3)
Type workers. The following description is based on the holotype and numerous paratypes from all parts of the distributional range given below.
Dimensions (holotype cited first) : HL 1.14, 0.97-1.23 mm; HW 1.26, 1. 04- 1. 40 mm; Cl 111, 108- 1 16; SL 1.11, 0.92-1. 14 mm; SI 88, 81-90; PW 0.60, 0.51-0.67 mm; WL 1.09, 0.89-1.20 mm; Pronotal Spine L 0.17, 0.15-0.20 mm; Propodeal Spine L 0.35,
0.27-0.39 mm; Petiole L 0.41, 0.36-0.44 mm; Postpetiole Height 0.38, 0.31-0.43 mm. General features as in Figures 11 and 12, and Plate 3 (upper figure). Similar to P. quadridentatus, as described above, but differing in the following features:
1. Generally larger size (HW 1.04- 1.40 mm, opposed to 0.88-1.14 mm in quadridentatus) , with much broader head (Cl 108-116 against 97-103), and lower Scape Index (81-90 opposed to 93-105).
2. Inferior carina of antennal scrobe vestigial or absent.
3. Pronotal and propodeal spines longer and more slender, in side view approximately sub-equal in length ; metapleural spines acute, rarely more than half as long as propodeals.
4. Pilosity similar, the body hairs thicker and longer, petiolar peduncle with a single hair on each side, first gastric tergite naked.
5. Color distinctive; head deep dark umber-brown. Mandibles, clypeus and antennae rich golden-brown. Remainder of body golden- brown, much like the self-colored form of quadridentatus , but more
1965]
Taylor — Pristomyrmex
49
Figures 11-14: Pristomyrmex wheeleri n. sp. — Holotype worker: fig. 11, head and pronotal dorsum, frontal view; fig. 12, mesosoma, petiole and postpetiole, lateral view. — Paratype female: fig. 13, head, frontal view; fig. 14, mesosoma etc., lateral view.
yellowish. Posterior half of first gastric tergite shaded medium to dark reddish brown.
Type locality. Vicinity of Binna Burra, S.E. Queensland. The type colony was taken in soil under a stone in rain forest, elevation c. 2800 ft (May 24, 1962, R. W. Taylor acc. 1281).
Worker variation. Negative allometry of head length and scape length, relative to head width, is suggested by the distribution of cephalic and scape indices in the study series. Other characters, including the clypeal structure and the development of the mesosomal spines, show no apparently significant variation.
Paratype females. A female from the type-nest (cited first), and six others, one from each locality given below except National Park and Woodenbong, have the following dimensions: HL 1.11, 1 . 1 7- 1.25 mm; HW 1.23, 1.3 1- 1.42 mm; Cl 112, 112-114; SL 1.04,
50
Psyche
[March
Figure 15 (above) : Distribution of members of the Pristomyrmex quadri- dentatus group. Land above 200 meters elevation shaded.
EXPLANATION OF PLATE 3 (opposite)
Workers of the Pristomyrmex quadridentatus group, to illustrate coloration. Top: P. wheeleri n. sp., from Binna Burra, S. E. Queensland. Middle: P. quadridentatus Andre, bicolored form, from Binna Burra (sympatric with •wheeleri) . Bottom: P. quadridentatus Andre, self-colored form, from New England National Park (Point Lookout) (allopatric with eivheeleri). Author del
Psyche, 1965
Vol. 72, Plate 3
Taylor — Pristomyrmex
52
Psyche
[March
1.08-1.12 mm; SI 85, 79-82; Scutum W 0.67, 0.78-0.84 mm; WL 1. 1 9, 1.26-1.36 mm; Propodeal Spine L 0.33, 0.35-0.39 mm; Petiole L 0.41, 0.44-0.48 mm; Postpetiole Height 0.39, 0.41-0.45 mm; Eye Diameter 0.24, 0.26-0.27 mm; Ocular Index 20, 19-20. Negative allometry of scape length and head length, relative to head width, is indicated. General features as in Figures 13 and 14. Distinguished from P. quadridentatus by very different proportions and coloration. Structure of clypeal apron variable, usually as in Figure 13; the median projection narrowed with reduced teeth in one specimen, bidentate in another, and completely edentate in a third. Transverse pronotal carina distinct. Pronotal spines variable, ranging from low rounded tubercles to small acute teeth (maximally about 0.03 mm long) , with no apparent allometric tendency. Parapsidal lines feeble to moderately distinct. Propodeal spines almost invariable, their bases bridged by a low carina. Metapleural spines acute, distinctly shorter than propodeals.
Color and pilosity as in worker, mesosomal suture lines and alary sclerites infuscated dark brown.
Distribution and Material Examined. All records of P. wheeleri are from S. E. Queensland and N.E. New South Wales, between about S. Lat. 27.20 and 29.00 (Fig. 15), as follows: QUEENS- LAND: A It. D’ Aguilar Range, 2000 ft, colony, May 4, 1962 (RWT acc. 819). Cunninghams Gap: 2500 ft, six colonies,
January 22, 1961 (B. B. Lowery) ; 2000 ft, colony, May 1, 1962 (RWT acc. 721). T amborine Mountain : no further data (A. M. Lea); colony, January 15, 1957 (B. B. Lowery); Curtis Falls, workers, leafmold berlesate, May 8, 1953 (T. E. Woodward). Binna Burra: 2800 ft, colonies, May 21-25, 1962 (RWT accs. 1192, 1281 ( holotype nest series) , 1290, 1293) J stray, forest floor, 2 hours after sunset, May 23, 1962 (RWT acc. 1210) ; 2600-3000 ft (P. F. Darlington). National Park , December 1919 (H. Hacker). NEW SOUTH WALES: W oodenbong ; JJnumgar Forest , c. 2000 ft; Tooloom Range , c. 2000 ft, dealate female; all May 1958 (Darling- tons). Mt. Warning, 2600-3000 ft, five colonies, September 13, 1962 (B. B. Lowery).
Type deposition. Holotype and numerous paratypes (worker and female) in MCZ (Type No. 31 154), paratypes in CSIRO; Queens- land Museum, Brisbane; Australian Museum, Sydney; National Museum of Victoria, Melbourne, and other Australian collections, also British Museum (NH), London, and B. P. Bishop Museum, Honolulu.
Ecology. All the above records are from rain forest. Sixteen of the
1965]
Taylor — Pristomyrmex
53
1 8 colonies listed were found in soil under or between stones, one was in a rotting log (RWT acc. 819), and one in a rotting wood frag- ment (RWT acc. 721). Nesting behavior thus contrasts with that of the sympatric bicolored form of quadridentatus , which has been found nesting only in rotting logs. Lowery’s six Cunningham’s Gap colonies each contained 18 to 36 workers and one queen. At Binna Burra I searched extensively for foragers of wheeleri but encountered only one (acc. 1210), which was taken two hours after sunset on top of a small rock, later found to shelter a colony. Foraging behavior presumably is unlike that of P. quadridentatus (see p. 48).
Relationships. See below.
DISTRIBUTION AND VARIATION IN THE QUADRIDENTATUS GROUP
All known bicolored series of Pristomyrmex quadridentatus were collected in immediate geographical and temporal sympatric associ- ation with P. wheeleri , while the self-colored form has never been taken in such association. This correlation invites the hypothesis that character displacement of quadridentatus has occurred wherever the species is sympatric with wheeleri.
Character displacement is the situation, usually involving several related species, where interspecific differences are more distinct in areas of sympatry than in areas of allopatry. This phenomenon has long been recognized, but was first crystalized by Brown and Wilson ( 1956) , who reported its occurrence in birds, frogs, fishes, crabs and various insects. Subsequent reports involve many animal groups, including the South African cyprinid fishes, Barbus marequensis and B. natalensis ( Farquharson, 1962), the North American salamanders Plethodon cinereus (Green) and P. richmondi Netting and Mittle- man (Highton, 1962), the Central American iguanids. Basiliscus plumifrons Cope and B. vittatus Weigmann (Maturana, 1962), the East African gekkonids Holodactylus africanus Boettger and H. cornii Scortecci (Laurent, 1964), several pairs of Indian nuthatches of the genus Sitta (Ripley, 1959) and the New Zealand pelecypods Bassinia yatei (Gray) and B. parva Marwick, which apparently underwent displacement on becoming sympatric in the Lower Pleisto- cene (Fleming, 1959). Few of these cases have been studied in detail aimed at elucidating the selective significance of the displacement, though it seems clear that secondary reinforcement of premating isolating mechanisms, or special adaptation to greater niche specificity resulting from competitive displacement, are involved in some cases.
This hypothesis is invoked to explain the origin of the bicolored form of P. quadridentatus on limited distributional evidence, which needs supplementation. A selective mechanism is difficult to envisage since there is no evidence that color is involved in interspecific relations
54
Psyche
[March
between these ants. Ecological differences between the sympatric forms are reported above; at Binna Burra quadridentatus nests in logs and forages openly at night, while wheeleri nests under stones and apparently forages cryptically. Nothing is known of the behavior of self-colored quadridentatus and such information is required. Bicolored quadridentatus and ivheeieri are perhaps the most spectacu- larly colored members of their genus; their coincident distribution suggests that the causal factors producing their color patterns may be inter-related. The two “forms” of quadridentatus might prove ultimately to be good biological species, but even if this is the case, character displacement need not be rejected in explaining their color differences. These ants are worthy of much further study.
Literature cited
Brereton, J. LeG.
1963. Point Lookout survey New England National Park. Aust. J. Sci., 26(1): 10-14.
Brown, W. L., Jr.
1953. Characters and synonymies among the genera of ants, Part I. Breviora, 11 :1-13.
Brown, W. L., Jr., and E. O. Wilson
1956. Character displacement. Syst. Zool., 5 : 49-64.
Darlington, P. J., Jr.
1960. Australian carabid beetles IV. List of localities, 1956-1958. Psyche, 67(4): 111-126.
Emery, C.
1922. Formicidae, Subfamily Myrmicinae. Genera Insect., 174c: 207- 397.
Farquharson, F. L.
1962. The distribution of cyprinids in South Africa. Ann. Cape Prov. Museums, 2: 233-251.
Fleming, C. A.
1962. Paleontological evidence for speciation preceded by geographic isolation, in The Evolution of Living Organisms , G. W. Leeper ed. pp. 1-459, Melbourne University Press, (pp. 225-241). Highton, R.
1962. Revision of North American salamanders of the genus Plethodon. Bull. Florida. State Mus., 6 (3) :235-367.
Laurent, R. F.
1964. Deplacement des characteres dans le genre Holodactylus (Repti- lia, Gekkonidae). Senck. Biol., 45 (3/5): 417-423.
Maturana, H. R.
1962. A study of the species of the genus Basiliscus. Bull. Mus. Comp. Zool., Harv., 128 (1): 1-33.
Ripley, S. D.
1959. Character displacement in Indian nuthatches (Sitta). Postilla, Yale, 43: 1-11.
Wheeler, G. C., and Jeanette Wheeler
1954. The ant larvae of the myrmicine tribe Myrmecinini. Proc. Ent. Soc. Washington, 56(3): 126-138.
A NEW SPECIES OF MEGALOMYRMEX FROM THE CHILEAN ANDES (FORMICIDAE, HYMENOPTERA)
By George Ettershank*
Department of Entomology, Cornell University
The genus Megalomyrmex has generally been thought of as associated with the wet tropical rain-forests of the New World. Members of the goeldii species-group in particular forage openly on the rain-forest floor and tend Homoptera on low bushes. It was thus a considerable surprise to receive a few specimens of a goeldii group species collected only 25 miles from the crest of the Andes in Chile. This is a new species, and its discovery extends the genus into a new habitat.
The specimens under study were found by Mr. E. Ackerman at Pachon Peak (Cerro Pachon), near La Serena, Coquimbo Province, Chile, at an altitude of 8500 dz 100 feet. The collection site is a rather flat hill crest with sparse low shrub cover; the substrate is andesite boulders resting on andesite rubble and soil. The colonies, Mr. Ackerman notes, were populous and by no means rare. A few live workers were sent to Dr. Caryl P. Haskins of the Carnegie Institution of Washington, who placed the workers in a laboratory nest; a few eggs were laid, from which three larvae were reared. All the material was later sent to me; this consisted of four live workers and three preserved larvae. My thanks are offered to Dr. Haskins for allowing me to study this interesting species.
Measuring techniques and abbreviations quoted below follow the system of Brown ( 1 ) while certain other terminology follows Etter- shank (2).
Megalomyrmex bicolor n. sp.
Holotype worker (measurements in millimeters; those for two parat}q>e workers are enclosed in parentheses) : total length, TL 4.8 (4.6, 4.4), head length, HL 1.04 (0.98, 0.94), length closed man- dibles, ML 0.24 (0.22, 0.20), alitrunk length, WL 1.34 (1.32, 1.26), head width, HW 1.00 (0.92, 0.88). Cephalic index, Cl = 100 HW/HL = 96 (94, 94), mandibular index = 100 ML/HL
^Present address: Department of Zoology and Comparative Physiology,
Monash University, Clayton, Victoria, Australia. This article is a partial result of work done under U. S. National Science Foundation Grant GB-2175, principal investigator, W. L. Brown, Jr.
Manuscript received by the editor January 4-, 1964
55
Psyche, 1965
Vol. 72, Plate 4
O- 5mm.
Ettershank — Megalomyrmex
1965]
Ettershank — Megalomyrmex
57
= 2 3 (22, 21). A fourth paratype worker was dissected and not measured.
This species is smaller than the other goeldii group species. The head (fig. 1) in dorsal view is rather more square in shape. The eyes are large, about 15 ommatidia high and 8 wide, and their black color contrasts strongly with the golden-brown ground color of the head. Antennae 12-segmented with a weak 3-segmented club (fig. 2). Clypeus strongly convex with a slight median sulcus, giving a weakly bicarinate effect. Median seta present, but variable in position around the midpoint of the oral margin of the clypeus ; one pair of intercarinal setae present ; first paracarinal setae well developed, higher paracarinal setae weaker. A few strong lateral setae present. Lateral areas of clypeus and oral margin of median area concolorous with mandibles. Mandibles with dental formula 1 + 4, masticatory margin more darkly colored, brown black, rest of mandible red-brown. Palpal formula 4, 3.
Alitrunk (fig. 3) of normal worker form. Promesonotal suture very distinct on the pleurae to the height of the spiracle, curving forward and then posteriorly; indistinct laterally above the spiracle and on the dorsum. Metanotal groove distinct, impressed on the dorsum and on the pleurae. Faint grooves separate the anepisternum from the notum above and the katepisternum below. Mesothoracic dorsum narrowing rapidly above, quite distinctly compressed at the metathoracic groove. Propodeum with angle distinct but rounded; declivity distinctly impressed. Propodeal spiracle small, round. Inferior propodeal plate's large, semicircular. Metapleural glands large, very distinct, surmounted postero-ventrally by a rounded protuberance bearing the gland opening.
Petiole and postpetiole of normal Megalomyrmex form, with nodes high and rounded (see fig. 3). Subpetiolar process dentiform, distinct; anterior subpostpetiolar process distinct, posterior process slight ; nodes not particularly broad above. Gaster biconvex in profile, attached at its most anterior end, in life normally carried with its long axis inclined downward so that the posterior end just clears the substrate.
Smooth shining ants, bearing a few faint and several distinct rugae on the mesokatepisternum and metapleuron. Posterodorsal flange of the postpetiole finely granulose. Head, alitrunk and gaster bearing sparse, long, white setae. Head (except as noted earlier), alitrunk,
Explanation of Plate 4
Megalomyrmex bicolor, n. sp., Worker: Fig. 1, head, dorsal full-face
view. Fig. 2, antenna. Fig. 3, alitrunk and nodes, lateral view. Larva: Fig. 4, lateral profile. Fig. 5, setae (see discussion in text) ; a, dorsal and lateral abdominal; b, of head and prothorax; c, of ventral abdominal tufts.
58
Psyche
[March
petiole and postpetiole golden brown, with a narrow margining of darker brown ventrally on the alitrunk ; gaster brown-black, the sheen of the rest of the body being reduced noticeably by a fine shagreening over all the gaster.
Malpighian tubules 5 (one worker dissected), the tips of the tubules being attached to the rectum (cryptonephric) .
Larvae: Three larvae were reared from worker-laid eggs and were presumably genetically male. A profile drawing of the larva is shown in fig. 4, and is essentially similar to that of M. symmetochus shown by G. C. and J. Wheeler (3). The head and prothoracic dorsum bear simple, slightly curved setae of the type shown in fig. 5b. The remaining body segments dorsally and laterally bear deeply cleft setae of the type seen in fig. 5a, while ventrally on each abdominal segment are paired clusters of setae of mixed types: short simple; long simple; and long deeply bifid, as shown in fig. 5c. In their characterization of Megalomyrmex larvae, the Wheelers (loc. cit.) state that only simple setae are present — short ones on the head, long ones on the body — and that the ventral and lateral surfaces of the abdomen are nearly naked. This diagnosis is based on M. symmetochus , a member of the modestus species group. Further comparative study is required, but perhaps this setal character represents another fundamental difference between the two species groups.
Disposition of Types: Holotype worker, three paratype workers (one dissected) and three larvae have all been deposited in the collection of the Museum of Comparative Zoology, Cambridge, Mass., under type number 31 139.
References
1. Brown, W. L.
1953. Revisionary studies in the ant tribe Dacetini. Amer. Midi. Nat. 50: 1-137; cf. pp. 7-15.
2. Ettershank, G.
1965. A generic revision of the world Myrmicinae related to Solenopsis and Pheidolog eton (Hymenoptera, Formicidae). Bull. Mus. comp. Zool. Harv. (in press).
3. Wheeler, G. C. and J. Wheeler.
1955. The ant larvae of the myrmicine tribe Solenopsidini. Amer. Midi. Nat. 54: 119-141.
THE HABITS AND DISTRIBUTION OF CRYPTOCERUS ROHWERI WHEELER (HYMENOPTERA: FOR MI Cl DAE)
By
William S. Creighton, City College, New York1
AND
William L. Nutting, University of Arizona2
When W. M. Wheeler described Cryptocerus rohweri in 1916 he stated that the specimens which Rohwer sent him had been taken by Chrisman in a canyon of the Santa Catalina Mountains of Arizona and that they had been nesting in the dead limbs of a palo verde tree ( 1 ) . There was no reason for Wheeler to question that the host tree was Cercidium torreyanum , but this question has since arisen and it is advisable to consider it here. The host plant identifi- cation appears to have been made from the dead limb sent in with the ants. Chrisman’s field notes stated only that his specimens were nesting in “palo verde”, which would cover either of the two species of Cercidium in the Santa Catalina area. But in this area C. torrey- anum is scarce and it usually occurs at elevations below those where rohweri has been taken. The abundant and widespread species is C. microphyllum and all seven colonies of rohweri which we took in the Santa Catalinas were nesting in this tree. The probability is that
Table 1. Distributional Data for Cryptocerus rohvoeri Wheeler
|
STATION |
ELEVATION |
NESTS |
NEST SITE |
COLLECTOR |
|
STA. CATALINA MTS. Buehman Canyon |
_ |
1 |
palo verde |
M. Chrisman |
|
Brush Corrals |
3700' |
1 |
palo verde |
M. Chrisman |
|
Sabino Canyon |
3000' |
2 |
hackberry |
E. D. Algert |
|
Sabino Canyon |
— |
1 |
not given |
W. D. Edmonton |
|
Catalina Springs |
— |
1 |
not given |
Hubbard & Schwarz |
|
Catalina Foothills |
2800' |
1 |
palo verde |
Wm. L. Nutting |
|
Catalina Foothills |
2800' |
6 |
palo verde |
Wm. S. Creighton |
|
Saguaro Nat. Mon. |
3100' |
strays |
unknown |
F. G. Werner |
|
BABOQUIVARI MTS. Baboquivari Canyon |
3500' |
1 |
mesquite |
Wm. S. Creighton |
|
Baboquivari Canyon |
3500' |
1 |
mesquite |
C. H. Musgrove |
|
A JO MOUNTAINS Alamo Canyon |
|
stray |
unknown |
E. D. Ball |
|
ATASCOSA MTS. Pena Blanca Sprs. |
3700' |
1 |
live oak |
Wm. S. Creighton |
Tmeritus Professor, Department of Biology, department of Entomology, College of Agriculture. Manuscript received by the editor December 28, 1964.
59
6o
Psyche
[March
Chrisman’s specimens came from C. microphyllum and not from C. torreyanum. The list in Table i is an expanded version of the list of localities for rohweri which M. R. Smith presented in his 1947 study of Cryptocerus (2).
While this list is unbalanced by the preponderance of records from the Santa Catalina Mountains it permits several conclusions. The range of rohweri covers a number of mountainous areas in southern Arizona. In each of these areas rohweri prefers to nest in canyon bottoms or on foothills at comparatively low elevations (2800-3700'). Although as yet unconfirmed, it is certain that the range of rohweri extends into northern Sonora. Pena Blanca Springs is (or perhaps better “was”, for the area has been dammed and flooded) about five miles north of the border of Sonora and the mountains there run south into Mexico. It is clear that rohweri will accept at least four different trees as nest sites ( Cercidium microphyllum , Prosopis juliflora, Q'uercus emoryi and Celtis sp.). Despite the large number of records from palo verde there are indications that rohweri has no special preference for this tree. In 1954 Creighton and Gregg showed that Crypt, texanus prefers to live in live oaks (3). This view was based not only on a preponderance of records for nests in live oak limbs but also on the fact that texanus nests in live oaks wherever these occur within its range. If rohweri prefers to nest in palo verde limbs it might be expected to do so over its entire range. As far as we have been able to determine it does not do so. In January of 1963 the senior author made an extensive survey of palo> verde trees in the area between Benson and Ajo. This survey failed to produce a single nest of rohweri. Negative evidence of this sort is not conclusive but at least it may be said that the high incidence of rohweri nests in palo verde limbs in the Santa Catalina Mountains is not maintained in other parts of its range. It is possible that our present fragmentary view of the range of rohweri is an outcome of the fact that the tree which it prefers as a nest site has not yet been recognized.
Most of the observations in the remainder of this paper are based on three captive colonies of rohweri. One of these was observed by the junior author from October 1961 until August 1963. During this period the colony was studied at Tucson, less than five miles from its original nest site in the Santa Catalina Mountains. The senior author was less fortunate for the two captive colonies which he observed were carried far out of their range. From January to mid- April of 1963 they were studied at Riverside, California. Thereafter until June 1964 they were studied at Rockport, Ontario. It would appear, however, that these expatriate colonies behaved in a normal
1965]
Creighton and Nutting — Cryptocerus
61
fashion, for there was no noticeable difference between their behavior and that of the Tucson colony. This colony was housed in a plastic petri dish 90 mm. in diameter. A short polyethylene tube connected this petri dish to a second one which served as a feeding chamber. This arrangement permitted a close watch on the development of the brood. The other two colonies were kept in sealed aquaria which contained oak block observation nests. This provided the ants with passages similar to those which they normally use and at the same time gave them the opportunity to forage outside the nest.
Most of the habits of rohweri are like those of texanus but there are some significant differences in the behavior of the two species. At maturity the rohweri colony is notably smaller than that of texanus. There are seldom more than seventy-five workers present and in most of the colonies that we have seen the total has been less than fifty individuals. This is undoubtedly due to the fact that rohweri is seldom, if ever, pleometrotic. Each of the nine colonies of rohweri which we have examined had a single queen. The nests of rohweri are established in abandoned burrows of wood-boring beetles (often those of small buprestids) which are cleared of the detritus left in them by the beetle larvae. While most of the branches selected by rohiveri consist of hard, sound wood it will nest in rotten branches as well. A limb housing one of the Santa Catalina colonies was so badly decayed that the ants were extracted by crumbling the wood between the fingers. As shown elsewhere (4) texanus ordinarily rejects nest sites in rotten wood. The burrows chosen by rohweri are of a size that permits the major to occlude the terminal opening. This occlusion is like that of texanus; the opening is blocked by the head and pronotum of the major, who crouches to admit the minor. An interesting variation of this response was observed in the junior author’s colony. This colony originally occupied burrows in a large, dead palo verde branch. Just inside the entrance of one of these burrows was a circular flange of detritus. The circular opening in this flange was slightly more than 2 mm. in diameter. This opening was occluded by the cephalic disc of the major, who stood in the passage behind the flange. The workers of rohweri pack themselves tightly into the outer portion of the nest passage, as do those of texanus, but show one response under these conditions that texanus does not display. The minor worker of rohweri can reverse its position in the passage by a twisting somersault. This begins with a lowering of the head, whose forward edge is thrust under the anterior coxae. Thereafter the body is swung forward and downward and during this arc it is twisted sidewise. The end result is that when the
62
Psyche
[March
minor regains its feet it is facing in the opposite direction. No major of rohweri was ever seen to behave in this way.
The foraging activities of rohweri are of interest since there is evidence that it deliberately forages on the ground. This is probably true of texanus as well but it has not yet been conclusively proved in that species. On one occasion Dr. F. G. Werner took six foraging workers of rohweri from white cholla in the Saguaro National Monument. Unless the ants were living in the cactus, which seems completely unlikely, they must have reached it over the surface of the soil. It is not clear why the foragers had visited the cholla. It was not in bloom and efforts by both writers to interest the captive colonies in cholla were unsuccessful. Workers in the aquarium colonies spent much time crawling over leaves and twigs of various plants with which they were kept supplied. When a worker fell from a leaf its righting reaction was completely different from that of texanus. When a worker of rohweri lands on its back it shows no fixed right- ing reaction. The body is violently contorted and the legs are flailed about until one of them anchors on something that enables the ant to pull itself over. The stereotyped righting reaction of texanus has been described elsewhere (4).
The junior author’s colony was fed on diluted honey, which was supplied through a wick from a reservoir. They were also fed on the juices of phalaenid caterpillars. It was necessary to tear the cater- pillars open before the workers would feed on them. Entire insects, either alive or dead, were avoided, as was pollen taken from honey bees. The colonies of the senior author were fed on pollen from the start. It was found that rohweri will accept a wide variety of pollen if it is smeared on the surface of leaves, although they seldom take it from the anthers of flowers. Of the various sorts of pollen fed to the colonies that of Ouercus agrifolia was clearly the most relished. As will be shown, the colonies were also fed with aphid honey dew. As a rule they preferred this to pollen but on one occasion, while the ants were feeding on honey dew, catkins of Quercus agrifolia were placed in the aquaria. The response to these was immediate and spectacular. The foragers feeding on honey dew deserted it for the oak catkins and practically every worker turned out to take pollen from them. They gathered such quantities of pollen that they returned to the nest with masses of grains in their jaws. These masses were held against the heads of the larvae who nibbled away the pollen grains. The original experiments with honey dew were disappointing. The ants paid little attention to coccids on the leaves of Quercus chrysolepis or to the rims of liquid which surrounded them. They were definitely
1965]
Creighton and Nutting — Cryptocerus
63
interested in Aphis fabi> a species that is abundant on Mesembryan- themum at Riverside. The ants tried to milk the aphids but the latter, who were clearly afraid of the rohweri workers, failed to cooperate. They would usually run away from the ants, which resulted in some of them being1 killed when the ants tried to catch them. Much better results were secured with Aphis sphaericola. This aphis produces such large quantities of honey dew that it will drip from the leaves on which the aphids are feeding. When leaves of Viburnum suspensum bearing Aphis sphaericola and coated with its honey dew were placed in the aquaria the foraging ants gorged themselves until their intersegmental gastric membranes were visible. This they did by lapping the honey dew from the surface of the leaves. The ants paid little attention to the aphids and made no attempt to milk them.
Since the colonies were well-supplied with food a number of eggs were soon present. When the major or female handles an egg it is positioned so that its long axis runs through the notch in the anterior edge of the cephalic disc. The larger eggs fit the rim of this notch closely when so positioned. Most of the eggs hatched in about 27 days. The voiding of the larval meconium was observed three times and took between 60 and 75 minutes. The movement of the black meconium, both within the larva and during its emergence is so slow that it is difficult to follow. Infrequent contractions of the posterior half of the larva probably reflect peristaltic movements of the gut within. Once outside the larva the meconium invariably attracted workers, both majors and minors, although they were never observed to assist the larva in any way. It was only after the meconium was presented that the larva was groomed. On one occasion the meconium was eaten by the attendant worker. On the other two it was carried to the dump in the feeding chamber. The period between the passing of the meconium and pupation was from six to ten days. The pupal moult was never observed. New pupae are ivory white and the first suggestions of pigmentation appear in the compound eyes. The color of the pupa darkens from yellow ochre to brown during the second and third weeks. There is a four- or five- day callow period after the adult emerges. In the senior author’s colonies worker brood developed from egg to adult in about three months (egg to larva ±27 days; larva to pupa ± 33 days; pupa to adult ± 23 days). The nests were kept at room temperature and the range, for the most part, lay between 6o°F and 70°F. Since the temperature range to which a free colony is subjected is far wider, these figures are useful only as an indication of the relative length of the several stages. Moreover, the
64
Psyche
[March
development of the larva may be drastically modified. In the junior author’s colony eight males emerged during the period from July 3 to July 24, 1963. These males developed from eggs which had been laid in June of 1962. These eggs developed into larvae in about a month and the larvae grew for about two months. But from Septem- ber 1962 to June 1963 the larvae showed no further development. In short, these eight males over-wintered as larvae. They emerged in July and it is interesting to note that the marriage flight of rohweri occurs in that month. The senior author took a female of rohweri which had just completed her marriage flight in Baboquivari Canyon on July 28, 1951.
Literature Cited
1. Wheeler, W. M., Proc. New Eng. Zool Club, 6, pp. 29-35 (1916)
2. Smith, M. R., Proc. Ent. Soc. Wash. 49, No. 1, pp. 29-40 (1947)
3. Creighton, W. S. & Gregg, R. E., Psyche, 61, No. 2, pp. 41-57 (1954)
4. Creighton, W. S., Psyche, 70, No. 3, pp. 133-143 (1963)
CONTRIBUTIONS TO A RECLASSIFICATION OF THE FORMICIDAE. IV. TRIBE TYPHLOMYRMECINI (HYMENOPTERA)
By William L. Brown, Jr.
Department of Entomology, Cornell University
The Typhlomyrmecini (spelling here emended) are a tribe of Ponerinae here considered to contain the single small genus Typhlo- myrmex. In this sense the tribe dates only from Brown, 1953. The name Typhlomyrmicini (sic) } however, goes back to Emery, 1 9 1 1 , who first proposed it as a subtribe of tribe Ectatommini to contain the three genera Prionopelta , Typhlomyrmex , and Rhopalopone. Brown (1950) showed that Prionopelta belongs to tribe Amblyoponini, while Rhopalopone is a synonym of Gnamptogenys in tribe Ectatommini (Brown, 1958). After these subtractions, the genus Typhlomyrmex could not be placed comfortably in any existing tribe, and its present taxonomic position is an expression of this fact.
At first sight, Typhlomyrmex workers look like rather ordinary small cryptobiotic members of tribe Ponerini, although the frontal lobes are not as prominently developed as in Ponerini, and the petiole is never quite “right” in form. The males and larvae clearly conform to Emery’s “Section Proponerinae,” including Amblyoponini, Ecta- tommini, and Platythyreini in the modern sense; (the cerapachyines all probably belong here as well), so that the resemblance of the workers to those of certain Ponerini (in Emery’s “Section Eupone- rinae”) is either convergent or else marks a side lineage from near the base of the stock that led to the Ponerini.
Among “proponerines”, Typhlo?nyrmex shows some similarities to Amblyoponini and to Ectatommini, but it can be distinguished from both by the wing venation of the sexes and the larval mandibles. The main similarity between Typhlomyrmecini and Amblyoponini, other than in “basic ponerine” traits, lies with the shape of the petiolar node of one Typhlomyrmex species, T. rogenhoferi. This node, because of its elongate form without a distinct posterior face, resembles that of an Amblyopone very closely in side view. In dorsal view, however, T. rogenhoferi proves to have a much thinner (bilaterally compressed) petiolar peduncle, and this makes it seem possible that its amblyoponine features could have been convergently acquired. Whether or not this is the correct interpretation, it is true that, aside from basic “pro- ponerine” characters, the Typhlomyrmex adult has little in common
65
66
Psyche
[March
with the Amblyoponini. It shares more characters with the small “degenerate” members of Gnamptogenys , but here again, though less certainly, I feel that the similarities may be convergent ones. The two main characters contributing to this opinion are the forewing venation of the larger Typhlomyrmex species (Mfi arising basad of cu-a) and the shape of the mandible in the larva (inflated basal part, suddenly narrowed to an acute apical blade). Even these characters do not weigh decisively against a possible origin of Typhlomyrmex from ectatommine ancestors, and it must be admitted that the con- vergence hypothesis is to some extent based on subjective impressions that remain to be tested.
Tribe Typhlomyrmecini Genus Typhlomyrmex
Typhlomyrmex Mayr, 1862, Verh. zool.-bot. Ges. Wien, 12: 736. Type species: Typhlomyrmex rogenhoferi Mayr, 1862, monobasic.
Typhlomyrmex Emery, 1911, Gen. Insect., 1 18: 33-34, characterization and catalog of species.
Nec Typhlomyrmex Gistel : J. Betrem and C. Jacot-Guillarmod have called my attention to a generic name Typhlomyrmex obscurely published by J. Gistel in his Mysterien der europaischen Insectennvelt (1856, cf. p. 447) for a “Myrmica typhlops L.” This species name is a nomen nudum of Lund (not Linnaeus!), published in 1831 in Ann. Sci. Nat., 23: 128. Because it was based on an unavailable species name, and is itself without description, indication or figure, Typhlomyrmex Gistel is considered to be a stillborn name (nomen nudum), and I am well satisfied to let it subside into permanent nomenclatorial limbo.
Worker: Monophenic (“monomorphic”) or feebly polyphenic;
size small (full length under 2 mm to slightly over 5 mm) ; pigment- poor, yellowish to ferruginous in color.
Head parallel-sided, or sides slightly converging anteriad, occipital margin straight to slightly concave. Eyes reduced to minute vestiges with or without pigment. Clypeus with a broad, convex median part and narrow, concave side pieces, the anterior median border often with a narrow translucent margin, in a minority of cases produced as a variously-shaped small median process. Frontal carinae forming small frontal lobes that lie close together and roof a small basal part of the antennal scape insertion (but not the basal collar of the scape) ; lobes not or only weakly pinched in behind, and not expanded as in typical members of tribe Ponerini. Between the lobes lies a narrow, often indistinct frontal fossa, and a shallow median furrow may run back from the fossa to the vertex, or even to the occipital margin ; it is not usually as distinct as in most Ponerini. Antennae short, with thick scapes that either fail to reach the occipital margin, or else surpass it just barely, When held straight back. Funiculus of 11
1965] Brown — Typhlomyrmecini 67
segments, with the last 3 or 4 enlarged and forming a more or less distinct club.
Mandibles triangular in basic plan, their dorsal surfaces convex in both directions; basal border distinct from masticatory border, al- though they may meet either in an angle or a broad curve. Masticatory border with minute crenulation, denticulation and/or small, uneven teeth; apical tooth large to very large, and tending to cross with its opposite number when the mandibles are fully closed. Shape and dentition varying with the species.
Under-mouthparts relatively bulky; labrum bilobed, with a broad median excision. Palpi segmented maxillary 1, labial 2.
Alitrunk robust, with rounded humeri and propodeum, slightly constricted at posterior mesonotum ; promesonotal suture distinct and apparently movable; metanotal groove distinct, only feebly impressed. A distinct line curving posteroventrad from the propodeal spiracle represents the upper edge of the metapleural gland atrium showing through the integument. Legs short and thick; tarsal claws small, those on the anterior legs toothed, the others simple; tibial calcariae of middle and hind legs 1 or 2, indistinctly pectinate or simple.
Petiole briefly pedunculate; node distinct, variable in form (Figs. 1, 3) ; subpetiolar process well-developed, acute or rounded at apex. Gaster porrect or slightly downcurved, of the usual ponerine type, with slight but distinct constriction after postpetiole; sternum fused to tergum in abdominal segment III (postpetiole) and IV, but not fused in V ( T. rogenhoferi worker, female) . Sting well-developed and acute, usually exserted in dried specimens.
Sculpture generally fine, varying in development with the species and, within species, allometrically. Cranium longitudinally striate or striolate above, especially mesad, mostly shading off to reticulate on the sides. Mandibles and central part of clypeus usually smooth and shining. Alitrunk, petiole and gaster smooth, with spaced punctation, or partly reticulo-striate or otherwise roughened, the sculpture always becoming weaker caudad. Pilosity fine, rather short, uneven, fairly abundant and widely distributed. A pair of long fine sensory hairs rises steeply from the clypeus.
Alate female, or gyne: Slightly (T. pusillus) to considerably
(T. rogenhoferi) larger than associated workers, and often darker in color, at least around the ocelli ; darkest in T. rogenhoferi , which is brown. Sculpture sometimes better developed than in workers, and petiolar node distinctly shorter and more transverse. Compound eyes large and hairy; ocelli developed.
Alitrunk somewhat box-like, with a rather flat dorsal surface;
68
Psyche
[March
notauli obsolete; parapsidal furrows present but inconspicuous. Vena- tion nearly “complete/’ with the median abscissae of Rs (Rsf2#3) missing, so that the cubital cell is undivided (single). The first abscissa of M forks off from Cu basad of crossvein cu-a, as in the army ants. (This pattern of venation does not hold for T. pusillus, in which the veins are reduced and their relationships modified.) The hind wing entirely lacks an anal lobe, but has the large discal cell, usually with 2 or 3 stubs of apical abscissae corresponding to Rs, M and Cu; another small cell may occur at the base of the discal cell behind, or may be incorporated into the discal cell. The hamuli number 3, and usually arise from a small darkened sclerotic patch a little beyond the midlength of the costal margin.
In other characters, gynes resemble workers.
Male: (Based on T. rogenhoferi and T. clavicornis ) smaller and more slender than the corresponding gyne, but the difference is slight in the smaller species; dark brown to black in color; habitus typical of proponerine males. Eyes large and hairy, occupying nearly half of the sides of the head. Ocelli distinct. Scapes straight, of moderate length, usually equal to about the basal 3 or 4 flagellar segments; flagellum 12-segmented, the segments all longer than broad and increasing very slightly in thickness toward the apex. Mandibles well- developed, opposable or crossing at closure, dentition a variably reduced copy of that of the corresponding worker. Palpi segmented 1, 2 (rogenhoferi) or 1, 1 (clavicornis). Clypeus with a large, swollen mid section and small sunken side pieces.
Alitrunk with notauli developed only as the anterior arms of the “Y” and obsolete medially; parapsidal furrows present but inconspicu- ous. Wings as in gyne (see above). Legs slender, all three pairs with tarsal claws toothed.
Petiole subclavate, i.e., with the peduncle rising gradually caudad toward nodal summit, which is rounded; ventral tooth or process present. Gaster with a slight constriction behind postpetiole; the latter segment has tergum and sternum firmly fused, but the next segment, abdominal IV, appears to have them connected only by thin cuticle or membrane. Genitalia only partly retractile, with parameres broadly rounded at apex; volsellae varying with the species (Figs. 7, 8) ; aedeagal valves ordinary, serrate. Hypopygium with a long, more or less digitiform, upcurved, hairy, median process (ventral view, Fig. 9)-
Head longitudinally striate, rest of body predominantly smooth and shining, with fine scattered punctures. Pilosity fine, mostly erect and rather short, abundant and generally distributed.
1965]
Broiun — Typhlomyr/necini
69
Larva: (After G. C. and J. Wheeler, 1952 (rogenhoferi —
robustus) and 1964 (pusillus). Thorax moderately stout and bent ventrally; slightly constricted at first abdominal somite; remainder of abdomen stout and ovoidal. Body densely covered with moderate- sized branching (mostly trifid) hairs; head with a few bifid hairs. Mandible distinctive, composed of a strongly inflated basal 2/5 and a very narrow, acute apical 3/5, the latter with 2 small median teeth in addition to the apical. This mandible is somewhat like those of amblyoponine larvae, except that the basal portion is relatively much wider than in Amblyopone.
Distribution and biology: So far as known, Typhlomyrmex is
restricted to the warmer parts of the Americas, from southern Mexico to northern Argentina. Within this region, T. rogenhoferi is the most widespread and by far the most often-collected species, being an in- habitant of rotten logs in forest. This species is common in the Amazon Basin, where I have seen nests of several hundred workers moving in file through the rot zone just beneath the bark of a log. I have examined several such aggregations in the field, but I was not able to find definite indications of the prey of these undoubtedly predaceous ants. In some sites in the Amazon Basin, where rogen- hoferi is moderately common, I found the species in very close prox- imity to termite colonies, but I never saw it actually taking or feeding upon a termite.
T. pusillus appears to be a soil dweller in cultivated and pampas areas as well as in forest (Kempf, 1961). It also seems to exist at higher elevations (e. g., in a coffee plantation at Venecia, near Me- dellin, Colombia). Probably it is strongly subterranean in foraging and nesting habits. The small series taken by P. F. Darlington at the mouth of the Amazon came from a rotten root in rain forest soil.
The remainder of the species are rare, and nothing is known of their biology. Probably their habits are strongly cryptic; the large pro- portion known from alate males and females indicates that most samples are taken during or after nuptial flight. T. major may be restricted to the south of Brazil and neighboring countries, while T. clavicornis is widespread in South America. T. prolatus is known only from the unique type, a female from Costa Rica.
Synonymic synopsis of Typhlomyrmex species Typhlomyrmex clavicornis (Figures 4, 5, 7) Typhlomyrmex clavicornis Emery, 1905, Bull. Soc. ent. Ital., 37: 112, nota, alate female. Type locality: Mapiri, Bolivia. Holotype examined in Coll. Emery, Museo Civico di Storia Naturale, Genoa.
Typhlomyrmex clavicornis var. divergens Forel, 1906, Ann. Soc. ent. Belg.,
0. 5 mm.
Psyche, 1965
Vol. 72, Plate 5
Brown — Typhlomyrmex
0.5 m nru
1965]
Brown — Typhlomyrmecini
71
50: 248, female, male (in cop.). Type locality: San Bernardino, Para- guay. Type examined in Coll. Forel, Museum d'Histoire Naturelle, Geneva. New synonymy.
Typhlomyrmex richardsi Donisthorpe, 1939, Ent. mon. Mag., 75: 161, male. TL: Mazaruni Clearing, British Guiana. Type series examined in
British Museum (Natural History). New synonymy.
This species is distinguished in all castes by means of the wide head; long, falcate apical mandibular tooth (Figs. 4, 5) ; and oblique basal borders of the mandibles, which fail to meet the clypeus when closed. The worker-female antennal club is prominent, as the name suggests, and the petiolar node is short. A single worker (taken at Bartica, British Guiana by H. O. Lang, together with winged females) has a head length, without mandibles, of 0.67 and a head width of 0.65 mm., which is within the size range of the smaller T. rogenhoferi workers.
The type series of T. richardsi consists of numerous males, ac- companied on one card by a female specimen ( the latter not mentioned by Donisthorpe). The differences cited among the synonymous species by Forel and Donisthorpe mainly concern mandibular form and the proportions of the antennal segments. On examining all the types and comparing them with digms from British Guiana, I was impressed by the similarity of the mandibles between members of the same caste from different series. The basal segments of the funiculus show noticeable variation in length among males, even in those on one card, and I do not think they make a good diagnostic character.
T. clavicornis is known from the above-mentioned widely separated localities in South America, ranging from Bolivia and Paraguay north to British Guiana. A female with forewings missing, probably fully alate when captured, comes from the Floresta di Tijuca, near Rio de Janeiro, February i960, C. A. Campos Seabra leg. The T. richardsi types were a part of a large series (apparently nearly all males) taken from a nest of the social vespid Polybia bistriata.
Typhlomyrmex major , new status Typhlomyrmex pusillus st. major Santschi, 1923, Rev. Suisse Zool., 30: 246, worker. Type locality: Blumenau, Santa Catarina, Brazil. Location of type unknown (not in Santschi Collection).
Explanation of Plate 5
Figures 1-6, Typhlomyrmex spp. Fig. 1, T. pusillus worker from El Rey, Salta, Argentina, lateral view of body. Fig. 2, Same, head in full-face view. Fig. 3, T. rogenhoferi, large worker, lateral view of petiole. Fig. 4, T. clavicornis, mandible of gyne from British Guiana, hairs omitted. Fig. 5, T. clavicornis from British Guiana, full-face view of male head. Fig. 6, T. prolatus sp. nov., fertiale holotype, anterodorsal view of right mandible. Drawings by Nancy Buffler, D. Alsop and the author.
72
Psyche
[March
I refer to this species a worker from Agudos, Sao Paulo State, Brazil (W. W. Kempf leg.) with head length (HL) 0.71 mm, head width (HW) 0.63 mm. In habitus, this worker is like a small specimen of T. rogenhoferi , but the short, pusillus- like petiolar node separates it at once. It is distinguished from pusillus by its larger size and longer antennae, the scapes of which reach or surpass the occipital border when they are held straight back. The cephalic striation is also coarser and more distinct than in pusillus.
A female from Petropolis, Guanabara, Brazil (T. Borgmeier leg.), apparently belonging to this species, has HL 0.76 and HW 0.65. As has already been indicated, the female described by Santschi as T. foreli may belong to T. major.
Typhlomyrmex prolatus species nov. (Figure 6)
Diagnosis (gyne) : A medium-sized Typhlojnyrmex with unusually elongate, narrowly subtriangular mandibles; basal border short and curving broadly into long, indistinctly denticulate masticatory border (Fig. 6) ; apical tooth very long and acute. Petiolar node as seen from above broader than long, with feebly concave (almost straight) anterior border and strongly concave posterior border. Postpetiole with a distinct, sharp median longitudinal carina on the anterior third of its dorsal surface.
Holotype gyne, further description: Total outstretched length
(TL) 3.8 mm, head length (HL) 0.72, head width without eyes (HW) 0.63, in full-face view, closed mandibles extend beyond median clypeal margin (ML) 0.31, straight-line length of right mandible from external point of insertion to apex 0.49, length of alitrunk (WL) 1.13, scape length 0.52, greatest diameter of eye 0.18, petiolar node length in dorsal view 0.20, width 0.33 mm. Cephalic index (HW/HL X 100) 88.
Head with parallel, feebly convex sides; occipital border with a shallow v-shaped concavity. Eyes feebly convex. Clypeus with the raised median area very smooth, with a broad strip hairless. Mandibles smooth, with small elongate punctures, becoming striatopunctate and opaque laterally toward insertions. Antennal scapes just barely reach- ing occipital margin in full-face view when held straight back from insertions ; antennal club rather distinctly 3-segmented ; flagellar segments 2-8 broader than long. Median furrow distinct, reaching anterior ocellus. Ocelli small but distinct.
Remainder of body much as in other species. Middle and hind tibiae each with a single simple calcar. Forewing as in T. rogenhoferi , but M forking from Cu even farther back toward wing base. In hind wing, there is only a single large cell.
1965]
Brown — Typhlomyrmecini
73
The median carina on the postpetiole is a curious feature; it is symmetrical and does not look like a pathological condition. It is followed by a flat central area, which may even be slightly impressed. It remains to be seen whether the corresponding worker also carries it.
On the head, the longitudinal striation is indistinct except on the cheeks, and dense opaque reticulo-punctulation prevails. Dorsal surface of alitrunk densely punctulate, opaque to subopaque; pro- podeum mostly smooth and shining; sides of alitrunk weakly shining, pronotal part densely punctulate, remainder finely longitudinally striate, with scattered punctures. Petiole smooth and shining, with sparse punctulation on sides. Gaster smooth and shining, but with dense punctulation. Scapes densely punctulate, subopaque, as are also most of legs; mesal surfaces of femora smooth and shining.
Pubescence short, reelinate or appressed, fairly abundant over most dorsal body surfaces, gaster and appendages; longer fine hairs on clypeus, mandibles, and sparse on gaster above and below (abundant at gastric apex). Short oblique hairs extend beyond pubescence on scapes, funiculi and legs.
Color light ferruginous, legs lighter and more yellowish; head infuscated around ocelli.
Holotype (and only known specimen) from the vicinity of San Jose, Costa Rica, in 1940 (H. Schmidt leg.). Deposited in the collection of W. W. Kempf, Sao Paulo, Brazil.
Typhlomyrmex pusillus (Figures 1, 2)
Typhlomyrmex pusillus Emery, 1894, Bull. Soc. ent. Ital., 26: 141, pi. 1, fig. 2, worker. Type locality: Bolivia.
Typhlomyrmex schmidti Menozzi, 1927, Ent. Mitt., 16: 268, female, male. Type locality: vie. San Jose, Costa Rica. Syntypes in Istituto di Entomo- logia della Universita, Bologna, Italy; 2 alate females examined. New synonymy.
This is the smallest species of the genus. It will probably turn out to be much more common and widespread than it seems at present ; its habitat in the soil and its very small size have made it scarce in collections.
Samples that I have referred to this species show so much variation that they may actually represent more than one species. Specimens from the south (northern Argentina, Santa Catarina) and the Colombian Andes (Venecia, near Medellin) average larger and more robust (HW 0.34-0.50 mm) than those from Amazon drainage and the Guianas. Among the smaller forms, most samples (Surinam: Dirkshoop and Maripaheuvel ; Brazil: near Belem do Para; Peru: Finca Santa Beatriz, Chanchamayo) have the petiolar node and
74
Psyche
[March
postpetiolar dorsum very finely and densely sculptured, usually strio- late with interspersed shallow punctulation, and opaque or nearly so, but two samples from Tambahredjo in Surinam have the petiole and postpetiole smooth and shining, with only the usual abundant fine punctures. Of special interest is the occurrence, in two of the above samples (Peru: Finca Santa Beatriz, Chanchamayo, io°57’S,
75°i2’N, iooo m, C. A. Portocarrero leg., No. 88-SB29, 11 July 1964. Brazil: Utinga tract, near Belem, Para, P. F. Darlington leg., No. 335, Aug. 24, 1962) of a small, square to sharply trapezoi- dal lamellar lobe or process on the median anterior clypeal margin. This process is variable in shape and size, and is transparent and thus very difficult to see unless the mandibles are open at least partly. It appears to represent a modification of the lamellar free clypeal margin seen in other samples. This margin is usually evenly convex, but some Surinam specimens show a tendency for its most central part to form a narrow, shallow truncate lobe. Such a character would normally separate good species among ponerine ants, but the overall variation in the small Typhlomyrmex calls for caution and more than the present meager material before we draw new species boundaries.
The wings of the female of this species are atypical for Typhlomyr- mex, in that M leaves Cu distad of cu-a, and the crossvein r-m is missing, so that the cubital cell is open at its apex.
The Santa Beatriz collection was made from a single chamber about 4 mm in diameter, located in the ground; Portocarrero found there 21 workers and one queen, plus 8 pupae, 1 larva, and 10 eggs. The Utinga collection came from a red-rotten root in rain forest.
Typhlomyrmex rogenhoferi (Figures 3, 8-1 1)
Typhlomyrmex rogenhoferi Mayr, 1862, Verh. Zool.-bot. Ges. Wien, 12: 737, worker. Type locality: “Amazonas.”
Typhlomyrmex Rogenhoferi r. robustus Emery, 1890, Bull. Soc. ent. Ital., 22: 40, worker. Type locality: Alajuela, Costa Rica. Syn. Brown, 1957. Typhlomyrmex robustus subsp. manco Wheeler, 1925, Ark. f. Zool., 17A (8): 2, worker. Type locality: Pablobamba, Peru. Syn. Brown, 1957. Centromyrmex sculpturatus Santschi, 1931, Revista Ent., Rio de Jan., 1: 266, dealate female. Type locality: Panama Canal Zone. Holotype examined in Naturhistorisches Museum, Basel. New synonymy.
Easily recognized by the relatively large size, triangular mandible and shape of the petiolar node in worker and female. The male can be recognized by size, the distinct angle between basal and masticatory borders of the mandible, and by the distinctive volsella of the genitalia.
Interesting Brazilian records for the species have been furnished by W. W. Kempf from his collection : Goias State, Goiaffiia,
Campinas ( Schwarzmaier leg.). Espirito Santo State, Santa Teresa
1965]
Brown — Typhlo?nyrinecini
75
Figures 7-11, T yphlomyrmex spp. Fig. 7, Right half of genital capsule of T. clavicornis male from British Guiana, viewed from inside, ventral side to right, semidiagrammatic. Fig. 8, same, T. rogenhoferi, aedeagal valve omitted. Fig. 9, hypopygial process of T. rogenhoferi, ventral view. Fig. 10, maxillary palpus (left) and labial palpus (right) of T. rogenhoferi. Fig. 11, right wing of T. rogenhoferi Figs. 8-11 are drawn from a male from Perene, Peru. Drawings by D. Alsop and the author.
76
Psyche
[March
(O. Conde leg.)- Minas Gerais State, Serra Caraga (K. Lenko leg.). Sao Paulo State, Fazenda Itaquere, Nova Europa (K. Lenko leg.). The species ranges from Bolivia to Veracruz State in Mexico.
Typhlomyrmex foreli
Typhlomyrmex foreli Santschi, 1924, Ann. Soc. ent. Belg., 64: 6, female.
Type locality: Rio Negro, Parana, Brazil. Location of type unknown
(not in Santschi Collection).
This species was described from a single gyne. I suspect it to be the gyne of T. major , but some items in Santschi’s description will leave doubts until the type can be re-examined. Probably the “court sillon median” on the clypeus can be dismissed as an illusion caused by the shiny surface here; I have noticed this in other species. Santschi says that the external margins of the mandibles are slightly concave, a description that will fit no specimen of any species of Typhlomyrmex in full-face view, but will fit all of them if viewed obliquely from above and slightly to the side.
The “bord terminal long, finement denticule avee une dent apicale bien developee” would fit the new species prolatus (see above), but would also fit T. major reasonably well.
Santschi’s statement, “The first article of the funiculus is not quite as long as the three following ones taken together” fits prolatus fairly well, while in a gyne I take to be major, the first funicular segment is shorter than this.
In spite of these difficulties, the like of which often crop up in connection with Santschi’s descriptions of ants, I think it probable that T. foreli and T. major are conspecific. The distributional evidence weighs for this decision, and Santschi mentions no posterior concavity of petiole and no median postpetiolar carina such as prolatus carries.
Summary of changes proposed in species-level taxonomy of Typhlomyrmex
clavicornis Emery
— clavicornis var. divergens Forel, n. syn.
— richardsi Donisthorpe, n. syn.
f foreli Santschi (possibly a synonym of T. major ) major Santschi, raised to species level prolatus sp. nov. pusillus Emery
= schmidti Menozzi, n. syn. rogenhoferi Mayr
— rohustus Emery
1965]
Brown — Typhlomyrmecini
77
= robustus manco Wheeler = Centromyrmex sculpturatus Santschi, n. syn.
Note: Brown (1953) cited Prionopelta marthae Forel as a syno- nym of Typhlomyrmex rogenhoferi. The synonymy was based on a specimen labeled as a cotype of P. marthae , found in the Wheeler Collection, and which is clearly a specimen of T. rogenhoferi. Studies in European museums in 1963 and 1964 show that this specimen is mislabeled, because P. marthae syntypes (“cotypes”) in the Forel Collection and elsewhere are true Prionopelta close to, and possibly conspecific with, the species currently called Prionopelta antillana. P. marthae must, therefore, be deleted from the synonymy of T. rogenhoferi and returned to genus Prionopelta.
Key to Typhlomyrmex species, workers
1. Petiolar node longer than high, without a differentiated posterior
face (Fig. 3) rogenhoferi
Petiolar node as high as or higher than long, with a differentiated posterior face (Fig. 1) 2
2. Head width <0.55 mm pusillus
Head width >0.55 mm 3
3. Mandibles triangular, with basal border fitting tightly against clypeus at full closure; apical tooth stout, not notably elongate
major
Mandibles more elongate, basal borders oblique and not closing up against clypeus; apical tooth notably elongate and very acute
( Fig. 4) clavicornis
Note: The worker of T. prolatus, at present unknown, probably would key to couplet 3, where it would undoubtedly be dis- tinguished from both alternatives by having mandibles like those of its gyne (Fig. 6).
Key to Typhlomyrmex species, gynes
1. Petiolar node without a differentiated posterior face; size large,
head width (without eyes) normally >0.85 mm rogenhoferi
Petiolar node with a differentiated posterior face; size smaller .. 2
2. Head width (without eyes) <0.58 mm; r-m crossvein missing
from forewing pusillus
Head width (without eyes) >0.58 mm; r-m present in forewing
; 3
3. Petiolar node as seen from above concave behind; mandibles
elongate but only weakly denticulate, of a particular form (Fig. 6) ; postpetiolar disc with a distinct anteromedian longitudinal carina prolatus
7»
Psyche
[March
Petiolar node as seen from above with a straight or convex posterior margin; mandibles not as in Fig. 6; no carina on post-
petiolar dorsum 4
4. Same as couplet 3 of worker key, above: major vs. clavicornis. Note: T. foreli is not included in the key. It may be the female of T. major.
Acknowledgements
I am indebted to my colleagues G. Grandi, F. Keiser, W. W. Kempf and C. Portocarrero for the opportunity to review important material in their care. This project was supported by National Science Foundation Grants G-23680 and GB-2175, and publication was aided by a grant from the Grace Griswold Fund, Department of Entomology and Limnology, Cornell University.
References Cited
Brown, W. L., Jr.
1950. Morphological, taxonomic and other notes on ants. Wasmann Jour. Biol., San Francisco, 8: 241-250.
1953 (1952). Composition of the ant tribe Typhlomyrmicini. Psyche,
59: 104.
1958. Contributions toward a reclassification of the Formicidae. II. Tribe Ectatommini (Hymenoptera). Bull. Mus. comp. Zool. Harv., 118: 173-362.
Emery, C.
1911. Ponerinae, in Wytsman, Gen. Insect., 118: 125 pp. and 3 pi. Kempf, W.W.
1961. A survey of the ants of the soil fauna in Surinam (Hymenoptera: Formicidae). Studia Ent., Petropolis, Brazil, 4: 481-524. Wheeler, G. C., and J. Wheeler
1952. The ant larvae of the subfamily Ponerinae — Part I. Amer.
midi. Nat, 48: 111-144, 5 pi. Part II. Ibid., pp. 604-672, 2 pi. 1964. The ant larvae of the subfamily Ponerinae: Supplement. Ann. ent. Soc. Amer., 57: 443-462, 19 figs.
A REVISION OF THE ANT TRIBE CARDIOCONDYLINI (HYMENOPTERA, FORMICIDAE)
I. The Genus Prosopidris Wheeler1
By Jonathan Reiskind Biological Laboratories, Harvard University
This is the first of a series of papers revising the tribe Cardio- condylini, a group of small myrm'icine ants (up to 3.0 mm in length), characterized by the possession of a pedunculate petiole, a wide postpetiolar dorsum, propodeal spines, and a general lack of pilosity. The tribe currently embraces the genera Cardiocondyla Emery (virtually cosmopolitan, mainly Old World), Xenometra Emery (West Indian) and Prosopidris (Papuasian). Xenometra includes a single poorly known species which is apparently parasitic on Cardio- condyla emeryi (Emery, 1917). Cardiocondyla and Prosopidris are free living, and form small inconspicuous colonies usually in the soil or in rotting logs.
In this contribution Wheeler’s subgenus Cardiocondyla (Prosopi- dris) is elevated to full generic status. The worker and female types of its single previously known species, P. sima Wheeler (Philippines), are redescribed and a second species, P. papuana (New Guinea), is newly described. The papuana material includes the first known male of the genus, along with females and workers. The male is highly ergatoid, like those of some Cardiocondyla species.
Measurements and Indices: In order to characterize properly cardio- condyline ants the following measurements and indices (with their abbreviations) will be used in this and succeeding papers:
Head Length (HL) — Maximum length of head in frontal view from clypeal apex to posterior border of occiput.
Head Width (HW) — Width of head measured in frontal view immediately behind the eyes.
Scape Length (SL) — Maximum measurable length of scape, not including its articular condyle.
Weber’s Length (WL) — Weber’s length of mesosoma ( = ali- trunk), measured in lateral view along a line connecting the place where the prono- tum joins the cervix and the apices of the
Research supported by U.S. National Science Foundation Grant No. GB1634.
79
8o
Psyche
[March
the flanges of the sides of the propodeal declivity (Brown, 1953). The lateral view- ing position of specimens must be standard- ized by lining up the tips of the propodeal spines.
Pronotum Width (PW) — Width of pronotum, measured in dorsal view, disregarding humeral angles. Propodeal Spine L(ength) — Maximum measurable length of pro- podeal spine, in lateral view, from its apex to the far edge of the propodeal spiracle. Petiole L(ength) — Length of petiole, measured in lateral view, from the apices of the flanges of the sides of the propodeal declivity to the dorsal posterior border of the petiolar tergite. Petiole Height — Height of petiole, measured in lateral view, on a line perpendicular to the petiole length ; from the dorsalmost point on the petiolar node to the ventral side.
Postpetiole W(idth) — Maximum width of postpetiolar dorsum. Postpetiole L(ength) — Midline length of postpetiolar dorsum. Cephalic Index (Cl) — Head width expressed as a percentage of
head length. HW X 100/HL Scape Index (SI) — SL X 100/HW Mesosomal Index — PW X 100/WL
Propodeal Spine Index (Spl) — Propodeal Spine L. X 100/WL Petiolar Index — Petiole L. X ioo/Petiole Height Postpetiolar Index — Postpetiole L. X ioo/Postpetiole W.
Genus Prosopidris Wheeler new status
Cardiocondyla subgenus (Prosopidris) Wheeler, 1935, Psyche 42: 40-41.
Type-species: Cardiocondyla (Prosopidris) sima Wheeler, Ibid: 42-43.
Wheeler’s subgenus Prosopidris is here raised to full generic status. This change of rank is based on the distinctive clypeus, the 11- segmented antennae, and the high mesosoma, which readily distinguish Prosopidris from Cardiocondyla , though the two genera are obviously closely related.
Wheeler’s (1935) subgeneric diagnosis requires some modification in light of the new species, Prosopidris papuana, described below. In the workers the mesometanotal impression is lacking or very feeble. In the female wing venation the stub of vein Rs, extending beyond the cubital cell, may or may not be present. The only known male of the genus is described below.
1965]
Reiskind — Cardiocondylini
81
Prosopidris papuana new species (Figures 1-8)
WORKER
Measurements: Based on the holotype and 4 paratype workers (holotype cited first) : HL 0.59 mm, 0.59-0.60 mm; HW 0.44 mm, 0.44 mm; Cl 75, 73-75 ; SL 0.49 mm, 0.49-0.52 mm; SI 111, m- 1 1 8 ; PW 0.33 mm, 0.33 mm; WL 0.70 mm, 0.68-0.70 mm; Meso- somal Index 47, 47-48; Propodeal Spine L. 0.18 mm, 0.18-0.20 mm; Spl 26, 26-29; Petiole L. 0.28 mm, 0.28-0.29 mm; Petiole Height 0.17 mm, 0.17-0.18 mm; Petiolar Index 159, 155-164; Postpetiole W. 0.24 mm, 0.24 mm; Postpetiole L. 0.15 mm, 0.15-0.16 mm; Postpetiolar Index 60, 60-65. Palpal formula Maxillary 5 : Labial 3 (paratype dissected).
Description: General form is shown in Figures 1 and 2. Head subrectangular, sides sub-parallel, slightly convex; posterior corners rounded ; occipital border slightly concave. Scape slightly curved distally, exceeding occipital border by a distance about equal to its maximum diameter. Antennae 1 1 -segmented, with a 3-segmented club, the apical segment about 3 times as long as the preapical. Eyes strongly convex, maximum diameter (in all specimens) 0.13 mm; their anterior edges at a distance of 0.7 X their diameter from the lateral genal border. Outer borders of mandibles convex; masticatory border with 5 teeth, the 2 apical enlarged. Clypeus produced ante- riorly, forming a shelf over the mandibular bases and bulging out- wards; anterior border convex, with thick lateral sections anterior to each antennal sulcus; posterior portion extending broadly between the frontal carinae.
Mesosomal profile as in Figure 2 ; lacking a promesonotal con- striction, humeral angles indistinctly rounded. Mesometanotal im- pression weakly developed on mesosomal dorsum as a very shallow concavity, lacking an incised sutural trace. Propodeal spines moderate- ly long, pointed, uncurved. Petiolar node from above, oval, distinctly wider than long. Postpetiole with convex anterior and posterior margins and convex sides. Dorsal width of first gastric segment 2.1 X postpetiole width.
Head and mesosoma strongly reticulate-punctate and mat; petiole, postpetiole and gaster moderately reticulate and shining.
Yellow pilosity on clypeus, mandibles and posterior end of gaster. Yellowish adpressed pubescence dense on antennae, moderate on head, petiolar node, postpetiolar dorsum, and gaster.
Entire body yellow to dark yellow.
82
Psyche
[March
Figs. 1-3. Prosopidris papuana new species. Figs. 1 and 2, worker (Holotype). Fig. 1. Head, frontal view. Fig. 2. Mesosoma and petiole, lateral view. Fig. 3, female (Paratype) ; mesosoma and petiole, lateral view. Scale line: 0.40 mm.
Type locality: Bisianumu, near Sogeri, Papua, 15-20 March, 1955 (E. O. Wilson acc. 655).
FEMALE
Measurements : Based on 5 specimens collected with the workers: HL 0.60-0.62 mm; HW 0.46-0.47 mm; Cl 76-78; SL 0.49-0.52 mm; SI 103-114; PW 0.46-0.47 mm; WL 0.83-0.85 mm; Meso- somal Index 54-57 ; Propodeal Spine L. 0.20-0.21 mm; Spl 23-25; Petiole L. 0.31-0.33 mm; Petiole Height 0.20-0.21 mm; Petiolar Index 146-167; Postpetiole W. 0.28-0.29 mm; Postpetiole L. 0.16 mm; Postpetiolar Index 56-59.
Description : General form of mesosoma as in Figure 3. Head, antennae, clypeus and mandibles similar to worker. Ocelli slightly elevated. Eyes strongly convex, maximum diameter (in all specimens) 0.16 mm; their anterior edges at a distance of 0.5 X their diameter from the lateral genal border. Propodeal spines longer and thinner than worker. Petiolar node and postpetiolar dorsum as in worker. Dorsal width of first gastric segment 2.2-24 X postpetiole width.
Head and mesosoma strongly and coarsely reticulate-punctate and moderately shining; petiole, postpetiole and gaster moderately reti- culate and shining.
Yellow pilosity as in worker. Yellowish adpressed pubescence dense
1965]
Reiskind — Cardiocondylini
8 3
on antennae, postpetiolar dorsum and gaster, and moderate on head and mesosoma.
Color as in worker, pterostigma pale brown, wing veins pale yellow; apical stub of wing vein Rs absent beyond cubital cell.
MALE
Measurements: Based on a single specimen collected with the
workers: HL 0.59 mm; HW 0.47 mm; Cl 81; SL 0.44 mm; SI 93; PW 0.36 mm; WL 0.75 mm; Mesosomal Index 48; Propodeal Spine L. 0.15 mm; Spl 20; Petiole L. 0.34 mm; Petiole Height 0.21 mm; Petiolar Index 162; Postpetiole W. 0.29 mm; Postpetiole L. O.15 mm; Postpetiolar Index 52.
Description : Ergatoid. General form as shown in Figures 4 and 5.
Head with sides slightly divergent behind eyes, then strongly converging; posterior corners rounded; occipital border slightly concave. Scapes straight, their apices falling slightly short of the occipital border. Antennae 12-segmented. Funiculus and clypeus as in worker. Eyes strongly convex, maximum diameter 0.13 mm; their anterior edges at a distance of 0.7 X their diameter from the lateral genal border. Mandibles narrowly falcate, ending in a point. Frontal carinae more extensive than in the worker.
Mesosoma high, in lateral view, with no promesonotal constriction, but with distinctly rounded humeral angles; mesometanotal im- pression and constriction clearly developed, but lacking an incised suture. Propodeal spines blunt; shorter and wider at base than in worker. Petiolar node and postpetiole dorsally, as in worker. Dorsal width of first gastric segment 2.0 X postpetiole width.
Pro- and mesonotum strongly reticulate-punctate; rest of mesosoma, head, petiole, postpetiole and gaster moderately reticulate. Whole ant very shining.
Yellowish pilosity on clypeus and posterior end of gaster. Yellowish adpressed pubescence very dense on antennae, pronotum, mesonotum, gaster, and dorsa of petiole and postpetiole.
Entire body yellow.
Genitalia: General form shown in Figures 6, 7, and 8. Basal ring entire ; the genital foramen inclined ventrally. Gonoforceps each with a lobate, hirsute ventrolateral subapical extension; a peculiar large ventrally directed curved tooth arising on the ventral part of the inner face (Figure 8). Volsella hook shaped, its apex directed ventrally, the base swollen with a small acute ventrally directed tooth (Figure 8). Penis valves rounded apically, each with a row of fine teeth on its ventral edge.
84
Psyche
[March
Figs. 4-8. Prosopidris papuana new species. Figs. 4-8, male (Paratype). Fig. 4. Head, frontal view. Fig. 5. Mesosoma and petiole, lateral view. Fig. 6. Genitalia, dorsal and ventral views. Fig. 7. Genitalia, left half, external lateral view, showing gonoforceps and basal ring. Fig. 8. Geni- talia, right half, internal lateral view, showing penis valve, volsella, gonoforceps and basal ring. Scale line: 0.40 mm., for Figs. 4 and 5; 0.15 mm., for Figs. 6, 7 and 8.
Type deposition: The holotype and most paratypes are deposited in the Museum of Comparative Zoology at Harvard College, Cam- bridge, Massachusetts (Type No. 31156), single worker and female paratypes in the Australian National Insect Collection, CSIRO, Canberra.
Prosopidris sima Wheeler
Cardiocondyla (Prosopidris) sima Wheeler, 1935, Psyche 42: 40-43, figs.
The worker and female are adequately described (with figures of the worker) in Wheeler’s paper. The measurements and indices applied above to Prosopidris papuana have also been applied to P. sima and are listed here:
1965]
Reiskind — Cardiocondylini
85
WORKER
Based on 5 workers from Wheeler’s syntype series (MCZ Type No. 20798), the lectotype (by present designation, so labelled) cited first, the others paralectotypes: HL 0.57 mm, 0.55-0.57 mm; HW 0.46 mm, 0.44-0.46 mm ; Cl 80, 79-8o ; SL 0.5 1 mm, o.47"°*52 rnm ; SI hi, 107-114; PW 0.34 mm, 0.33-0.34 mm; WL 0.68 mm, 0.67-0.68 mm; Mesosomal Index 50, 48-50; Propodeal Spine L. 0.20 mm, 0.18-0.20 mm; Spl 29, 26-29; Petiole L. 0.28 mm, 0.26-0.29 mm; Petiole Height 0.20 mm, 0.19-0.20 mm; Petiolar Index 141, 1 3 3- 1 54J Postpetiole W. 0.23 mm, 0.23-0.24 mm; Postpetiole L. 0.16 mm, 0.15-0.16 mm; Postpetiolar Index 69, 64-71. Dorsal width of first gastric segment 2. 1-2.3 X as wide as postpetiole.
FEMALE
Based on three paralectotypes: HL 0.59 mm; HW 0.46-0.47 mm; Cl 78-81; SL 0.49-0.51 mm; SI 104-m; PW 0.46-0.47 mm; WL 0.80 mm; Mesosomal Index 57-59; Propodeal Spine L. 0.21-0.22 mm; Spl 27; Petiole L. 0.31-0.33 mm; Petiole Height 0.20-0.21 mm; Petiolar Index 148-167; Postpetiole W. 0.25-0.26 mm; Postpetiole L. 0.16 mm; Postpetiolar Index 63-65. Dorsal width of first gastric segment 2. 3-2. 4 X as wide as postpetiole.
Type Locality: Dansalan, on Lanao Island, Philippine Islands.
The following chart presents the major distinguishing character- istics of the two Trosopidris species considered above:
Prosopidris sima Wheeler
Prosopidris papuana n.sp.
Worker
Anterior clypeal border straight. Mesosoma almost smooth, shining. Yellow-brown.
Anterior clypeal border evenly con- vex.
Mesosoma mat, highly reticulate- punctate.
Yellow to dark yellow.
Female
Anterior clypeal border flattened.
Propodeal spines thicker than in worker.
Pronotum shining, almost smooth. Pronounced lateral propodeal striae converging on propodeal spine. Stub of vein Rs extending beyond the cubital cell.
Anterior clypeal border evenly con- vex.
Propodeal spines thinner than in worker.
Pronotum highly reticulate-punctate. No clear propodeal striae.
Stub of vein Rs beyond cubital cell lacking.
86
Psyche
[March
The author wishes to thank Dr. Robert W. Taylor of Harvard University for his invaluable assistance.
References
Brown, W. L. Jr.
1953. A Revision of The Dacetine Ant Genus Orectognathus. Mem. Queensland Mus. 13 (1): 84-104.
Emery, C.
1917. Questions de nomenclature et synonymies relatives a quelques genres et especes de Formicides. Bull. Soc. ent. Fr. 1917: 94-97. Wheeler, W. M.
1935. New Ants from the Philippines, Psyche 42: 38-52.
“QUEENLESSNESS,” WORKER SIBSHIP, AND COLONY VERSUS POPULATION STRUCTURE IN THE FORMICID GENUS RHY T'lD OPONERA
By Caryl P. Haskins1 and Roy M. Whelden2
INTRODUCTION
William Morton Wheeler, in his Colony Founding Among Ants (!933)> called special attention to the fact that in a number of formicid genera, and particularly in the socially primitive subfamilies Ponerinae and Cerapachyinae, typical alate female forms have never been described. In such genera as Onychomyrmex, Eusphinctus, Acanthostichus, Megaponera , and Plectroctena, this normal female may be replaced by a wingless ergatogyne, intermediate in structure between queen and worker. The same condition obtains among certain species of the archaic subfamily Myrmeciinae, as Wheeler also pointed out. In some cerapachyine species, normal females typically coexist with ergatogynes, and such caste duality occurs elsewhere also.
In certain respects, the shift from a “queen-like” toward a “worker- like” form of reproductive female is curious and striking. At first sight it may even suggest a reversal of the trend tending to emphasize the dichotomy between more “vegetative” and more actively “for- aging” forms which, as Wigglesworth (1954) and Kennedy (1961) have pointed out, is so characteristic of evolution in insects generally, whether at the level of “successive polymorphism” in the juvenile and adult phases of the individual, or of “alternative polymorphism” among adult populations of such forms as aphids, migratory locusts, and the social insects. Among the ant genera cited, however, evolution from alate to ergatoid reproductive may only superficially appear to lie in that direction, for the ergatoid is clearly at least as effective a reproductive as the winged female. In most species with such females, moreover, it seems likely that the ergatoid has been directly derived by a stabilization of a queen-worker intercaste, as Brown (i960) has suggested, and merely replaces the normal queen, with no drastic change in the general economy or structure of the colony. Even in the ponerine genus Leptogenys sens, str., where the laying female is no longer morphologically distinguishable from the worker, the course of evolution still seems relatively clear. As Wheeler (1933) pointed out, a normal female is present in the related Lobopelta langii , and
’Carnegie Institution, Washington, D.C.
2Haskins Laboratories, New Durham, New Hampshire
87
88
Psyche
[March
in L. ergatogyna the wingless laying female still possesses well de- veloped ocelli and a normal female thorax. In the genus Myrmecia similar series, ranging through increasing female microptery toward total aptery, can be assembled, culminating in the ergatoid-like gynes of, for instance, M. tarsata. In all these species, however, there is little evidence of any significant modification in the social structure of the colony, which remains at base a typically matrifilial community.
Certain other ponerine genera referred to by Wheeler, however, may present a rather different picture. In the genera Diacamma , Streblognathus , and Dinoponera, and in some species of Rkytido- ponera, for example, distinguishable ergatogynes have not been reported. A male of a Philippine species of Diacamma has been described in copula with a form indistinguishable from a worker by Wheeler and Chapman (1922), suggesting that in this organism the normal reproductive may have been replaced by what is in effect a worker-producing worker. Our overall knowledge of the first three genera, however, is hardly sufficient to support even speculation about their social situations at present. For the genus Rhytidoponera, however, something more is known or knowable, and it may quite possibly represent a situation of considerable interest to the student of social evolution in the Formicidae.
The Ponerine genus Rhytidoponera comprises an extensive but relatively compact series of species inhabiting the Australian and parts of the Melanesian and Malaysian areas, ranging from New Caledonia in the east through New Guinea and neighboring parts of Melanesia to Timor, the Moluccas, and the southern Philippines in the west and occcupying a large portion of the Australian continent and of Tasmania (Brown, 1954, 1958; Wilson, 1958, 1959a). They are members of the widely distributed ponerine tribe Ectatommini, bearing considerable resemblance in many respects to the generalized New World tropical genera Acanthoponera and Ectatomma , as well as to the pantropical Heteroponera, with the Old World components of which they may well have shared common ancestry. The females of Ectatomma and Acanthoponera, so far known, are of the normal winged form. Those of Heteroponera may be winged or ergatoid.
In at least three species of Rhytidoponera, R. impressa, R. purpurea , and R. chalybaeae , typical winged queens are the rule. Normally a single queen is found in each colony examined in the field, and com- munities appear to be initiated by isolated dealated females following a normal Ponerine dispersion and mating flight. These species are confined in distribution to well-watered and warmer areas, ranging from New Guinea and eastern Queensland rain forest southward
1965]
Haskins and Whelden — Rhytidoponera
89
along a belt down the eastern edge of Australia at least as far as south central Victoria (Brown, 1958).
In a second and much larger group of Rhytidoponera species, an exactly opposite situation obtains with respect to the queen. Many of these species include among the largest and most conspicuous members of the genus, are widely distributed, abundant, and well known, especially in the drier areas of Australia, and have been extensively collected over long periods. Yet in none of them has a reproductive morphologically or functionally distinguishable from a normal worker ever been described.
Finally there is a third group, designated by Brown (1958) the Rhytidoponera metallica complex, which may be the most interesting from the standpoint of social evolution. The type species is one of the most widely distributed and ubiquitous of Australian ants; an inhabit- ant of thickly populated as well as remote situations over a very large area both temperate and subtropical; and so familiar as to have been known to a wide public for many years by the popular name of “greenhead” ant. Alate typical queens of this species have been described, and are represented in limited numbers in some collections, notably that of the Harvard Museum of Comparative Zoology. Wheeler described a single dealate and possibly colony-founding female of R. inornata, a member of the complex from southwestern Australia in 1931 (Brown, 1958). Brown (1958) has described a dealate female of another related species, R. aspera , collected by H. Hacker in southeastern Queensland and also in the Harvard Museum of Comparative Zoology. A single perfect female of R. victoriae, taken by Brown at Seaford, Victoria, is in the same col- lection. But it is striking that so few typical females have been identified in a complex of species as extraordinarily abundant and well-collected. It is clear that the vast majority of colonies in nature must exist without such females.
Even more interesting is the fact that in no species of Rhytidoponera , including those of the metallica complex, has a queen-worker inter- mediate ever been recorded. This could suggest that evolution to the loss of the typical female took a somewhat different course from that in the Lobopelta-Leptogenys complex or even in Heteroponera. Instead of the alate female reproductive being morphologically modi- fied toward a stabilized intermediate between queen and worker while continuing the same functional role in the colony, the original queen caste may have disappeared entirely and one or more laying workers substituted as the usual reproductives. If, as Carroll Williams (in Brown, i960) has suggested, worker development in ants is due to a
90
Psyche
[March
precocious decline of juvenile hormone titer in the maturing larva, or if on the other hand as Brian (1959, 1961) proposes, to a sharp rise in the concentration of ecdyson near the critical period or periods of caste determination in larval ontogeny, it seems conceivable that mutations have accumulated in evolution affecting neurosecretory products or processes or timing, such that the threshold for worker- queen determination is passed only rarely in species of the R. metallic a complex, and never in many of the larger species. Alternatively, it is conceivable that dimorphic female reproductive forms originally existed, as they presently do, for example, in species of N eophyrcaces, one form being ergatoid and the other unmodified, and that further evolution resulted in the loss of the latter and so close an approxi- mation of the worker form, by the former that it is no longer dis- tinguishable except through its reproductive capacity. If the latter course has been the actual one, however, we should perhaps suspect that the fertile “workers” would be comparatively rare in the Rhyti- doponera colony, at least as rare as are the laying true queens in colonies of relatively primitive pleometrotic species. It appears, as will later be indicated, that they are in fact much more abundant.
Whatever the channel of evolution at the level of individual morphology, its end result poses some intriguing questions about the direction of evolution in Rhytidoponera at the level of the society. Some years ago Sturtevant (1938), and later Williams and Williams ( 1 95 7 ) > emphasized the evolutionary significance of the close family relationships which typically obtain among the members of the matrifilial colony so characteristic of the higher social Hymenoptera. Very recently W. D. Hamilton (1964), in two important papers, beginning with Haldane’s (1955) concept of the evolutionary sig- nificance of genetically based altruistic behavior, has derived a quantity in social evolution that he defines as “inclusive fitness.” It may be regarded in certain respects as an analogue at the social level of the concept of Darwinian reproductive fitness at the level of the individu- al. Like the latter it should be found to maximize in evolution. This maximization, in the social insects, has obviously involved an extra- ordinary evolution of worker altruism, at both structural and behavioral levels. Now as Hamilton demonstrates, it can be expected that such an evolution will be positively correlated with the maintenance of close genetic relatedness among the members of a colony. In the absence of parthenogenesis, the closest genetic relation- ship between queen and worker possible in a colony is that of mother to daughter, and among workers that of siblings. The pronounced evolutionary trends toward the inclusion of but two generations in the
1965]
Haskins and Whelden — Rhytidoponera
91
colony structure and toward the restriction of fertilized reproductives in the colony to a single or a few individuals, so evident in a great number of ant species, both evidently contribute to maintaining this maximal degree of relatedness within the community.
Exceptions to both trends, to be sure, are known. On the one hand, they are well illustrated by the puzzling situations explored in species of the Formica obscuriventris group (King, 1949, 1955 j King and Sallee, 1951) and the Formica rufa group (Sturtevant, 1938; Chauvin, Courtois, and Lecompt, 1961) where it appears that young reproductives, even of different species, may be adopted by large colonies, thus prolonging the existence of the colony well beyond the second generation and introducing both multiple queens and what must be a remarkable degree of unrelatedness among the worker personnel of individual colonies. Exceptions to the second trend may be presented by the numerous pleometrotic species of ants, though it is still unclear, as in the Polybiine and Polistine wasps, how usually such multiple reproductives are in fact sisters, and how often or how elaborately special behavior patterns may be adapted to restricting reproduction in practice to a single dominant female, or to preventing non-sisters from coexisting in the colony.
In this context, the course of social evolution in the genus Rhyti- doponera is of special interest, as is the probable degree of average relatedness among the workers of a single community of both normal and “queenless” forms. In the so-called “queenless” species, do workers in fact regularly give rise to workers? Is the same true of those species possessing functional typical queens? Is worker pro- duction accomplished in one or both groups through parthenogenetic thelytoky, so common among lower nonsocial Hymenoptera, and frequently reported in the Cape honey bee? Such thelytoky has been described in ants by several investigators over a long period of years, including Reichenbach (1902), Crawley (1912), and Comstock (1903) for Lasius niger, Haskins and Enzmann ( 1945) for Aphaeno- gaster picea and A. lamellidens, Soulie (i960) for Cremastog aster , and Otto (i960) for Formica polyctena, while Ledoux (1949, 1954) has reported extensively on a specialized social adaptation of thely- toky in the workers of Oecophylla. If such thelytoky obtains in Rhytidoponera the pattern of relatedness among colony members might be quite different than if laying workers possess developed spermatheca and are fertilized in the manner normal to ordinary ergatoid queens, by active low-flying males from other colonies. Such males are indeed characteristic of all species of Rhytidoponera , both those possessing and those lacking typical queens. Do laying workers
92
Psyche
[March
of the “queenless” species occur only one to a colony, or are they present in some numbers? If the latter, are they commonly actual siblings, or how closely are they, on average, related? Is there a tendency, in the “queenless” species, to confine worker production to a single individual even if a number of potential worker-producers are present? If such laying workers are in fact fertilized, are their mates normally derived from the same or from other colonies? Are such workers singly or multiply inseminated? Is a single individual inseminated more than once during its lifetime? What is the average contribution of male progeny by the non-fertilized members of the colony, and how is the production of males regulated ? How resistant is the genus to extreme inbreeding? Is it the rule that a proportion of each successive brood of workers brought to maturity in a colony is fertilized and that these individuals remain with the parent colony, so prolonging the life of the community well beyond the normal two generations, or do newly fecundated workers typically leave the parental nest? How indeed are new colonies normally formed?
Such questions as these must be answered before any critical assessment of the direction of social evolution in Rhytidoponera can be undertaken. The results reported in the present paper, derived in the course of some ten years of investigation of the genus both in the field and in the artificial nest, represent the early stages of an obser- vational attempt to provide answers to a very few of them.
SOURCE OF WORKER AND MALE BROOD IN A SPECIES OF RHYTIDOPONERA POSSESSING NORMAL QUEENS Rhytodoponera purpurea
Rhytidoponera purpurea is a typical member of the R. impressa group, in which normal queens are characteristic. A single such queen is typically found in each colony taken in the field. According to Brown (1954) the species occurs in New Guinea and ranges in Australia through the rain forests of the Cairns-Atherton Tableland region of northern Queensland.
On December 27, 1963, a typical, populous colony of R. purpurea , comprising the parent female, some 250 workers, and numerous larval and pupal brood including sexual males and females, was collected near Kuranda in northern Queensland. The following day a similar colony was taken at Millaa Millaa on the Atherton Tableland. In early January these colonies were housed in a type of modified earth- containing glass Lubbock nest used throughout these investigations. The colony from Kuranda was divided at the time of nesting into several isolated groups of workers with broods of cocoons and larvae. Only one such group had access to the brood female. The colony
1965]
Haskins and Whelden — Rhytidoponera
93
from Millaa Millaa was divided into two portions, one with and one without the brood queen. An interval was then allowed to permit the maturation of larvae and pupae in the nests at the time of capture. In the fragments of both colonies lacking parent queens, as well as ia those where the normal females were present, oviposition soon oc- curred, and fairly copious broods were shortly reared. By late March abundant pupae were present in several groups. In that month, and over a following period until mid- January, 1965, samples of cocoons were regularly withdrawn from these colony-fragments and opened, and their contained pupae scored for sex and caste. The results, in which the numbers of pupae in all the queenless fragments of Colony No. 1 are summed for each date, appear below:
|
No. of Colony |
Dates of Assay of Pupal |
1 and Young Adult Samples |
|
|
Fragment |
Examination Workers |
Males |
|
|
1 A (Kuranda — |
4/29/64 |
23 pupae ; 3 callows 0 |
|
|
with parent |
6/ 4/64 |
14 pupae ; |
0 |
|
female) |
7/12/64 |
15 pupae; 1 callow 0 |
|
|
9/20/64 |
15 pupae |
0 |
|
|
12/ 5/64 |
34 pupae |
0 |
|
|
1/19/65 |
30 pupae |
0 |
|
|
Total: 131 |
worker pupae; |
4 callow workers; 0 males. |
|
|
1 B-E (Kuranda — |
4/29/64 |
0 |
47 pupae; 26 adults |
|
fragments of |
5/10/64 |
0 |
3 pupae |
|
colony without |
6/ 4/64 |
0 |
12 pupae |
|
brood queen) |
7/12/64 |
0 |
10 pupae |
|
9/20/64 |
0 |
27 pupae |
|
|
12/ 5/64 |
0 |
34 pupae |
|
|
1/19/65 |
0 |
47 pupae |
|
|
Total: 180 male pupae; |
26 male adults; 0 workers. |
||
|
2 A (Millaa-Millaa |
— 4/29/64 |
23 pupae; 2 callows 0 |
|
|
with parent |
6/ 4/64 |
39 pupae |
0 |
|
female) |
6/22/64 |
1 pupa |
0 |
|
12/ 5/64 |
32 pupae |
0 |
|
|
1/19/65 |
29 pupae |
0 |
|
|
Total: 124 |
worker pupae ; |
2 callow workers; 0 males. |
|
|
2 B (Millaa-Millaa - |
- 4/29/64 |
3 callows* |
15 pupae |
|
fragment without |
6/ 4/64 |
0 |
5 pupae |
|
parent female) |
6/22/64 |
0 |
4 pupae |
|
9/20/64 |
0 |
8 pupae |
|
|
12/ 5/64 |
0 |
14 pupae |
|
|
1/19/65 |
0 |
14 pupae |
|
|
Total: 60 male |
pupae; 3 callow workers.* |
*These callow workers, found fresh-hatched on April 29, 1964, almost certainly represent the final fragment of maturing brood collected with the original colony, and seem with little doubt to have been progeny of the fertilized brood female.
94
Psyche
[March
Thus, summing the output of the two colonies together over the period of approximately a year after observation was begun, those fragments containing a brood queen produced a total of 255 pupae which were identified as workers, 6 identifiably callow workers, and no males. Those fragments of both colonies containing workers only brought to maturity a total of 240 pupae identified as males and 26 identified young adult males, a total of 266 males. In Colony No. 2 B, three callow workers were also brought to maturity at the time of an early count. It seems a safe assumption, however, that these indi- viduals represented the final, delayed residual of queen-laid brood “inherited” from the partitioning of the colony some three months earlier.
Thus it seems very clear that in R. purpurea worker brood is entirely derived from the fertilized brood female in typical Formicid fashion. Workers, however, can produce and rear a prolific male brood, at least in the absence of the parent female. Whether the male brood which typically appears seasonally in large numbers in normal wild colonies is queen- or worker-derived, or both, is an interesting and important question for future investigation. It applies with equal cogency, of course, to the vast range of “normal” Formicid species.
SOURCE OF WORKER AND MALE BROOD IN SPECIES WHERE NORMAL QUEENS ARE RARE OR ABSENT
Rhytidoponera metallica
Between December 23 and 25, 1963, a number of vigorous colonies of Rhytidoponera metallica were collected at various points in the Blackall Range in Queensland, Australia, some sixty miles north and thirty miles east of Brisbane. No perfect females were found. These colonies were housed and maintained in modified glass Lubbock nests of the same design as those used for R. purpurea. After a preliminary incubation period of approximately six months, to allow brood resident in the colonies at the time of capture to mature, samples of cocoons were withdrawn at intervals, opened, and the contained pupae scored for sex and caste. Callow workers that were obviously fresh-hatched were scored at the same time. The result are given in the table below (p. 95)-
Thus a total of 644 worker pupae or young adults were produced in the five “queenless” colonies of R. metallica over a period of little more than six months, and only 1 1 males. It seems clear that worker production by morphological workers is a normal feature of this species.
1965]
Haskins and Whelden — Rhytidoponera
95
|
Dates of |
Assay of Pupal and Young Adult S; |
arnples |
|||
|
Colony No. |
Examination |
Workers |
Males |
||
|
! |
7/12/64 |
20 |
pupae ; 1 fresh-hatched |
callow |
0 |
|
12/15/64 |
24 |
pupae; 13 fresh-hatched |
callows |
0 |
|
|
1/23/65 |
33 |
pupae ; 1 fresh-hatched |
callow |
2 pupae |
|
|
2. |
7/12/64 |
43 |
pupae |
2 pupae |
|
|
12/15/64 |
44 |
pupae; 36 fresh-hatched |
callows |
0 |
|
|
1/23/65 |
46 |
pupae |
4 pupae |
||
|
3. |
7/12/64 |
20 |
pupae |
0 |
|
|
12/15/64 |
38 |
pupae; 6 fresh-hatched |
callows |
0 |
|
|
1/23/65 |
69 |
pupae ; 4 fresh-hatched |
callows |
0 |
|
|
4. |
7/12/64 |
27 |
pupae |
0 |
|
|
12/15/64 |
67 |
pupae ; 28 fresh-hatched |
callows |
0 |
|
|
1/23/65 |
66 |
pupae ; 4 fresh-hatched |
callows |
2 |
|
|
5. |
12/15/64 |
31 |
pupae ; 23 fresh-hatched |
callows |
1 |
|
Queensland appears |
to lie near the northern limit of the |
natural |
range of R. me tallica. Ample confirmation that the same situation obtains elsewhere in its range, however, was provided by counts made from a single colony of the species taken at Sutherland, N. S. W., on June 2, 1952, and observed continuously in the artificial nest over a ten-year period. This colony contained no typical females when collected, though much later in its history some were produced, as will be described later. It was kept as a single unit in the standard glass modified Lubbock type of nest until July 8, 1956, when it was split into three portions, one of which perished rather shortly. The second and third were maintained in Lubbock nests until January 1, 1962, when the second also died out. The third portion survived somewhat longer, but eventually perished on July 30, 1962.
Throughout the ten years of observation, these two colony frag- ments were kept in closed foraging arenas, to which no males could enter from outside and from which no individuals matured within could escape. The ants established and maintained regular kitchen- middens within these arenas, outside the nests proper. The ambient humidity of the arenas was maintained low, and the contents of the middens therefore remained well preserved and readily recognizable for considerable periods. Thus periodic removal of the middens and examination of their contents could provide a rather accurate picture of the quality and type of brood produced.
Until June 13, 1954 (two years after observation was begun), only workers were brought to maturity. A count of cocoon fragments accumulated in the middens at intervals during this period thus
96
Psyche
[March