Crematogaster carinata Mayr 1862

Formicidae, Hymenoptera, Insecta, Arthropoda, Animalia

worker lateral view

worker face view

Image catalog (click here).

Images of lectotype (click here).

Images of lectotype of C. limata parabiotica (click here).


Costa Rica to Brazil (Rio de Janeiro), Bolivia.


Abundant long flexuous setae on face and mesosoma, relatively short propodeal spines, short petiole with anteroventral angle or tooth, and no anteroventral tooth on the postpetiole.

Description of worker

Color red brown; mandibles, antennal club, and tarsi usually lighter yellow; workers monomorphic in size.

Mandibles smooth and shining; clypeus usually with 5-6 longitudinal carinulae, these may be more abundant, making clypeus uniformly striate, or they may be nearly absent, making clypeus smooth and shining, especially medially; head about as long as wide, subquadrate, with emarginate posterior border; antenna with distinct two-segmented club, or third segment from end somewhat enlarged, blurring distinction between two and three-segmented club; scapes with abundant long erect setae; when scapes laid back from antennal insertions, they slightly surpass margin of vertex; face largely smooth and shining, with variable extent of striated region between antennal insertion and eye, and whorled above antennal insertion; face covered with abundant long flexuous white setae, no appressed pubescence; in face view abundant setae project from lateral and posterior margins.

Promesonotum in profile forming evenly convex surface, varying from flattish to forming peak at juncture of pro and mesonotum; propodeal suture deep in dorsal view but obscured in profile due to lateral carinulae that bridge the suture; posterior mesonotum curves smoothly into horizontal dorsal face of propodeum; propodeal spines short, projecting posteriorly such that they are more or less in same plane as dorsal face of propodeum; dorsal and posterior face of propodeum appear well differentiated in lateral view, the dorsal face confluent with the horizontal spines, the posterior face sloping down to petiolar insertion, but faces less differentiated medially; pronotal dorsum with variably developed longitudinal carinulae, strongest laterally, becoming weaker medially, interspaces smooth and shining; mesonotal dorsum with two strong, subparallel lateral carinae, interspace smooth and shining or with faint longitudinal carinulae; dorsal face of propodeum striatorugose, rugulae extending onto spines, posterior face smooth and shining; side of pronotum smooth and shining; katepisternum and ventral portion of side of propodeum variously punctatorugose; dorsal portion of side of propodeum smooth and shining; mesosomal dorsum with abundant long flexuous white setae, setae on pronotal humeri longest; femora and tibiae with abundant long erect setae.

Petiole in side view subtrapezoidal, varying in length and degree of narrowing anteriorly, weakly punctate to nearly smooth; anteroventral tooth well developed, often forming a right-angle to short acute tooth; dorsal face of petiole smooth and shining, elongate, widest posteriorly, regularly tapering anteriorly, with a long flexuous seta on each posterolateral tubercle and varying number of short setae along posterior border; postpetiole lacking ventral tooth, globular in dorsal view, with abundant erect setae; fourth abdominal tergite smooth and shining, with abundant long flexuous erect white setae, no appressed pubescence.


HL 0.561, 0.528, 0.657; HW 0.604, 0.563, 0.722; HC 0.565, 0.507, 0.664; SL 0.485, 0.512, 0.612; EL 0.130, 0.139, 0.175; A11L 0.215; A11W 0.101; A10L 0.113; A10W 0.089; A09L 0.068; A09W 0.064; A08L 0.051; A08W 0.051; WL 0.615, 0.585, 0.744; SPL 0.077, 0.109, 0.148; PTH 0.148, 0.137, 0.166; PTL 0.207, 0.213, 0.250; PTW 0.144, 0.138, 0.174; PPL 0.169, 0.129, 0.178; PPW 0.162, 0.119, 0.166; CI 108, 107, 110; OI 23, 26, 27; SI 86, 97, 93; PTHI 71, 64, 66; PTWI 70, 65, 70; PPI 96, 92, 93; SPI 13, 19, 20; ACI 0.64.

Description of queen

A normal queen (dorsal face of propodeum drops steeply from postscutellum and much of propodeum appears ventral to scutellum and postscutellum) with general shape, sculpture, and pilosity characters of the worker.

Natural History

Crematogaster carinata is a biologically complex species and there is almost certainly some degree of genetic structuring that is not readily revealed by external morphology. This species and its close relative C. levior often have large polygynous colonies that blanket large sections of forest, with various degrees of mutual foraging and nesting with other unrelated ant species. Forel observed this species nesting with Dolichoderus debilis in Colombia, and coined the term parabiosis for the mutual sharing of nest space and foraging columns by multiple species (Forel 1898). Further examples of parabiosis involving carinata (as parabiotica) were described by Wheeler (1921a).

In the Atlantic lowland forests of northeastern Costa Rica, including coastal strand vegetation, C. carinata forms large polygynous, perhaps even unicolonial colonies. Columns of workers are spread over second growth vegetation and multiple crowns of trees, with small clusters of physogastric queens, brood, and workers dispersed in dead twigs and branches and under epiphytes. There is no obvious colony center, and it is difficult to discern colony boundaries. At La Selva Biological Station it is one of the most common species in the canopy. It occurred in 24 of 52 canopy fogging events. In an old treefall I once found a small aroid with a carinata nest in and around the root ball. The ants had covered the roots with a mass of carton material to form a nest volume about the size of a large orange. Inside were abundant workers, brood, and 43 dealate queens. Some of these queens had torn remnants of wings, suggesting they never left the nest for a nuptial flight and perhaps never mated. In general when multiple queens are found in nests they are evenly distributed in the nest volume, individually or in pairs. Brood is segregated by size. I observed a similar occurrence of large polygynous colonies in the Santa Marta region of Colombia, the same site as Forel's original observations of parabiosis. In contrast to these observations of polygyny, discrete nests with single physogastric queens seem to be the rule in the Pacific and southeastern Atlantic lowlands of Costa Rica, and at some sites in Venezuela.

On occasion I find aggregations of workers only, with no brood. All the dead sticks in a patch of forest understory will be filled with workers. The nest aggregation will be bounded, with all nests within about a meter of each other, but no colony center or area with brood can be found. In different parts of its range, in both polygynous and monogynous forms, Crematogaster carinata may form parabiotic associations with other ants. In the La Selva forest canopy it cohabits large ant gardens with Odontomachus panamensis. Because the ant gardens are conspicuous and packed with Crematogaster workers and brood, it can appear that the Crematogaster are a specialized associate with Odontomachus, but closer inspection reveals that the ant gardens are nodes of higher Crematogaster density in a sea of thinly spread Crematogaster nests. It is possible that the ants in Odontomachus nests are genetically and behaviorally differentiated from those in the diffuse polydomous colonies, but I have not found any evidence for this based on external morphology.

In the Pacific lowlands of Costa Rica I have twice seen parabiotic associations between carinata and Dolichoderus species. At Carara I observed a parabiotic association between carinata and D. debilis. Several nests were in a cluster of dead branches. Crematogaster workers and brood were distributed in multiple chambers, and one chamber contained a single physogastric queen. The nests of the two species were contiguous and interdigitated, with interconnections among chambers, but they were still largely segregated. In general the Crematogaster occupied smaller and more peripheral chambers, while the Dolichoderus occupied larger chambers in the center of the branches. In some peripheral chambers I found workers of both species together, but these chambers never contained brood. Any chamber with brood always contained only one species. In Corcovado National Park I observed parabiotic foraging involving carinata and D. inermis Mackay. Workers of both species were using the same foraging trails, and both species occurred together in clusters of non-foraging workers. Davidson has multiple observations of parabiotic foraging between carinata and D. debilis in Peru (pers. comm.), and many similar observations are reported by Forel (1898) and Wheeler (1921a).

Wherever I have observed carinata it makes use of carton construction to a variable extent. Construction may be as small as a 2cm diameter shelter over a single scale insect. Several times I have seen a 5-10cm wide globular mass of carton extending from the end of a rotten stick or investing epiphyte roots, extending the nest volume. In very humid areas these small carton nests may sprout epiphyte seedlings and in some cases form small ant gardens. Workers form large and more fully developed ant gardens only when in the presence of a larger parabiotic associate.

Workers may be found foraging day or night, and they are generalized omnivores. I have seen them scavenging dead or wounded insects, they recruit to carbohydrate and protein baits, and they are frequent visitors at extrafloral nectaries.


Costa Rican species in the limata complex are brasiliensis, carinata, limata, and tenuicula. All have abundant erect flexuous setae on the face, moderate length to short posteriorly directed propodeal spines, and elongate tapering petioles. The four species can be difficult to separate. They differ primarily in the nature of the ventral processes of the petiole and postpetiole. Crematogaster carinata has a squared-off anteroventral petiolar process and no postpetiolar process. Both brasiliensis and tenuicula have ventral postpetiolar processes. Crematogaster limata usually lacks a petiolar process but can be very like carinata in some cases. Crematogaster limata is a larger ant with longer propodeal spines. Crematogaster carinata may also be confused with foliocrypta, but foliocrypta has appressed rather than erect tibial pilosity.

Crematogaster carinata is also very similar to the South American species Crematogaster levior Longino. The two species are not cleanly separable on morphological grounds. Crematogaster levior is a specialized parabiotic associate of Camponotus femoratus (Fabricius). Crematogaster levior and Camponotus femoratus inhabit ant gardens throughout Amazonian South America. All Crematogaster material collected from these ant gardens has the worker pronotum completely smooth and shining. In contrast, I have identified as carinata all material with longitudinal carinulae on the pronotum. However, on purely morphological grounds there is continuous variation from strongly carinate forms, to forms with faint traces of carinulae, to the completely smooth pronotum of levior. Crematogaster levior is polygynous and polydomous like some populations of carinata. Crematogaster levior is always parabiotic with Camponotus femoratus, while C. carinata is facultatively parabiotic with other large ants in the genera Dolichoderus and Odontomachus.

Further evidence for separate species has been found by D. Davidson at her study site at Cocha Cashu Biological Station in Amazonian Peru. In early studies of ant gardens and their ants (Davidson 1988, Seidel et al. 1990, Davidson et al. 1990) I helped identify the Crematogaster and I failed to differentiate levior and carinata. More recent observations on behavior and defensive chemistry suggest discrete sympatric forms, with levior inhabiting ant gardens with Camponotus and carinata found outside those gardens (Davidson pers. comm.). Davidson has discovered that levior has lost its chemical defense and instead relies on the defensive capabilities of C. femoratus. In contrast, carinata has retained chemical defenses. Reflecting the differences in defensive chemistry, levior does not elevate the gaster when disturbed; carinata does. Crematogaster carinata is often found living parabiotically with Dolichoderus debilis, and in this case it is the Dolichoderus that has lost its chemical defense, relying on the Crematogaster. Davidson's observations on carinata in Peru not withstanding, one should not assume that carinata exhibits the same behavior and defensive chemistry throughout the range. It would be worth investigating whether carinata at La Selva Biological Station, which lives in ant gardens with the aggressive species Odontomachus panamensis, exhibits behavior and chemistry more like levior. In contrast, the carinata from Costa Rica's Pacific slope, which is monogynous and parabiotic with non-aggressive Dolichoderus, might be like Peruvian carinata.

Wheeler's (1921a) discussion of parabiosis included observations of carinata living parabiotically with Dolichoderus, and levior living parabiotically with Camponotus femoratus.

Within carinata, there is pronounced variation in the degree of anterior narrowing of the petiole in side view, strength of pronotal carinulae, size of propodeal spines, and some details of queen size and petiole shape. Crematogaster levior is more likely to be a "good" species, in terms of phylogeny and/or genetic similarity. Crematogaster carinata, with its greater geographic range and morphological variability, is more likely to be a set of allopatric or parapatric species, or even a set of broadly sympatric cryptic species.


C. carinata Mayr 1862. Costa Rica to Brazil (Rio de Janeiro), Bolivia.
= parabiotica Forel 1904. Colombia.

Literature Cited

Page author:

John T. Longino, The Evergreen State College, Olympia WA 98505 USA.

Date of this version: 19 June 2003.

Previous versions of this page:
Go back to top

Go to Ants of Costa Rica Homepage