Cyphomyrmex salvini Forel 1899

Attini, Formicidae, Hymenoptera, Insecta, Arthropoda, Animalia

worker face view

worker lateral view

worker face view

worker lateral view

worker face view

worker lateral view


Guatemala south to Ecuador (type locality Panama). Costa Rica: widespread in wet to moist habitats, lowlands to cloud forest.


Mandibles with 5 teeth; preocular carina curving mesad toward frontal carina; antennal scrobe poorly defined; lateral vertex margins produced as acute teeth that project posterolaterally; pronotum with lateral and median tubercles.

Natural History

This species occurs in wet to moist forest habitats throughout the country, from lowlands to cloud forest. It is by far the most common Cyphomyrmex species in wet forest leaf litter. It does not usually occur in more open habitats, where it is replaced in abundance by C. rimosus. The nests are in between leaves in the leaf litter, in small pieces of rotten wood on the ground, and in rotten logs. Nests are also quite common in the subarboreal zone, in dead wood suspended in vegetation or under epiphytes, but usually within 2m of the ground. Nests may be partially or wholly constructed of accreted organic soil. Colonies are small, from tens to hundreds of workers. The few nests I have examined in their entirety have been monogynous; I have never seen a polygynous colony.

The center of the nest is the fungus garden, which usually contains two or three large caterpillar droppings and a few dozen fragments of dead insects. The dead insect parts are often brightly colored beetle elytra, and the workers appear to selectively harvest fragments that are very shiny or with metallic coloration (Figure 1). When exposed the nests look like small glittering piles of jewels, dotted with the green yeast-form fungus. Workers usually forage nocturnally and may be seen carrying over their head a caterpillar dropping or beetle elytron that is often several times their own size.

Figure 1. Partial contents of a Cyphomyrmex salvini nest, showing abundant dead insect fragments relative to workers (upper right). Click here for additional images of nest.

I observed a nest in Corcovado National Park that was a mass of accreted organic soil on the side of a tree about 1m high. The nest was 20x10cm, elliptical in outline, and 3-4cm thick, with numerous seedlings sprouting from the surface. The nest contained a single dealate queen. Inquilines and other nest occupants included one Rogeria tonduzi worker, isopods, millipedes, annelid worms, nematodes, unidentified insect larvae, mites, snails, and one each of 3 species of Staphylinidae. Another similar nest of accreted soil also contained a small nest of Pachycondyla unidentata.

Subarboreal nests of C. salvini can have spectacular panic evacuations in response to army ant raids. Individual workers, each carrying a larva, explode from the nest and may actually rain onto surrounding vegetation. The workers disperse and rest on leaf tips for a period of time before making their way back to the nest.

I have sometimes found larvae of C. salvini to be parasitized by Diapriidae. In Corcovado National Park I found a nest in flat chambers under the loose dorsal bark of a dead log. The area of the nest was roughly an ellipse 20x8cm. There were many alate queens, males, and over 100 workers. When I collected from the disturbed colony I noticed workers carrying large, blackened larvae. I later discovered that all these blackened larvae were larval skins covering developing diapriid parasitoids. All the healthy brood of the colony was white. I kept part of the colony alive and isolated some parasitized larvae. Seven to eight days later wasps emerged from both ant tended and isolated larvae. Each larva contained a single wasp. Wasps emerged by cutting a slit above the larval head capsule. I sent material to Lubomir Masner, who identified the wasps as genus Acanthopria.


Cyphomyrmex salvini is almost certainly a complex of sibling species, or at least shows some degree of morphological sorting by microhabitat. At various times I have tried to split Costa Rican salvini into as many as six or seven morphospecies. In a lowland rainforest site like La Selva Biological Station, a relatively small form with short scapes is collected abundantly in Winkler samples of sifted leaf litter from the forest floor, but is almost never collected by searching for nests (middle worker images above, specimen code INBIOCRI002278772). In contrast, a larger form with longer scapes is rarely collected in Winkler samples, but is commonly found when searching for nests (top worker images, specimen code INBIOCRI001254199). These nests are usually in dead wood lying on top of the litter or in the subarboreal zone. This size difference is also reflected in the queens. It appears that these two forms partition the forest floor, one occurring below the leaf litter, the other just above it.

Some collections have a relatively enlarged promesonotum compared to others, and this form seems more abundant in mid-elevation sites.

The Monteverde cloud forest contains one common species of Cyphomyrmex, and it is an almost black, somewhat more wrinkled version of lowland salvini (bottom worker images, specimen code INBIOCRI001281765). There seems to be a fairly abrupt transition from the dark cloud forest form to lighter-colored and somewhat smoother forms just downslope in either direction. The form of salvini in the southern Pacific lowlands (Corcovado National Park and vicinity) is a bit different from either of the La Selva forms, appearing intermediate.

In spite of this degree of spatial patterning, all of the characters show overlapping distributions and I have given up trying to separate them for now.

Page author:

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

Date of this version: 22 February 2004.
Previous versions of this page:
Go back to top

Go to Ants of Costa Rica Homepage