An isolated second abdominal segment constitutes the characteristic petiole (blue) in ants. Pachycondyla stigma worker (Scanning Electron Micrograph, Roberto Keller/AMNH)

The easiest way to know you are looking at an ant is to pay attention to its waist: if it consists of one or two nicely isolated segments you can be sure you made a positive identification. The basal condition for the family, common to all ants, is to have the second abdominal segment in the shape of a node or scale and distinctly isolated from the rest of the abdomen to form a petiole (remember that the first abdominal segment is coupled to the thorax as the propodeum). The functional advantage of such novel architecture seems to be an enhanced articulation between body segments, and thus greater mobility for a posterior part of the body that bears the ant’s weapons in the form of a sting or other specialized chemical producing  organs like the acidopore.¹

Petiole (blue) in a Metapolybia cingulata vespid (Scanning Electron Micrograph, Roberto Keller/AMNH)

Adetomyrma sp. worker showing the broad posterior attachment of the petiole (in blue; Scanning Electron Micrograph, Roberto Keller/AMNH)

In reality, the presence of a petiole is not as clear cut as we would like from a systematic point of view. Wasps in other families may have a petiole as isolated as many ants (see the vespid Metapolybia above), and the petiole in some ants can have such a broad posterior attachment as to be quite similar to the usual condition found in the rest of the stinging wasps and bees (see Adetomyrma above). Still, our current understanding of phylogeny, both in terms of the position of Formicidae within Aculeata as well as the internal relationships within ants, suggests that the the petiole originated anew in the common ancestor of the group. In the case of Adetomyrma, even though its parent clade Amblyoponinae is believed to be close to the root of the ant tree, the genus is well nested within the subfamily and so the “unpetiolated” condition must be explained as a secondary derivation from a petiolated condition, as Phil Ward discussed when he first described this taxon ².

Petiole (blue) and postpetiole (purple) in a Manica rubida worker (Scanning Electron Micrograph, Roberto Keller/AMNH)

Petiole (blue) and postpetiole (purple) in a Aenictus binghami worker (Scanning Electron Micrograph, Roberto Keller/AMNH)

More interesting among ants is the subsequent modification of the third abdominal segment into a similar constricted node to form the postpetiole, thus resulting in ants with two segmented waists. This appears to have occurred at least seven times in parallel, in the subfamilies Aenictinae, Agroecomyrmecinae, Ecitoninae, Leptanillinae, Leptanilloidinae, Myrmicinae, and Pseudomyrmecinae (see phylogeny below). Moreover, some ants have a condition that seems intermediate between an undifferentiated segment and a true postpetiole (e.g., Cerapachyinae, Myrmeciinae). The similarity of this feature across the different unrelated groups is striking. It was once thought, for example, that the two segmented waist in myrmicines and pseudomyrmecines was a case of homology rather than independently derivation. Again, the functional advantage of having an extra “hinge” is greater flexibility of the metasoma. A weak, yet curious correlation is that ants that have evolved a two segmented waist retain, for the most part, a powerful sting, whereas in derived ants with single segmented waists the sting is either vestigial (e.g., Dorylinae) or is completely absent and has been replaced by chemical spraying glands (e.g., Dolichoderinae and Formicidae).

The origin of the petiole in ants can be traced back to the common ancestor of the group. Modification of the the third abdominal segment into a postpetiole have ocurred in parallel in the clades marked in purple. Subfamily names in yellow contains species with further tubulation of the abdominal segments IV to VI (ant phylogeny simplified from Brady et al. 2006, Moreau et al. 2006, and Rabeling et al. 2008)

Some structural details are of interest here. The metasomal segments are arranged like the cylindrical sections of a hand telescope, with each segment entering the previous one in a series. Abdominal sclerites (the skeletal plates that form each segment) have a well-marked anterior section corresponding to the part that articulates inside the preceding piece, which can be recognized by its smooth and shiny surface lacking hairs (colored in yellow in the images below). Barry Bolton³ introduced the terms presclerite for this anterior section and postsclerite for the remaining posterior one. All the metasomal segments are divided into these two sections but, as Robert Taylor⁴ pointed out, sometimes the segments bear a strong constriction right at the boundary between the presclerite and the postsclerite sections. He termed this “tubulation”, and noted that the transformation of a segment from an undifferentiated structure into a petiole or/and postpetiole entailed nothing but an extreme case of tubulation.

Presclerites (in yellow) along the segments of the metasoma in a Leptogenys sp worker. Note how the IVth abdominal segment is "tubulated" but there is no true postpetiole (Scanning Electron Micrograph, Roberto Keller/AMNH)

The concept of tubulation has been important in discussing the different degrees of constriction found in the third abdominal segment across the ants, from untubulated ones to full postpetiole, mainly from the point of view of phylogenetics and classification. But one overlooked but nevertheless highly interesting aspect of tubulation is its occurrence beyond the second and third segments of the abdomen and what does this pattern suggests about the underlying developmental process of segment modification.

Metasomal segments showing serial tubulation of segments in a Dorylus helvolus worker. Presclerites in yellow (Scanning Electron Micrograph, Roberto Keller/AMNH)

The evidence from comparative anatomy points towards tubulation as a case of morphological diversification through homeosis: once a genetic mechanism was established for the constriction of the second abdominal segment in the common ancestor of ants, this mechanism seems to have been coadapted, independently, for the formation of the postpetiole in the third segment, explaining not only its recurrence in phylogeny but also the almost identical nature of this structure in various adult workers of distantly related clades. Tubulation further occurs in the fourth, fifth and sixth abdominal segments in the genera Dorylus, Leptanilloides and Sphinctomyrmex (also independently as far as we know) forming a pattern of serially homologous constrictions of abdominal segments along the body axis.

Leptanilloides biconstricta worker. Note the elongated body and the serial tubulation of the metasoma (Scanning Electron Micrograph, Roberto Keller/AMNH)

Since tubulation is prominent among the minute, subterranean groups of ants (e. g., Leptanillinae and Leptanilloidinae), and since this seems to be the final frontier in the discovery of new ant forms, I predict that it won’t be long until an ant with a three segmented waist shows up in our winklers sacks.

Update June 9th, 2009: Added figure with phylogenetic tree.

Notes and references