PNAS yesterday carried breaking news by Rust and colleagues of extensive new fossil Indian amber deposits dating to about 50 million years ago and holding exquisitely preserved fossils. How important is this find?

It’s huge.

Not only does the discovery add an older record to bridge the excellent Dominican (15 mya) and Baltic (40 mya) deposits with the various and spotty Cretaceous deposits (65-140 mya ), but it provides insight on the history of the Indian subcontinent as it migrated from the south. The fossils will give biologists grist for years of testing hypothesis about what happened to India on its northward journey, and for examining broader patterns of evolution in the contained taxa. This is big, exciting stuff.

But enough praise for the research. I always find some nitpick to rant about, and Rust et al are no exception.

On the surface, the PNAS paper indicates the Indian paleofauna had non-intuitive connections to northern Europe and Australasia. Or, in the authors’ words, “India just prior to or immediately following contact shows little biological insularity.”

But this conclusion depends very much on the integrity of the taxonomic identifications. And I think they’ve botched their exemplar ant:

Fig 2B: a Myrmicine?

From the paper:

Despite the biased comparisons to the Baltic amber paleobiota, phylogenetic affinities of the identified insects in Cambay amber also lie with Recent taxa from southeastern Asia and Australasia… In ants, a highly sculptured, distinctive myrmecine [sic] (Fig. 2B) is closely related to a small group of Recent southeast Asian and Australian genera, Lordomyrma, Meyriella [sic], and Austromorium.

The fact that myrmicine and Mayriella are misspelled is a fat red flag that the specimen wasn’t examined by any ant specialists. Indeed, this ant shares only a suite of rather plastic traits with the fingered Australasian myrmicines: coarse areolate sculpture, propodeal spines, and the like. These are superficial characters that aren’t ideal for deeper phylogenetic inference.

Let me offer an alternate interpretation.

The mystery ant is not a myrmicine at all, Australian or otherwise. It’s an agroecomyrmecine. This is an ancient, presumably carnivorous lineage known only from Baltic amber (Agroecomyrmex), the Florissant fossil beds of Colorado (Eulithomyrmex), and a handful of rare extant Neotropical species (Tatuidris).

I just did a set of quick-and-dirty measurements from the mystery specimen and from several other Antweb specimens of head length, eye placement on head, and mesosomal length:

The Mystery Ant is more similar in eye placement and mesosomal proportions to Agroecomyrminae than to Rust et al’s hypothesized Australasian relatives

Several traits indicate an agroecomyrmecine affinity:

  • The position of the compound eye high on the head
  • The relatively broad and stubby mesosoma
  • The bulky post-petiole (=abd III) with a broad attachment to Abd IV
  • An apparently expanded Abd 4 tergite (not easy to see in the fossil).

Wheeler’s illustration of the Agroecomyrmex types from Baltic amber

If true, this alternate identity of the fossil is even more exciting. Agroecomyrmecinae is one of the oddest, rarest, and least understood of all ant lineages. An excellently-preserved fossil on a 50 million year old Indian subcontinent adds considerably to our knowledge of the group. Of course, this identification also nixes Rust et al’s Australasian-India connection, but I’ll happily trade that for another agroecomyrmecine.

My taxonomic whinging aside, though, the Cambay amber is a major discovery. Keep an eye out for the more detailed descriptions of the contained fossils that are sure to come.



Carpenter, F. M. 1930. The fossil ants of North America. Bulletin of the Museum of Comparative Zoology of Harvard College 70: 1-66.

Rust et al. 2010. Biogeographic and evolutionary implications of a diverse paleobiota in amber from the early Eocene of India. PNAS Published online before print doi: 10.1073/pnas.1007407107

Wheeler, W. M. 1915. The ants of the Baltic amber. Schriften der Physikalisch-Okonomischen Gellschaft zu Konigsberg 55: 1-142