Understanding the global ant generic diversity


**New** The website is now available at this address: click on the link



In autumn 2007, while driving back from a field excursion with Gary Alpert and my advisor, Rob Dunn, I overheard a discussion concerning various ant genera.  As I had missed the beginning of the conversation and knew of the genera in question to be common to North America, I asked naively if they were referring to ants local to Massachusetts, Gary’s home state. Their answer surprised me: “No, we are talking about ants native to the Philippines.”

As Massachusetts and the Philippines are not ecologically similar areas, I was surprised to hear about genera that not only covered large scale distributions but also existed within vastly different biomes. With my new knowledge on ant distributions, I tried to locate maps that would indicate the world-wide and regional presence of specific genera.  I quickly found that the maps I desired didn’t exist.  

A few weeks later, I was scouring through literature on the geographic presence of different ant genera and beginning to formulate distribution maps.

 Why create diversity maps at the genus level?

According to Antbase.org, there are currently 12,451 known species of ants (2008).  Many of these species are only described, with little to no recorded data on regional distribution. However, these species represent just a fraction of the actual number of ants estimated to be present globally. The Formicidae family is believed to have over 20,000 to 30,000 species. Therefore, species-based diversity maps would be difficult to create without a wide range of uncertainty and error. In addition, there are currently existing complexes of species that will most likely be divided into further taxonomic subsets in the future (i.e., Aphaenogaster rudis and Pachycondyla chinensis), thus making a stable species-based map unrealistic.

Unlike species-based diversity maps, genus-based maps are more stable. Relatively few new genera are discovered, and major taxonomic revisions are less frequent. Furthermore, at the genus-level, identification errors are less likely, and datum compiled on regional genera distribution has both greater precision and accuracy. This perspective on genus-level classification is further supported in a recent article by Phil Ward (2007):

“The genus-level classification of ants is more stable, with about 288 extant ant genera currently recognized (Bolton et al. 2006). Most of these genera are reasonably well demarcated and readily identified (Bolton, 1994),  but there are more than a dozen ant genera which are either weakly or very broadly defined, and which appear to be non-monophyletic by virtue of their exclusion of derivative (satellite) genera.”

Genus-level distribution maps reduce taxonomic error and, ultimately, will maintain database integrity with the discovery and incorporation of new species.

What geographic scales and boundaries are most useful for maps?

As scientists, we often like to divide the world into distinct biological biomes.  However, when creating distribution maps, partitioning the globe into countries, provinces, regions, states, etc, which all reflect historical notions, serves more practical for widespread use (though not necessarily the study of specific biomes). Although political borders may have no place in a strictly biological context, it is necessary to recognize that in widespread publications, a political map format is often more transmittable to our audiences. Scientists may limit their work to state lines or country borders, and regionally published ant species lists are in the vast majority of cases defined by political borders and not by biological ones (probably because the latter is often harder to draw a line around). As my main objective was to develop useful global maps, and as species presence lists are often based on regionally published records, defining the geographic distribution of ants based on political borders proved most logical. However, in each case, I have tried to restrain regions to the smallest political unity possible. By doing so, I have attempted to restrict the variation between biomes inside each specific area. Most regions are defined by country borders. In some countries, like Australia, Canada, China, Japan, Russia, and the USA, a smaller political scale was used (i.e., states and provinces). This tentative is also in process with Argentina, Brazil, Mexico and India (potentially Colombia), but the absence of complete regional data often makes this process difficult.

Islands can serve as fantastic natural laboratories for biological invasions and biogeography theories, and many works are based on studies conducted at these isolated sites.  I was able to locate and incorporate complete or partial list of species for 200 islands into my genus-based global diversity maps. In these maps, islands are considered separated from the rest of the mainland generic distribution.

Why now?

Utilized wisely, the internet can provide a wealth of knowledge for database formulation. Never in human history has so much data been available to an individual. Publications are now easily available online and websites created by both experts and novices are quickly accessible. Studying the ants of Africa, Australia, Costa Rica, Europe, Japan, the United States, and many other parts of the world is now as simple as a click of the mouse.

After several months of sorting through literature reviews in numerous languages, working with fantastic websites with local and regional lists of ant presence, and via direct contact with specialists (thanks for your help in fighting with various taxonomical revisions and the presence of exotic species), I was able to generate 292 genera-specific maps that represent 303 political divisions (and almost 198 islands).

In the future, all 292 genera-specific maps will be available on this website along with a complete listing of genera found in political entities world-wide. An initial version of global genera richness and two genera-specific maps are below. The first map indicates regional quality of database information: good (green), insufficient (yellow), very poor (red), or no information/no known presence (white).

Fig.1: Regional data quality estimations. Green-coded regions indicate sufficient data and most accurate estimations, yellow indicates regions with insufficient data, and red indicates areas with poor data quality. Note: Distribution data is continuously being updated and this figure may not provide the most current estimation qualities.



Fig.2: Global generic diversity estimation. Darker areas represent regions with higher generic diversity. This figure should be used in conjuncture with fig.1 (data estimation qualities).



Fig.3: Known distribution of Cephalotes. Red indicates genus presence; green indicates regions of probable presence.



Fig.3: Known distribution of Polyrhachis. Red indicates genus presence; green indicates regions of probable presence.



Updated July 21, 2008