Summary1. According to the pyrodiversity paradigm, a wide range of fire regimes is required to maintain biodiversity in fire-prone landscapes. However, the requisite level of pyrodiversity has seldom been tested and may actually be very low. 2. Here, we examine the sensitivity of tropical savanna ants to variation in fire regimes using results from a long-term fire experiment near Darwin, Australia. Six experimental fire regimes, with varying fire frequency and seasonality, have been applied to 18 one-ha plots in three replicated blocks since 2004, with ants sampled prior to experimental burning and then annually after up to 2009. Our primary focus is on the extent to which different patterns of ant richness and composition are associated with each of the six treatment regimes, or whether there is such high overlap that differences only become apparent when experimental treatments are grouped to provide strongly contrasting fire regimes. 3. When treating each of the six fire treatments separately, we were unable to detect a significant influence of fire on any ant community response variable. We were only able to detect significant ant responses when we grouped the experimental treatments into two contrasting fire frequency classes, low (burnt at most once over the 5 years) vs. high (burnt every 1 or 2 years). Even then, these responses were only evident after 3 years of fire treatment. 4. Our findings demonstrate that ant communities have very high resilience in relation to fire, with differences evident only between strongly contrasting regimes. Such resilience appears to be characteristic of savanna ants throughout the world. 5. Synthesis and applications. A large range of finely tuned fire regimes is unlikely to promote regional ant diversity. Rather, only very limited pyrodiversity (a combination of frequently and infrequently burnt areas) would appear to be sufficient for maintaining diversity at a landscape scale. It is important that fire management for biodiversity conservation focuses on the demonstrated requirements of target species, rather than be based on an assumption that 'pyrodiversity begets biodiversity'.

No abstract

Because different sampling techniques will provide different abundance values, it is currently difficult to compare results among many studies to form holistic understandings of how abundance influences ant ecology. Using three sampling methods in the same location we found pitfall traps best confirmed yellow crazy ant A. gracilipes presence recording the fewest zero values (9.1%), card counts were the least reliable (67.1%), and tuna lures were intermediate (30.1%). The abundance of A. gracilipes from card counts ranged from 0 to 20, in pitfall traps from 0 to 325, and the full range of tuna lure abundance scores (0–7) were sampled. We then determined the relationships between these three standard ant sampling techniques for Anoplolepis gracilipes abundance. Irrespective of the data transformation method, the strongest relationship was between pitfall traps and tuna lures, and the least strong was between pitfall traps and card counts. We then demonstrate the utility of this knowledge by analysing A. gracilipes abundance reported within published literature to show where the populations in those studies sit on an abundance spectrum. We also comment on insights into the relative utility of the three methods we used to determine A. gracilipes abundance among populations of varying abundance. Pitfall traps was the most reliable method to determine if the species was present at the sample level. Tuna lures were predominantly reliable for quantifying the presence of workers, but were limited by the number of workers that can gather around a spoonful of tuna. Card counts were the quickest method, but were seemingly only useful when A. gracilipes abundance is not low. Finally we discuss how environmental and biological variation needs to be accounted for in future studies to better standardise sampling protocols to help progress ecology as a precision science.

Epopostruma is an uncommon genus of myrmicine ants endemic to relatively mesic regions of southern and eastern Australia. Here we describe a new species recently recorded from the ‘Top End’ of Australia’s Northern Territory, E. topendi sp. n. from Melville Island and Nitmiluk National Park. The new species is known from one specimen from each of two sites in the Australian monsoonal tropics, the only records of the genus from this region. This addition brings the number of described species of Epopostruma to twenty.

Because different sampling techniques will provide different abundance values, it is currently difficult to compare results among many studies to form holistic understandings of how abundance influences ant ecology. Using three sampling methods in the same location we found pitfall traps best confirmed A. gracilipes presence recording the fewest zero values (9.1%), card counts were the least reliable (67.1%), and tuna lures were intermediate (30.1%). The abundance of A. gracilipes from card counts ranged from 0 to 20, in pitfall traps from 0 to 325, and the full range of tuna lure abundance scores (0-7) were sampled. We then determined the relationships between these three standard ant sampling techniques for the abundance of yellow crazy ant Anoplolepis gracilipes. Irrespective of the data transformation method, the strongest relationship was between pitfall traps and tuna lures, and the least strong was between pitfall traps and card counts. We then demonstrate the utility of this knowledge by analysing A. gracilipes abundance reported within published literature to show where the populations in those studies sit on an abundance spectrum. We also comment on insights into the relative utility of the three methods we used to determine A. gracilipes abundance among populations of varying abundance. Pitfall traps was the most reliable method to determine if the species was present at the sample level. Tuna lures were predominantly reliable for quantifying the presence of workers, but were limited by the number of workers that can gather around a spoonful of tuna. Card counts were the quickest method, but were seemingly only useful when A. gracilipes abundance is not low. Finally we discuss how environmental and biological variation needs to be accounted for in future studies to better standardise sampling protocols to help progress ecology as a precision science.