The exponential increase from one species to many tens of thousands occupying a great variety of ecological modes could have been driven by either (i) the expansion of habitats occupied as tetrapods moved from the waterside to exploit new sectors of Earth's surface, such as forests, plains and deserts, or (ii) competition within a more-or-less constant amount of habitat, leading to specialization within communities so that more species live together, with each exploiting an ever-narrower range of resources.

There are 207 habitable ecological modes for tetrapods to use, but only 75 have been populated (36%). The number of families occupying each realized mode has increased from an average of one to four families, and is well correlated with the taxonomic diversity of the different tetrapod classes; amphibians in the Palaeozoic, reptiles in the Mesozoic and birds and mammals in the Cenozoic (figure 2a). Increasing mode utilization is also associated with dominant faunas: amphibians maximally added 0.5 families to a mode per Myr until the end of the Palaeozoic; in the Mesozoic, reptiles added up to 1.0 families until the Cretaceous–Tertiary extinction; and in the Cenozoic, birds and mammals peaked at adding 3.0 families (figure 2b). Maximum decrease in mode utilization is consistent across all classes at 0.5 families per mode per Myr with four exceptions: stages that follow the end-Permian extinction (252 Ma), end-Triassic extinction (200 Ma), end-Cretaceous extinction (65 Ma), and the Grande Coupure (Eocene-Oligocene boundary; 34 Ma) show a significantly larger loss (figure 2b).

Tetrapods have a great ability for adaptation and their taxonomic and ecological diversity have seemingly been shaped over the last 400 Myr primarily by expansion. Initially tetrapods moved into empty ecospace where no other large animal life existed. They filled empty modes of life further away from the water and began to burrow, climb, fly, take advantage of specialized feeding strategies and then continued to invade new habitats evolved by other organisms such as forests, canopies and grasslands. At the same time abiotic processes such as continental breakup, geographical barriers, latitudinal temperature differentiation and changing climate contributed to greater complexity of Earth's surface, creating endemism.

The data show multiple lines of evidence for the role of expansion as the main driver of tetrapod diversification: (i) tetrapods have only explored a third of habitable modes of life; (ii) tetrapods have occupied an exponentially increasing number of modes; (iii) ecological diversification has been driven at an increasing rate by the different tetrapod classes; (iv) successively dominant tetrapod classes have increased the maximum rate of mode utilization; and (v) Tetrapoda exhibit ecological incumbency, observed by a limit at which mode utilization decreased, except at times of mass extinction.

Tetrapods have filled 36 per cent of habitable modes. In contrast, Bambach et al. (2007) found that marine animals have explored 78 per cent of habitable modes, categorized by tiering position, motility level and feeding strategy. This may be because the ocean is in essence a giant Petri dish, whose diversity is saturated and can only diversify further by packing more species into already existing modes or subdividing niches. The terrestrial realm does not appear to have such restrictions or perhaps this limit has not yet been reached.

The records of taxonomic and ecological diversity of tetrapods are closely linked and are good reflections of the expansion of tetrapods. Tetrapod taxonomic and ecological diversity has increased dramatically through time (figure 1) and the high correlation between these two measures is in keeping with observations of the marine realm, where the ecological and taxonomic diversity histories of marine animals are broadly parallel (Bambach et al. 2007).

Tetrapod ecological diversity, like taxonomic diversity, is driven by the four tetrapod classes (figure 2a). Mode utilization by multiple families has risen from a single Devonian amphibious piscivore to four families filling each mode today. During each of the three eras the rate of mode utilization was contained within a range that increased with successively dominant tetrapod classes (figure 2b), since each had a greater ability to share a mode of life among multiple families, doubtless related to their key adaptive features. Palaeozoic amphibians were largely tied to waterside habitats because of breeding constraints, but amniotes exploited a wide variety of new, entirely terrestrial habitats, and the endothermy of mammals and birds allowed these groups to conquer cooler habitats and explore alternative behaviors such as nocturnality. The innovative adaptations of mammals and birds led to a dramatic increase in the rate of mode invasion, tripling the number of occupied modes in 30 Myr.

Although the rate at which tetrapods have expanded mode utilization has increased, the typical limit at which families decreased the use of ecospace is a loss of 0.5 families per mode per Myr with the exception of stages immediately following the end-Permian, end-Triassic, end-Cretaceous extinctions and the Grande Coupure; they show a rate of loss and then addition outside the ‘normal range’ (figure 2b) of the dominant fauna. These four mass extinctions removed incumbent families, and new invading tetrapod groups used their key adaptations to refill ecospace at a greater rate than their predecessors.

Given the unrestricted access tetrapods have to ecospace, perhaps there is little need for competitive interactions to shape diversification. Though traditional views cite inter-clade competition as a driver of evolution (Colbert 1955; Romer 1966) there is little evidence of competition guiding large-scale biotic replacements (Gould & Calloway 1980; Sepkoski 1996). Benton (1996) first indicated that competitive replacement played a minor role in the evolution of tetrapods. He estimated that seven-eighths of tetrapod familial diversification was unrestrained expansion into ecospace rather than the result of biological interaction; the remainder were candidate competitive replacers (CCRs), families that could have originated by competitively displacing another. This study considered ecological, geographical and stratigraphical information, but the resolution of the fossil record does not allow observation of behavioural characteristics (e.g. sleep/wake cycles or reproductive cycles) or spatial partitioning (e.g. animals living at different levels of a forest canopy). If these further partitions of ecospace could be added, the count of CCRs would probably decline further.

Though the tetrapod record does not provide evidence for direct competition, there is evidence of competition in the manner of incumbent replacement, in which established groups can exclude competitors, even if those competitors possess advantageous key adaptations, until the incumbents are removed from their foothold by a major environmental disruption such as a mass extinction, at which time the key adaptations of the invading clade allows them to colonize the area before the incumbents can reestablish themselves (Rosenzweig & McCord 1991). The data support the growing evidence that, except following mass extinctions, tetrapod diversity was primarily achieved by unrestricted expansion into empty ecospace, that is by the filling of unrealized modes of life, and multiplying into already realized modes. As taxonomic diversity has increased, there have been incentives for tetrapods to move into new modes of life, where initially resources may seem unlimited, there are few competitors and possible refuge from danger. And as ecological diversity increases, taxa diversify from their ancestors at a much greater rate among faunas with more superior, innovative or more flexible adaptations.