Main

But this model lacks a detailed analysis of extinction/immigration rates in the framework of precise palaeomagnetic and biostratigraphical data. In this context, a comparison of the late Miocene-early Pliocene records of artiodactyl and rodent communities from Spain and Pakistan is of special interest. First, the amount of palaeomagnetic and biostratigraphical data from both areas2,3,4,5,6,7,8 provides a high resolution of the shifts in faunal communities, and second, according to Cerling et al.1, the C3/C4 transition took place all over the world except in western Europe.

Our comparison of extinction and immigration rates from sites in Spain and Pakistan (Fig. 1) reveals that first, in Pakistan's Siwalik sediments, not one but two faunal changes occurred, affecting especially artiodactyls. The first one, at about 10-9 Myr ago, largely preceded the C3/C4 transition (8-6 Myr ago); the second one occurred at its end (6.5 Myr). However, none of them exactly coincides with the initial expansion of C4 biomass (8-7 Myr ago), despite a slight increase in rodent extinctions.

Figure 1: Faunal turnovers in the Eurasian Upper Miocene epoch under C3 (Spain) and C4 (Pakistan) conditions.
figure 1

Dotted lines, appearance of genera; continuous lines, disappearance of genera. Hp, hippopotamids; Ho, hominoids; T, tragulids. Data on Pakistan's Siwalik sediments derive from ref. 8; the palaeomagnetic scale is from ref. 9. Considered are herbivore communities with frequent changes at the generic level (artiodactyls and rodents); excluded are groups with low generic diversity such as proboscideans, rhinos and equids. Between 12 and 5.5 Myr ago, two faunal turnovers took place, the older one between 10 and 9 Myr ago and the younger one between 7 and 6 Myr ago. 60% of all faunal changes (immigrations/extinctions) that happened between 12 and 5.5 Myr ago occurred in these times of faunal turnover, when the rate of extinction/immigration increased up to four times the ‘normal’ value. With the exception of the extinction pattern of Siwalik rodents in the late Miocene epoch, there is a high correlation between the faunal turnover maxima of both the Spanish and the Siwalik herbivore communities, despite the transition from C3 to C4 plants at 8 Myr ago in Pakistan.

Second, both faunal changes in the Siwaliks are contemporary with faunal turnovers in Spain, although in western Europe there is no indication of C4 diet at any time. In Spain, these faunal changes turn out to have been even more dramatic than in the Siwaliks as the extinction/immigration rate is twice as high. In Spain, woodland or forest indicators (hominoids and tragulids) became extinct about 9 Myr ago, whereas in the Siwaliks hominoids survived their European counterparts by about 1 Myr, and tragulids by as much as 5 Myr.

In the light of these data, we cannot find synchronicity or even a causal link between faunal and vegetation change. Faunal turnovers of different magnitudes took place before and after the appearance of C4 plants, suggesting that C3/C4 transition and faunal turnovers occurred independently.

The “important global ecological change” instead looks like a succession of faunal and vegetation changes, scattered over a long period of at least 5 Myr. We feel that there is insufficient information to permit a clear interpretation of these phenomena. They probably occurred under the influence of a variety of events such as alpine orogenesis with Himalayan and Tibetan uplift, monsoonal dynamics, increasing seasonality, and other factors not yet fully investigated.

Reply - Cerling et al.