Ann Gibbons reports on a recent conference investigating the interaction of climate change and Plio-Pleistocene human evolution "Where's the beef? Early humans took it." I like her description of Lars Werdelin's work:
After comparing fossils of 78 species of carnivores that lived during five different periods of time between 3.5 million years ago (when large carnivores were at their peak) and 1.5 million years ago, Werdelin found that all but six of 29 species of large carnivores (animals that weighed more than 21.5 kilos) had gone extinct in that time. Moreover, the mass extinction began just before H. erectus appeared in the fossil record 1.9 million years ago. He also found that the community of carnivores alive 2.5 million to 2 million years ago ate a much broader range of food—with species within a community filling a wider range of dietary niches. By 1.5 million years ago, just hypercarnivores that ate only meat, such as lions and leopards, had survived while omnivores that scavenged and ate a wider range of foods, like civets, had disappeared. "Even I was surprised by the dramatic drop," Werdelin says.
It will be interesting to see more details of this work as it is published. The picture described here seems fairly different from Werdelin's 2005 paper with Margaret Lewis , in which they concluded there was no evidence of a turnover pulse of carnivore species between 3 million and 2 million years ago. In that paper, they did note that the present carnivore community is dominated by species that have no Pliocene record in Africa. It's not clear how much of this turnover resulted from dispersals of carnivores into Africa from Eurasia, and how much was in situ origins of new genera. That paper was focused on the role of herbivore species turnover on the carnivore community, showing that the two records do not match each other.
With the present focus on hominins as potential competitors, maybe the expansion in scope to a greater number of omnivores made the difference to the analysis. On the other hand, it's hard to see how recent carnivores can be purer meat-eaters than extinct sabretooths like Megantereon. Lewis and Werdelin's 2007 book chapter  (in a Springer volume with no access for me, naturally) does link the origin of Homo erectus 1.8 million years ago to carnivore turnover.
While the appearance of stone tools at 2.6 Ma has no apparent effect upon carnivorans, the appearance of Homo ergaster after 1.8 Ma may have been at least partly responsible for the decrease in the carnivoran origination rate and the increase in the extinction rate at this time. The behavior of H. ergaster, climate change, and concomitant changes in prey species richness may have caused carnivoran species richness to drop precipitously after 1.5 Ma. In this situation, even effective kleptoparasitism by H. ergaster may have been enough to drive local populations of carnivorans that overlapped with hominins in dietary resources to extinction. Possibly as a result, the modern guild, which evolved within the last few hundred thousand years, is composed primarily of generalists.
We should probably add to the picture the evidence for dispersal into and out of Africa, which is unclear at the moment for carnivores . If humans had an important effect on the carnivore guild, we should expect codispersal of some carnivore genera with humans in the Early Pleistocene. It's possible that such codispersal did occur into Eurasia, but first appearance dates are not the greatest evidence to build such a hypothesis.
On that note, there are some incredible carnivore materials from Malapa that may really add to the picture of carnivore-hominin relations. The first of these were published last fall by Brian Kuhn and colleagues, including Werdelin . It will be exciting to see more of this work come out, as I'm sure that the preservation of a wide array of carnivore materials is really shifting how we can think about the relative diets and ecological roles of these species. It's another case where paleontologists can now leverage the vast record of time covered in East Africa by applying the detailed information from the exceptionally preserved Malapa deposit.
UPDATE (2012-04-25): Adam Van Arsdale writes that Dmanisi provides even more evidence about carnivore-human interactions: "Early Homo and the carnivore guild".
The Dmanisi fauna in general, including the carnivores, are only just beginning to be more widely published. A 2011 paper by Hemmer and colleagues discusses a possible large cheetah-like carnivore found at the site. This 2010 paper by Sotnikova and Rook looks at Canid evolution in Eurasia more broadly, but discusses the abundant Canid material from Dmanisi in some depth.
Kate Wong gives some more information about Werdelin's presentation: "Rise of Humans 2 Million Years Ago Doomed Large Carnivores".
- . Plio-Pleistocene Carnivora of eastern Africa: species richness and turnover patterns. Zoological Journal of the Linnean Society. 2005;144(2):121 - 144.
- . Patterns of Change in the Plio-Pleistocene Carnivorans of Eastern Africa: Implications for Hominin Evolution. In: Hominin Environments in the East African Pliocene: An Assessment of the Faunal Evidence. Hominin Environments in the East African Pliocene: An Assessment of the Faunal Evidence. Dordrecht: Springer Netherlands; 2007. pp. 77 - 105.
- . Hominins without fellow travellers? First appearances and inferred dispersals of Afro-Eurasian large-mammals in the Plio-Pleistocene. Quaternary Science Reviews. 2011;30(11-12):1343 - 1352.
- . Carnivoran remains from the Malapa hominin site, South Africa. PloS one. 2011;6(11):e26940.