Rate and delay of extinctions

Here's an interesting paper: it suggests that major catastrophes need not always lead to immediate extinction pulses, but that there can be a lag of as much as 2 million years (in the case of the rise of the Panama isthmus).

In hindsight this is not so surprising. Ecosystems generate a lot of their own resources as by products of autotrophic organisms (basically, photosynthesisers and lithotrophs), and so as long as there is a sufficient influx of energy into the system, an ecosystem might be able to persist for some time, buffered by the productivity of the ecosystem itself, if it is rich enough. Ecosystems are not directly correlated with external factors, and are hard to perturb sometimes. But eventually a crisis point will be reached and a rapid collapse will follow if conditions don't ameliorate.