I mentioned an important feature earlier; that is that if you want to influence a decision, it's important that the information be available when the decision is being taken and that the information be actually relevant to the decision. That requires, in our experience, a greater degree of priority-setting and of influencing of scientific priorities than exist external to government currently.
There is a certain type of science that's best done through granting councils, in curiosity-driven research. However, for the type of science that Environment Canada needs for at least part of its business, which is the risk management/risk assessment that we do in partnership with Health Canada, it's a very risky strategy to hope that someone will become curious about an issue and produce the information you require in time for a public decision. We believe that needs to have a strong in-house component, and that's why we do it in-house.