Thank you.
Actually, I want to continue this line, in part to follow up on some of the questions Mr. Dion had.
You're sampling at a national level, and as we read, whenever we take a sample and we start cutting it down, it becomes less and less representative. So in fact it's not unusual for us to use two methods.
The national sample that you're taking just wouldn't work within my province of New Brunswick. It begins to look less certain, and hence we turn to the actual census numbers when we're dividing up boundaries. Is that right?