Well, it probably depends on whether they expect the stuff to melt and fill in the gaps.
When you take one large roughly flat surface and put it against another and apply heat, it is easy to end up with a trapped air bubble. I've taken dozens of heatsinks off to discover large portions of the surfaces were not in contact with the proc.
IF, and this is a big IF, they designed it right, the missing areas could possibly allow air to escape before the melted stuff filled in the holes... They use a similar approach when setting some kinds of tiles.
Of course, I might be giving someone a little too much credit really....
IMHO... people put way too much effort into applying thermal grease. A few air bubles are not going to cause your system to crash. Hell.... the average user could run conroe with NO thermal paste and I doubt they would nitice a difference. Trust me on that one... I've done it. I often use a big heat sink (no fan) for testing and there isn't any thermal past on it (gets too messy when testing multiple systems). Squirt a glob of thermal paste on the middle of the IHS and let the pressure of the HSF take care of the rest. Once the grease heats up it'll fill any gaps.