Well we continue to operate crews that have a low channel count that are doing some 2D work, particularly back east. So you could have a 2D crew that has 1000, 1500, 2000 channels on it, you can move into the Barnett Shale region, for example, and you may continue to operate with 4000 or 5000 channels. As you get into some of the Haynesville stuff and some of the Eagle Ford and the Niobrara, you can get into some crew sizes that'll be 8000, 10,000, 11,000 channels. I think the largest one we're going to have operating will be about 11,000. So part of it is the desire to get the wide-azimuth, richer-azimuth, richer offset, distributed surveys for more analytical work on the shale plays. I think that certainly has an impact in it. The other thing that impacts channel count, Collin, can be just the size of the survey. You may not necessarily utilize more active channels from a technical imaging standpoint, but you may require more channels to efficiently acquire the survey and to actually operate in a cost-effective, efficient manner. So you've got several things driving that. Certainly, the geophysical imaging standpoint is certainly a driver, particularly in some of these shale plays that we discussed. But you've also got operational issues that are related to just the vast size of the survey itself. It has a lot to do with the shape and size of the survey. From an average standpoint, there's really not an average. But if you just pulled a number out of the air, it's like asking what's the average size of a survey. That's a tough question. But if you were using 110,000 channels in 2008 over 16 crews, you're using now 130,000 channels over 12 crews, let's say. So from a channel-count standpoint, I think the industry's using more capacity from a channel-count standpoint than we are from a crew-count standpoint. So if you just draw some numbers out, I don't know if the average is 4000 or 5000, 6000 up to 7000, 8000, 9000. But it's somewhere, 50% increase, something like that on an average basis.