 We can think about negative evidence in all sorts of different ways. Very often we emphasize the positives when actually there's a whole series of negatives as it were. We think about the room of error and forget to mention that we didn't find an iron age settlement or an energy settlement or a saxet or whatever it is. We focus on the positives we forget about the negatives very often, although in terrain modeling and in site modeling, such as we see here, we do very often use those residues as inverse residues in order to calculate the probabilities based on our positive evidence. I've just cited two things from what Roger mentioned earlier on. One is the conceptual notion of negative evidence and the other is the question of confidence and just say one or two things about them. The first thing we need to emphasize is that negative and positive are not a binary issue. There's greater complexity to it than simply binary. If we think of positive and negative, we often think of them as binary catapults. But actually what we're doing in archaeology has an empirical science. It means that actually we've got a matrix with four boxes in it rather than simply two. Because over here we've got true and over here we've got false. This is also the case in many other disciplines of course. Simple when you know what you're talking about. You have to come back and switch it on and on. We've got positive and negative and we've got true and false. The problem with a lot of archaeological work as it is in many other disciplines as well is determining which of these four boxes we are in. It's very often the case that we have a false negative in archaeological work. For example Geo-Psychological Survey which doesn't seem to reveal anything when actually there is a lot of stuff there. What we're aiming for of course is to be in these two boxes at the top. Studies have been done on this and archaeology doesn't come out too badly. Compare for example with areas of medicine like antiseptic protection and so on which follow the same basic rules. What we're interested in of course is developing if you like techniques and technologies which move these in this direction and these in this direction. In order to, as it were, heighten what's actually going on and develop our numbers a little bit more. Always therefore we've got to think what is the number that we're using here and how can we relate it to a more matrix based approach to thinking about positive and negatives. What that leads us onto though is to think about confidence and the notion of the confidence they get with it. And as Roger mentioned earlier on a lot of this is very often about the use of different kinds of techniques. Choosing them appropriately either sequentially or in parallel to develop approaches which allow us to be more confident in knowing where we are in this box. And particularly in getting true positives and true negatives out at the end. So as we increase the number of techniques for example and we could put it on the bottom here and we might expect that confidence increases somewhere down here in this general direction there. It's not over a linear arrangement and the selection of techniques becomes very important particularly if you nest those techniques or create them in a sequential way. So that your confidence may actually stay low down here that starts to step up as you increase the techniques. And you can easily think of it in terms of for example a doctor doing an analysis if you walk into a surgery with some sort of ailment and the doctor immediately starts to diagnose what's wrong with you. And there's simple techniques at first, patting your back and using the stethoscope and asking you to cough and looking in your eyes. These are very simple techniques. It's field walking, it's your visit which is the sort of things that we might do in our therapy. Eventually when they can't find out what's wrong with you and stick you inside an MRI scanner then we're talking about some really heavy weight science to find out. And this is what we're talking about in ideology when we start doing some really intensive work to reduce the risk that we're into the false negatives. That's the real difference. So we can think about it in these kind of terms. But then there's more complexity as Rodney noted earlier on. Scale becomes importance. And a concept that Marty Talbot about a few years ago was the concept of legibility which is quite an interesting idea. It involves scale but also the ease with which we can recognize archaeological patterns and archaeological features. So if we continue to have confidence on this scale on legibility as we're on this scale then we can think about how development might allow us to develop a confidence rating as legibility changes and as legibility increases. The important point behind all this though is the point that Roger made at the end of things. It's very bad at articulating the way that we work with our data in terms of telling the audience or the reader what it is we've actually done and what confidence we have as the person who's done it in the results that we're presenting. Have we done these sort of analysis? Are we confident that it's a true positive or a true negative? Or are we just hedging our bets as it were in the background to think about some of this stuff? The articulation of the confidence we have in our own approaches is just as important for positive as it is for negative evidence. Now in that sense positive and negative evidence become part of a continuum of understanding social spaces and things that we're talking about in the papers today in a way that perhaps allows us to be much more nuanced about understanding the archaeological record. So those are just a couple of things which I hope stand on some of the things that Roger introduced us to this morning first thing but we didn't have much time to develop and elaborate back then. But it brings us I hope more or less back on schedule as it were to move forward with the programme.