 or evening and thanks to the organizers for bringing us here. What we're going to present is a way to approach mobility from two different strategies. Always trying to have the little industry as the principal proxy. Our technological context for this approach is based on the eastern Mediterranean peninsula, basically on the region of Valencia and of the low Aragon region, and it's based on the late Mesolithic period which goes from 7700 to 6700 VP, pretty much of course, and our main objectives are to approach how hunter-gatherers moved, and how can that be expressed by the early record in two methodologies which will be on one hand how curiously or can express mobility or how diversity in archaeological assemblages can express mobility. We did a preliminary work on this last year in Maastricht where we let's say we will lift it at some points, and now we try to bring that work farther. At Maastricht we were more focused in Gothina Cave, we'll see a map just in the next figure, we're more focused on Gothina Cave, and we also noticed that we would find different results comparing older excavations to modern excavations. We also found some possible biases that could be happening in our methodology, and we tried to look for some possible measure which was not relative to our sample, like a fixed measure of the residential versus logistical crimes. So this time what we did is we had a global focus, we adapted our methodology to a reviewed sample, we assessed the causes for these biases, and we also concentrated on our specific context. This meaning that we, one of the things we had to do was to assess the differences between old and modern excavations, and also we wanted to assess how geographic geological variability could affect our sample and results. One of the first, well this is the geographical frame, and one of the first problems that we have is that we needed some data, and that data was not always accessible, that as a matter of fact the original plan was to study 28 sites, and we had to reduce that to nine sites because for all of the other sites we didn't have enough data to study them in the standards that we needed, and this is because of many reasons. In one hand some sites have not even been excavated, they just published but some surface collections let's say, some of the sites have been excavated but not even published. Some sites have been excavated but the record is not enough, and sometimes the information and the publications doesn't have all of the conditions that we need. Talking about what do we need in order to do this, we need a lithic industry, as by class, the volume of the sites, and also we wanted to have the proposed function to see how all results were compared with the function proposed by the original excavators of the site, so you can see between these two things. So this is just a statistical try of the first of the 16 sites where we see that a liar, which will be the site of masquerade map which could not be included in the accuracy versus expediency methodology because as we say it was an uplier and it was, well we could just not include this site, but it could be included in the other analysis and we will have an explanation of why this could be there and we'll go to that. So first we assessed this issue by the accuracy versus expediency. The rationale, which Julien explained better than I could before, the rationale is that due to the difficulties of transporting big amounts of raw material and function-specific camps are expected to present more curated assemblages, this implying that higher research frequencies regarding the total volumetric density. Being more specific, shorter camps are expected to present more curated assemblages, whereas longer camps are represented by more expedient assemblages. In order to perform this methodology, we were based on some older papers, like from Clark and Baton to 2017, where a linear model on locked-in scale was established and where we would compare the retouch frequencies against the lithic volumetric density. The position of the sites and levels in the model is expected to show its relative degree of accuracy and therefore allowing us to explore their mobility roles. So this is the first comparison that we did. As we can see, that correlation does exist. We have a significant correlation, an inverse correlation, but if we look at the sites, we can have some inconsistencies regarding previous knowledge. I don't know if you can actually see pretty well the names of the sites that we have here, for example. According to this, it's supposed to be a long-run camp, and instead we know because how the site is and where it is that it's difficult and actually was a long-run camp. Also, we have here a banana, an open-air, a fairly big camp, which we were not expected to be there. We believe that the reasons why this is happening is probably having to do with the way in which we gather the volume information. Because we are consulting bibliography, we don't work, but then we have managed all these sites. And sometimes, the volume, we just have the area and the average depth, for example. And that's not reliable for... We know that the depth is not always uniform. So we believe that can be one of the problems that we have. Also, we also knew that there were differences between all and new excavations. In this case, we decided to take in the year 2000 as a turning point and comparing older versus newer excavations. We see that correlations increase in fact when we compare all sites between them and new sites between them, although we do not have statistical significance for new sites. And our sample is small, but this could be hinting that actually there is a difference for some reason with this. In the beginning, we thought that this difference was because of the chips of the lithic industry. We thought that it would not be picked up in the oldest excavations. So we tried to just eliminate it from the sample. But our correlations, well, it says that they increase, but actually that's a mistake. They do not. They decrease. But our correlations decrease when we remove the chips. So we believe that's not the problem, probably. So the bias factor maybe is due to some on-site selection. Maybe it's due to some smaller accuracy because there are not special analysis on older sites. So we are not sure now of why this is happening, but we know that there is a bias between modern excavations and older excavations. So other thing that we wanted to compare was because of the nature of the geometric microliths. Because how it is a function-specific item is less affected by reduction. And we know that it is, for the mesolithic period, it can be a hunting proxy. We wanted to compare it against the lithic volumetric density. In this case, we see that correlation increases too. So this is strengthening this approach. Then we tried to do further comparisons with the geometrics. Because we know that since they are produced after doing blades, we thought that they should have to be an inverse correlation between geometrics and blades. But that correlation is not getting significant results, and it's not being as we expected. So we decided to do a new correlation between the geometrics and the blades, not between the volume, but just between the geometrics and the blades. And we see that positive correlation, although with no significant value either. So we probably believe that we have to increase our sample in order to have higher significant values in order to be able to compare this. So for this method, we have seen that global correlations are indeed obtained. We see that some inconsistency can appear when we go to the specificity of the sites. And we believe that this can be because possible on accuracy of volumes and lack of control of deposition rates. We believe that this system can achieve its optimal performance when these two factors are controlled. And then we can have a real a better approach on this. So as for the other method that we tried, the diversity and expression of functionality. The rationale is because a wide range of activities is supposed to be developed in long term camps regarding non specific short term camps. A high limited diversity is expected. This way, we suppose that multifunctional camps would indicate long term camps. And we applied two diversity indexes, the Shannon Weaver and the Gene Simpson. This because the we know the problems of the Shannon Weaver with the quality of the sample. So we tried to include the Gene Simpson index, which weights the sample to our predominant classes. So because we knew our sample could have some problems regarding the timing of our excavations, that's why we decided to imply Simpson index. Then we applied also the Jackett index to see how the sites really relate between themselves, geographically and chronologically. We didn't use types for this, but we use classes because this is a way one in one hand is a way of circumventing the typology debate. And in the other hand, classes are usually supposed to have similar functions. So we are looking for functionality. So we see that the results here, we see how they do not differ significantly between the different tests. And we saw the clusters here, they do belong to site or shelters and they do belong to sites and signal biotopes. And it's consistent with prior knowledge. Cozina is also in the same range and it's a very specific case. And also Benamera and Falgera would need, Falgera is, according to these results, Falgera is a hunting spot. And Benamera is an opener site. And I think we should also further study that. Then as for the Jackett results, we see two clusters, which imply basically all of Cozina, Benamera and Boticería 2. And then we see another cluster, which implies many of the shelters we saw before in the diversity measures analysis. Then the third cluster is just for the same site, for different levels of the same site. So wanting to assess if there was regional and chronological variability, we see if we check the HB, that's the Xamarin, the Simpson, the Lutus, if we check them, we see that there could be geographical variability. But then, if we analyze it further, then we know that, as we said before, this site, what it was that we showed is singular biotopes. And this site, both Cozina is a site with a very different collection with more than 2,000 geometrics, maybe not comparable to the other sites. And this site Benamera is an opener site, the only one that we have. So those are two very specific sites. Then if we see the data from the DACA clusters, we see that these sites do relate with Boticería 2 up here. So that could be indicating that there is, at this point, there wouldn't be geographical or chronological similarity. So, well, about this system, these measures are broadly consistent with prior knowledge. And however, we know that diversity measures can be expressing things other than functionality. It can be expressed in style, it can be expressed in different things. So we must be aware of this and apply further studies on that. Finally, one of the important things is that the data or disposal may be insufficient to perform optimal analysis. And that we do not have similarities of sites, they don't depend on geographical or chronological distributions. And that we are hinting to a logistical mobility model, where we have some camps and some other spot specific camps. We need more sample, we need a bigger sample. And we definitely need more open air sites, just to see how Benamera was behaving and what was happening there. Also diversity measures could be used as that threshold we spoke about before. But we need them to be, we need a bigger sample for that. Finally, about the methods, the Curriciverse Experiency methods, they offer consistent relations, but it's not easy to obtain a reliable sample with information on our disposal. About similarity, they offer consistent measures. But we must take into account different interpretations. That's why we should combine both. About the sample we said before, we need a larger reliable sample. And about mobility, we are speaking probably about logistical mobility patterns. And about chronological and geographical variability. We do not think they have an influence on our results in this case. And about mass reach, well, the things that we have done is implementing more sites, proposing a mobility measure, and bringing new lines of research and methodology to the issue. And thank you very much.