Richard, I agree that we should think about the large EM segmentation. But I don't think 4.5 cm segmentation is too big, but is (may be) too small. For example, in JLC design, they have chosen 40 mrad, which corresponds 10 cm segmentation. Because the fine segmentation in Large detector will be too expensive. (Instead of the fine segmentation of Cal, they put the fine segmentation pre-shower in front of the Cal.) We know fine segmentation with large radius provides a good performance. But if we need think about the costs, I think we need more discussion about the segmentation for Large. - Masako