at 0°, we obtain for this sample S= 1.129M? at 4.21°K, while the 2D hard-disk gas calculation yields S = 0.969Nk. The b for which the formula agrees with the experimental entropy is 8.8 A 2 , which lies well within its range of uncertainty. The entropy comparison indicates that the assumptions of a 2D gas state at 4° and of essentially complete order at low temperatures are correct. The transition at x g = i thus appears to involve a change from a 2D gaslike state at high temperatures to an ordered array at low temperatures; and since He 4 at "critical coverage" follows the logarithmic dependence of latticegas theory, the extinction of thermal mobility must in some sense be proportional to the degree of spatial order. If mobility is so intimately related to the spatial order, it is a collective property of the system and cannot be treated in terms of noninteracting single-particle Bloch states. The transitions therefore bear some resemblance to the metal-insulator transition of certain three-dimensional electronic systems. 11 We wish to thank C. E. Campbell, M. Schick, and P. R. Zilsel for continuing stimulating conversations and the free exchange of ideas. Their