This directory contains: 1. The 11 distinct matrices of proxy data (plain text format) corresponding to the 11 distinct intervals of the stepwise reconstruction described by MBH98. These files have names of the form dataXXXX.dat where XXXX is the start year of the reconstruction interval. 2. The 11 corresponding files listing the corresponding set of proxy indicators, and the weights applied to those proxy indicators in the analysis. These files have names of the form datalistXXXX.dat where XXXX is the start year of the reconstruction interval. 3. Data matrix (plain text format) containing all 415 individual proxy (and instrumental/historical) series contributing to the proxy indicators networks used by MBH98 (filename is dataall.dat) and a listing of the proxies (datalistall.dat) Note that PCs of the various proxy sub-networks are computed independently based on the series available over each independent sub-interval of the stepwise reconstruction (i.e., 1400-1980, 1450-1980, 1500-1980, 1600-1980, 1700-1980, 1730-1980, 1750-1980, 1760-1980, 1780-1980, 1800-1980, 1820-1980). A decreasing number of PCs of these sub-networks are retained increasingly further back in time, as dictated by application of objective selection criteria (see METHODS subdirectory). The total number of indicators used by MBH98 in the various intervals of the stepwise reconstruction are: back to 1820: 112 indicators back to 1800: 102 indicators back to 1780: 97 indicators back to 1760: 93 indicators back to 1750: 89 indicators back to 1730: 79 indicators back to 1700: 74 indicators back to 1600: 57 indicators back to 1500: 28 indicators back to 1450: 25 indicators back to 1400: 22 indicators 4. Updated list (mbh98datasummary.txt) of all individual series used by MBH98, with start/end dates of series, and references. A few additional notes: a. As noted by MBH98, a modest fraction (about 25%) of the proxy series used by by MBH98 terminate before the 1980 upper end of the calibration interval (between 1972 and 1980-see file #4 above). To complete the network through the 1980 calibration period end, such series were, as noted in the MBH98 supplementary information extended by MBH98 to the 1980 boundary by persistence of the final available value. To avoid confusion, such infilled values have instead been flagged by a missing data indicator ("NaN") in the files provided here. The MBH98 reconstruction is not sensitive to the use of these infilled series, as a nearly identical reconstruction is achieved if a 1902-1971 calibration (pre-dating any infilled indicators) is used rather than the 1902-1980 calibration period. In addition to single series used, this applies to PC series of certain proxy sub-networks: i. ITRDB AUSTRALIA/NZ: 15/15 series end before 1980 (terminate between 1974 and 1976) ii. ITRDB SOUTH AMERICA: 11/11 series end before 1980 (terminate between 1974 and 1976) iii.ITRDB N. AMERICA: 55/251 series end before 1980 (terminate between 1971 and 1980) iv. STAHLE ET AL SW TEX/MEX : 6/22 end before 1980 (terminate between 1978 and 1980) v. VAGANOV NETWORK: 1/60 series ends before 1980 (1977) Consistent with the above protocol, the PC series for (ii) and (iii) are flagged as missing after 1974. For simplicity. PCs of the other networks are provided through 1980, based on infilling by persistence of the modest subset of constitutent series termination prior to AD 1980. b. There were minor typos in the listing of proxy data used in the original supplementary information, including the listings of some series that were not used, some incorrect start dates, and a few incorrect references. These have been fixed in the revised listing (file #4) provided above. c. Two series (Western Tree Ring density and ring width series) used from Bradley and Jones (1993) were set equal over the interval 1962-1980 owing to missing data after 1961 in one of the two series. The MBH98 reconstruction is not sensitive in any way to this infilling; a nearly identical reconstruction is achieved without the use of either series.