The purpose of this archive is to make widely available in digital form the climatic time series which form the basis for the research presented in the publication, 'ON THE STRUCTURE AND ORIGIN OF MAJOR GLACIATION CYCLES. 1. LINEAR RESPONSES TO MILANKO- VITCH FORCING' which appeared in Paleoceanography, Dec, 1992 (Imbrie et al.,1992). Users of the archive should have a copy in hand to identify the files which follow. Imbrie et al.,1992, references the papers where the data was first published. These references contain important information, such as species corrections, etc., which can not be included in this archive. At the bottom of each file the variables in each column are identified. All files contain core identification, in most cases at the start of each record. In some files values of -9999 appear in the data column. This is a SPECMAP convention indicating a missing value. Files where this is the case contain a note at the end of the file. In deriving an age model, Martinson's inverse method ('An Inverse approach to Signal Correlation', Douglas Martinson, William Menke, Paul Stoffa, JGR vol. 87, pp4807 - 4818), was used for cores V19-27, RC13-110 and in tuning DSDP 607. For these cores the depth and age columns in the data file are the age model. For the Chinese loess magnetic susceptibility data, a land based record, an entirely different method of assigning ages was used from that em- ployed for marine cores. No age model is given for this data. The same is true for the Vostok ice core. For the two Vostok records in Figure 7, the chronology is that suggested by correlations of dust (Petit et al.,1990) and air O-18 (Sowers et al., 1991). For core RC13-110, the C-13 record here differs from the one published in that all values based on Uvigerina samples have been eliminated. After assigning ages to each observed core depth, the data is interpolated at evenly spaced time intervals. For most cores this is a linear interpolation based on the two values bounding the in- terpolation time. For others a Gaussian interpolation has been used, a procedure explained in Shackleton and Imbrie, Climate Dynamics, vol. 16, p. 217 - 230, 1990. This method takes a weighted average of all the data within a Gaussian filter width centered at the interpolation time. These time series can be distinguished from linearly interpolated records by the filter width (in kyrs) given at the end of the file. Some records have been patched. The Southern Ocean SST record is a composite of RC11-120 and E49-18. This is true for site 607 where, for both O-18 and SST, V30-97 data covering the past 243 kyrs is patched to data from DSDP cores 607 and 609. The ages assigned to V30-97 are based on the SPECMAP age model given in Archive 1. Equation F13 was used for estimating temperature in all site 607 cores. A gap in AII 107 - GGC 131 was patched with data from AII 107 - GGC 65. For most of the cores, the position of the patch is identified by the core identification at the start of the record. The depths are termed 'adjusted' or 'composite'. Files which appeared in Archive 1 are not repeated here. These files are K708-1 (O-18, SST), V19-30 (delta del 13C), Curry - Crowly stack (delta del 13C). The version of SPECMAP stack in- cluded in Archive 2 is in units of ppm rather than standard deviations. A file erroneously omitted from Archive 1, the plank- tonic foram counts for V25-56 is included here (see Archive 1 for details of formatting, etc.). For several of the time series new age models were developed. The ages assigned in cores RC11-120, E49-18, AII 107, CHN 82, RC13-229, and K-11 differ slightly from the previously published ones. (The ages for RC11-120, E49-18 are as published in Howard and Prell, 1990). Also included are files which are not data. The mid-month insolation for June at 65 deg N (from Berger, 1978a) is given, as well as a listing of thorium ages corresponding to C-14 dates for the interval 5ka to 29.3 ka (thor), based on an equation provided by Edouard Bard of the University of Marseille, France: age(thor) = -0.00585xage(14C)squared + 1.39xage(14C) + 1.807 where the thorium and 14C ages are in ky. For short records, ages are given in both 14C and thorium years. These files are O-18 for PS21 295, SST for MD84 551, Fairbanks' sea level curve (delta del 18O), and the Laurentide ice margin record of Figure 13. All long time series were filtered in the three principal Milankovitch frequency bands. Bandpass filters with Tukey cosine tapers were used. For all three frequencies, the bandwidth was 0.01/ky. The central frequencies were 1/100ky, 1/41ky, and 1/23ky. For those using the ARAND statistical package, the number of triangles was 1 for all three filters. The number and value of the weights making up the filters depend on the interpolation interval of the time series. In this archive, filter weights are given for delta-t's of 2ky, 3ky, and 5ky. A convolution program, such as SMOOTH in the ARAND package, is required to perform the filtering. The files in the sequence they appear on disk (or tape) are as follows: File 1: Description of contents File 2: SPECMAP stack (ppm) vs time 0t>13ka(thor) File 46: RC27-61 dust flux delta t=2ky File 47: Site 607 SST delta t=2ky File 48: Southern Ocean SST delta t=2ky File 49: V19-30 O-18, C-13 delta t=2ky File 50: V19-27 O-18, C-13 delta t=3ky File 51: RC13-110 O-18, C-13 delta t=3ky File 52: P6408-9 % foram fragments delta t=5ky File 53: AII 107 Cd/Ca delta t=5ky File 54: CHN 82 Cd/Ca delta t=1ky File 55: DSDP 552 C-13 delta t=3.4ky File 56: Site 607 C-13 delta t=3.4ky File 57: ODP 677 C-13 delta t=3.4ky File 58: RC13-229 C-13 delta t=3ky File 59: Mid-month insolation June 65 deg N (Berger 1978 solution) 0