Supplemental Material 7. Methods 7.1. Sclerosponge Collection and Growth Models For this study, two specimens of C. nicholsoni were collected from Exuma Sound and Tongue of the Ocean (TOTO), and analyzed for high-resolution Sr/Ca and „18O data. In order to construct age models, three U/Th dates from the Exuma Sound sclerosponge were used to calculate an average growth rate and Sr/Ca variations were used to count annual temperature cycles in the TOTO sclerosponge (Swart et al., 2002). Although growth rates in sclerosponges have been shown to vary (Willenz and Hartman, 1999; Lazareth et al., 2000; Swart et al., 2002; Haase-Schramm et al., 2003), the constant growth rate assumption for the Exuma sclerosponges provides a good estimate of the average growth rate over the time period of a century, as it did for the TOTO sclerosponge (Swart et al., 2002). 7.2. Sr/Ca measurements Sections of sclerosponge skeleton cut parallel to the growth axis with a thickness of ~5mm directly from thicker sections of sclerosponge prepared for micromilling (isotope analysis). These thin slabs represented mirror images of the slabs used for isotope analysis. They were scrubbed with a nylon brush in Milli-Q water (Millipore Water Systems), sonified for 5 minutes in Milli-Q water, and then triple rinsed with ultrapure 1% HNO3 and Milli-Q water sequentially. Each laser sample was taken as a series of spots, 1.4 mm in length, perpendicular to the growth axis and separated by approximately 20æm (Exuma) or 10æm (TOTO). The spot size was identical to the sample resolution to avoid aliasing from discontinuous sampling. This sampling resolution represents average temporal resolutions of 8 and 14 samples per year for Exuma and TOTO, respectively. A helium gas stream was used to transport the vaporized material to the ICP-MS. The carrier gas stream was mixed with a wet aerosol (1% w/w HNO3) from a self-aspirating PFA nebulizer (20 lúmin-1) in a dual inlet quartz spray chamber. An aragonite reference material from fish otoliths was used as a standard and measured at intervals throughout the analyses (Yoshinaga et al., 2000). Analyses of the aragonite reference material (n = 153) run during assays of the Bahamas sclerosponges provided a mean value of 2.878 ñ 0.008 mmolúmol-1 (1s), after correcting for mass bias using an internal laboratory standard. 7.3. Oxygen isotope measurements All isotope samples were milled in continuous transects. Samples were drilled from the Exuma Sound sclerosponge at a resolution of 50 mm (nearly 3 samples per year if growth rate is assumed constant) and were analyzed by a Kiel III automated carbonate device attached to a Delta Plus isotope ratio mass spectrometer (IRMS). The d18O values were calculated relative to VPDB against an internal laboratory standard with a standard error of 0.06 % (n = 46). Samples from the TOTO sclerosponge were drilled at a resolution of 34 mm, corresponding to an average temporal resolution of 5 samples per year, and analyzed via a Fairbanks common acid bath automated carbonate device attached to a Finnigan MAT 251 IRMS. Oxygen isotope values from these analyses were calculated relative to VPDB against an internal laboratory standard with a precision of 0.09 % (n = 48). 7.4. U/Th Measurements Approximately 0.5 g of sclerosponge aragonite was drilled along the visible growth bands of each dated sclerosponge and dissolved in warm HNO3 concentrate. The samples were each spiked with 229Th and 233U-236U. Thorium and U were precipitated with Fe(OH)2 in alkaline solution using NH4OH and the U and Th were then separated with HBr and HCl on an anion exchange column of BioRad AG 1X8 resin (100-200 mesh). Blanks showed 14-30 pg Th and 13pg U. Uranium and Th isotope analyses were carried out on a Finnigan MAT 262 RPQ + mass spectrometer at GEOMAR, Kiel, Germany. Thorium and U were loaded with 1N HNO3 on an out-gassed Re filament and measured using the double filament technique in the "peak jumping" mode. Measurments of the international standard A112 yielded a d234U of -30.3 +/-6.3%. See supplemental table for a summary of the results. 8. Error Propagation in Salinity Reconstructions The significant errors propagated through each of the measurement techniques and calibrations into equation 4 illustrate the sensitivity of salinity reconstructions to both the salinity - „18Osw and temperature - Sr/Caarag relationships. We have used salinity - „18Osw data from the APNAP cruises between 1986 and 1988 (Ganssen and Kroon, 2000). The residual error in this empirical linear relationship is magnified by the gradient in eq. 4 and is responsible for 43 percent of the error in our salinity reconstruction. The other significant source of error is the residual error of 0.7 degrees C in the in situ Sr/Ca - temperature relationship of Rosenheim et al. (Rosenheim et al., 2004), however it reduces the error calculated by seasonal extrema calibrations and decreases the total amount of error generally seen in this type salinity reconstruction (Schmidt, 1999). 9. Supplemental References 8. Supplemental References Ganssen, G.M. and D. Kroon, The isotopic signature of planktonic foraminifera from NE Atlantic surface sediments: implications for the reconstruction of past oceanic conditions. Journal of the Geological Society, London 157, 693-699, 2000 Haase-Schramm, A., F. B”hm, A. Eisenhauer, W.-C. Dullo, M.M. Joachimski, B. Hansen and J. Reitner, Sr/Ca ratios and oxygen isotopes from sclerosponges: temperature history of the Caribbean mixed layer and thermocline during hte Little Ice Age. Paleoceanography 18, doi: 10.1029/2002PA000830, 2003 Lazareth, C.E., P. Willenz, J. Navez, E. Keppens, F. Dehairs and L. Andre, Sclerosponges as a new potential recorder of environmental changes; lead in Ceratoporella nicholsoni. Geology 28, 515-518, 2000 Rosenheim, B.E., P.K. Swart, S.R. Thorrold, P. Willenz, L. Berry and C. Latcokzy, High resolution Sr/Ca records in sclerosponges calibrated to temperature in situ. Geology 32, 145-148, 2004 Schmidt, G.A., Error analysis of paleosalinity calculations. Paleoceanography 14, 422-429, 1999 Swart, P.K., S.R. Thorrold, B.E. Rosenheim, A. Eisenhauer, C.G.A. Harrison, M. Grammer and C. Latkosczy, Intra- annual variation in the stable oxygen and carbon and trace element composition of sclerosponges. Paleoceanography 17, 1045, 2002 Willenz, P. and W.D. Hartman, Growth and regeneration rates of the calcareous skeleton of the Caribbean coralline sponge Ceratoporella nicholsoni: A long term survey. Proceedings of the 5th International Sponge Symposium "Origin and Outlook": Queensland Museum Memoirs, J.N.A. Hooper, ed. v. 44, 675-685, 1999 Yoshinaga, J., M. Morita and J.S. Edmonds, Fish otolith reference material for quality assurance for chemical analyses. Marine Chemistry 69, 91-97, 2000