The chemical characteristics of aggregating material in the marine environment are largely unknown. We investigated neutral aldose (NA) abundance and composition in aggregation of marine snow and other organic matter (OM) size fractions in the field. Four sample sets were fractionated using membrane filtration and ultrafiltration into the following size fractions: particulate material, high-molecular-weight (HMW) material, and low-molecular-weight (LMW) material. We also collected three sample sets of marine-snow aggregates. Each sample set contained small, medium, and large aggregate size fractions and each size fraction consisted of 25-50 aggregates. For 7 marine-snow samples and for each water-sample size fraction, we determined monomeric and polymeric NA concentration, NA yield (amount of NA-C normalized to organic carbon), and composition; total organic carbon (TOC) concentration; transparent exopolymer particles (TEP) concentration, and TEP propensity (TEP concentration after inducing TEP formation in filtered samples). This is the first study to include compound-specific NA determinations on these four marine OM size fractions. The mass balances of organic carbon and NA indicated that there were no serious contamination or loss problems. Concentrations, yields, and NA mol fractions in water samples were similar to results from other studies. Glucose and galactose had the highest relative abundance in all size fractions. The NA yield increased with increasing molecular weight or particle size for all fractions except marine snow. The NA yield increased in the order: LMW<marine snow< HMW< particles. Marine snow had a higher average NA yield than the LMW fraction, but lower than particle and HNW-fractions. This indicates that OM in marine snow could have been diagenetically derived from particulate and HMW-fractions, that is, marine snow may include material from the particulate and the colloidal phase. TEP concentration or TEP propensity was positively correlated with concentrations of all individual NAs as well as the sum NA concentrations, indicating that TEP contains neutral sugars in addition to the acidic polysaccharides stained in the determination of TEP concentrations. Despite the relatively low NA yield in marine snow, marine snow was enriched in NA when compared with seawater, with enrichment factors of 34-225 (average 125). By combining data from this study with data from other studies, we estimate that < 10% of carbohydrates in marine snow comprise NAs. There was no clear correlation between marine-snow aggregate size and NA yield, that is, there appears to be no general age difference between small and large marine-snow aggregates. NA composition was similar among different marine-snow size fracionscollected during the same day, indicating that aggregation/disaggregation reactions resulted in homogenizing NA composition in marine-snow aggregates of all sizes. The NA composition of marine snow was different from that of other OM size fractions, indicating either that bacterial degradation has modified the composition of marine snow to a larger extent than other OM size fractions or that marine snow is formed through the aggregation of selected subcomponents of OM. (C) 2007 Elsevier B.V. All rights reserved.