Dissolution was high throughout the profile—in most cases, only 10–20% of the valves considered as pristine (Fig. 4). High relative percentages of A. baicalensis and of benthic Fragilaria (sensu lato) were positively correlated (r=0.29 and r=0.26, respectively) to samples with large proportions of pristine valves. There were no significant relationships between this index of dissolution and any of the other dominant taxa. Fragmentation of the valves was also high. On average only 45% of the total count was represented by whole valves (Fig. 4) and high percentages of non-fragmented valves were positively correlated with the percentages of pristine valves. The percentages of pristine valves showed high variation (20–70%) and by contrast with the dissolution index, they were significantly correlated with the variations in percentages of many dominant taxa.
All diatoms in the analysed section were extensively affected by dissolution with only c. 1% of valves in a “pristine” condition. Diatom concentrations were generally extremely low throughout MIS 3 and across the MIS 3/2 transition with samples containing a mixture of extant and extinct species (Fig. 4).
S. grandis is by far the species that contributes the most to the total biovolume accumulation rate with a peak value just above 4×106 μm3 cm−2 year−1. Its relative contribution is over 50% in most samples (Fig. 5). Only taxa such as S. carconeiformis, C. ornata and the vegetative cells of A. skvortzowii contribute significantly (with maximum around 0.6×106 μm3 cm−2 year−1). Other taxa, despite having large relative percentages in parts of the interglacial contribute little to the total BVAR. Biovolumes for benthic taxa were not calculated but their contribution to the total biovolume accumulation rate can be considered negligible as most of these taxa are of small size and the intervals in which they dominate in relative percentages correspond to low diatom valve accumulation rate.
The diatom succession at Academician Ridge is similar to the one from Continent Ridge and the two records, despite having very different sampling resolution, can be easily compared on the basis of their main floristic changes (Fig. 6). BVAR at Academician Ridge is about half that of Continent Ridge. There are also marked differences in the relative abundance of some taxa. At Academician Ridge, S. grandis and A. baicalensis are more abundant and S. formosus and C. sp. cf. operculata are less abundant than at Continent Ridge. The large peak in vegetative cells of A. skvortzowii found at Continent Ridge (DAZ 2) is absent at Academician Ridge, and is the most striking difference between the two records.