Of the various mechanisms suggested (Donovan & Jones, 1979), that most likely to be responsible for the eustatic sealevel changes in the early Silurian is variation in the amount of water incorporated in land-based ice-sheets. This conclusion is based first on the apparent rapidity of many of the sea-level changes (see, e.g. Loydell, 1994) and secondly on the direct evidence of glacial deposits present in the Silurian sequences of Brazil (Grahn & Caputo, 1992) and elsewhere in South America (see Hambrey, 1985).Obviously, the third (and possibly fourth) order cycles being discussed here are superimposed upon the longer term second order and first order cycles (Haq, Hardenbol & Vail, 1988) linked to global tectonic processes. It is likely, for example, that the particularly high sea-levels of the Telychian and early Sheinwoodian reflect the coincidence of interglacial 'greenhouse' conditions with an increase in the volume of oceanic ridges. Abstract -Global sea-level fluctuated markedly during the early Silurian, probably as a result of the waxing and waning of ice-sheets in the South American portion of Gondwana. The highest sea-levels of the Silurian are recorded by the Telychian upper crispus-lower griestoniensis and spiralis-lower lapworthi biozones. Other highstands occurred in the early Aeronian, during the convolutus Zone (mid Aeronian), guerichi Zone and late turriculatus Zone (early Telychian), and early Sheinwoodian. Low sea-levels characterized much of the argenteus and sedgwickii zones (Aeronian), the utilis Subzone (late guerichi-early turriculatus zones, early Telychian), the late Telychian (commencing in the mid lapworthi Zone) and, after a period of apparently only small amplitude sea-level fluctuations in the late Sheinwoodian and earliest Homerian, the mid-late Homerian, in particular the early nassa Zone. Facies (and faunal) changes in the Lower Silurian do not support the P and S model of Jeppsson and others, but are consistent with the sea-level changes proposed herein. Mid Telychian marine red beds appear to have been deposited during a minor sea-level fall immediately after a period of very high sea-levels, rather than during a transgressive episode as previously suggested. Comparison of the sea-level curve presented herein with those constructed in the past is hampered by the lack of precision currently possible in the correlation of early Silurian deep water (graptolitic) and shallow water (shelly) sequences. Improving the precision of this correlation should be a priority for future research.*