Over the past decade, Argo floats have provided an unprecedented number of profiles of the global oceans (to 2000m depth), far surpassing the number collected historically fromship-based hydrography. The original design of the Argomission specified nominal 3 x 3 degree spacing, with 10 daysampling interval, of the oceans between 60 °N and 60 °S,excluding the high latitudes and marginal seas. The exclusionof the high latitudes was due to the inability of early floats tosample under sea-ice. Technological advances in float designin recent years now give us this capability. Advancements havecome through re-design of hardware (i.e. armoured ice floats),software (ice-avoidance algorithm and open-water test) andcommunications (Iridium), allowing the transmission of storedwinter profiles. Observing circulation in seasonally ice-coveredseas is challenging. To date, most observations have beenmade during ice-free summer periods and consequently thewinter circulation beneath the sea-ice is not well understood.Despite this, Argo has already made a significant contribution tohigh latitude research with successful deployments of floats inthe polar oceans of both hemispheres. As of December 2008,over 100 floats had been deployed above 60 °N and over 200below 60 °S. Approximately 60% of these floats are still active(the failure rate of early floats was high as the ice-capabletechnology was being developed and tested). Mortality rates ofnewer ice floats are now equivalent to those deployed in lessdemanding conditions. In fact, a number of floats deployed inthe Weddell Sea have survived for 7 years (surpassing 225profiles) equal to some of the longest-lived floats deployedglobally. The high latitudes are important deep water massformation regions. The Southern Ocean connects the globalocean basins and regulates the meridional overturningcirculation. The exposed Arctic Ocean will have importantconsequences for ocean and atmospheric circulation, moistureand heat fluxes. Therefore, both polar regions play a critical rolein setting the rate and nature of global climate variabilitythrough their moderation of the earth's heat, freshwater andcarbon budgets. Recent studies have shown that certainregions at high latitudes are warming more rapidly than theglobal average. Some of the most important climate changesignals are seen near ice shelves and within the sea ice zone.In the Arctic, reductions in sea-ice extent and changes infreshwater fluxes, deep water mass properties and convectionhave been observed. Similarly strong reductions in sea-icecoverage are occurring near the Antarctic Peninsula while smallincreases appear in the Ross Sea. At the same time decreasingsalinity on the Ross Sea shelf is thought to be linked toincreased glacial melt. The Argo network has been crucial fordocumenting the recent changes in the open ocean; robust andlarge-scale freshening of the Southern Ocean has beenobserved from Argo and historical hydrographic data. Butsampling at these higher latitudes is less systematic than forthe rest of the globe. Therefore, observations of high latitudeoceans in both hemispheres should be a top priority. Inconsidering sampling strategies for the high latitudes werecommend extending the Argo network beyond 60 °S and60 °N through the deployment of ice-capable floats at thenominal density (3 x 3 degrees). In addition, regional arrays ofacoustically-tracked floats will provide a more focused effort onbasin scales. An established array of sound-sources (RAFOS)and acoustically-tracked floats in the Weddell Sea is alreadyyielding valuable information on ocean circulation and structurebeneath the sea-ice. A similar array should be established tosample the Ross Sea gyre. In the Arctic, an array of lowfrequency (< 100 Hz) sound sources would be required toprovide basin-wide geo-location for profiling floats. Now that wehave come to review the past decade of progress within Argo,we find there is considerable support and justification for theofficial extension of the Argo array into the seasonally icecoveredseas. Sustained, comprehensive observation of thepolar oceans is required to adequately monitor global climatechange signals. This can only be achieved in a broad-scale andcost-effective way by using autonomous platforms like Argoprofiling floats. It is thus imperative that a commitment is madeto enhance and maintain a profiling float array in the highlatitudes. The extension of the core Argo array beyond 60degrees in both hemispheres will ensure that it remains one ofthe most important and truly global components of the oceanobserving system.