An understanding of the microclimatology of the ice free areas in Antarctica is needed to better quantify the impacts of human activities on this fragile environment and to help monitor global climate change processes. This research was undertaken to characterise the surface and sub surface factors influencing Antarctic soil microclimates and quantify the sub surface thermal processes resulting from these conditions. A measurement programme was developed and implemented at three contrasting sites within the Dry Valleys of the McMurdo Sound Region, Antarctica during the 1994-95 summer season, and samples were collected for further analysis in New Zealand. Climatic variables measured included the surface radiation and energy balance, soil and air temperature, and windspeed and direction. Important soil physical and thermal properties such as moisture content, bulk density, particle density, grainsize distribution, salinity, specific heat and thermal conductivity were also measured. The sites investigated were located at Scott Base on Ross Island in the coastal climate zone, at the foot of Mount Brooke in the Coombs Hills which has a high mountainous climate, and in the Northwind Valley (Greenville Valley) in the Convoy Range which is climatically midway between the two. Apart from being climatically different, the soils examined at each site were also different. Basalt derived soil was examined at Scott Base, dolerite derived soil was studied at the Coombs Hills, and a sandy soil derived from sandstone was examined in the Northwind Valley. Similar amounts of shortwave radiation were recorded at each site under comparable conditions, although surface albedo varied significantly which determined the amount of radiation reflected. Consequently, the Scott Base site, which had the darkest surface and lowest albedo, had the greatest net radiation inputs, and the Northwind Valley site had the lowest due to its pale colour and high albedo. Despite this, the Northwind Valley soil had the largest soil heat flux. The reason for this is not fully understood, but involves the soil temperature gradient which is the main determinant of the soil heat flux. The warmer climate at Scott Base resulted in generally warmer soil temperatures which produced an active layer greater than two times that seen at either of the other sites. Due to its relatively high moisture content, ice-cement was found at about 0.25 m which delineates the summer O°C isotherm, and indicates the permafrost is at a similar depth. At the other two sites which were a lot colder, there was no permanent ice cement found due to their dry nature. The active layer was approximately 0.10 m deep at the Coombs Hills site, and was only 0.08 m at the Northwind Valley site. The permafrost at these sites would also be at about these depths. Soil temperatures, which were strongly negative, fluctuated by up to 5°C at a depth of about 0.20 m at these two sites. Under snow cover, the soil heat flux at both the Scott Base and Coombs Hills sites became negative which means energy was being lost from the profile, and the penetration depth of diurnal temperature variations was reduced. The main factors influencing the soil thermal regime are incoming solar radiation, air temperature, windspeed, and surface albedo. Soil properties including moisture content, thermal conductivity and heat capacity are less important in determining the variations observed.