"INTRODUCTION To provide a healthy indoor environment for building occupants@ an adequate outside air supply is necessary to dilute and remove indoor air contaminants. Outside air is normally provided by mechanical ventilation in commercial buildings and by natural ventilation in most residential buildings. However@ the energy required to condition the outdoor air in buildings is often a significant portion of the total space conditioning load. Large commercial buildings normally operate continuously at a slightly pressurized condition in which air infiltration is expected to have less impact on building conditioning load. In most small office buildings and residential buildings@ the air infiltration has a significant impact on the space conditioning load since the buildings are not pressurized and air conditioning systems operate intermittently. In a radiantly heated and cooled office@ the infiltration not only has an impact on building heating and cooling load; it also strongly affects the indoor humidity level and hence the ability to operate the radiant cooling panels without condensation. The magnitude of the infiltration into this type of the building must be known in order to size the radiant panels and dehumidification equipment for this type of system properly. Significant research has been done on building air flow models and infiltration measurement procedures. Building air flow models can be classified as single zone models and multizone models. A widely used multizone model is COMIS which was developed by Feustel and Raynor-Hoosen (1990)@ and improved by Feustel (1999). The single zone LBNL model developed by Sherman and Grimsrud (1980) has been used very widely. Walker and Wilson (1993) also proposed a well recognized enhanced single zone model. For small office buildings and residential buildings@ the single zone model is applicable. According to Persily (1986) and Walker and Wilson (1998)@ these models can have an error of 40% or higher. In order to obtain an applicable yearly infiltration profile@ these models should be combined with air infiltration measurements. Commonly used infiltration measurement procedures include the tracer gas method@ CO2 concentration decay method@ blower door test method@ etc. The office studied in this paper is a small university office area which includes space for faculty@ graduate student and staff offices and a meeting room at a university in Pittsburgh@ PA. This office space@ with an area of 580 m2 (6228 ft2)@ uses a radiant heating and cooling system combined with a solid desiccant ventilation unit. Gong and Claridge (2006) studied the indoor humidity levels of this radiantly heated and cooled space and noted that infiltration was the main factor affecting operation of the cooling system without condensation on the panels and energy efficiency of the radiant system. In order to accurately simulate the heating and cooling consumption of the space and precisely size the equipment in the energy supply and energy distribution systems@ the infiltration level of the space needs to be carefully measured and studied. This paper provides an overview of previous infiltration studies of the space@ analyzes blower door measurement results and evaluates the infiltration by the LBNL model. This paper also compares infiltration measurements in the space by the tracer gas method@ CO2 concentration method@ blower door test method@ and logged humidity data. It then combines the differing measured results with simulation to investigate reasons for these differences and reach a conclusion on the ""real"" infiltration value in this office. Some possible leakage points are also identified from site measurements."