During the testing of development engines and components, intrusive instrumentation such as Kiel head Pitot probes and shrouded thermocouples are used to evaluate gas properties and performance. The size of these instruments can be significant relative to the blades and their impact on aerodynamic efficiency must be considered when analysing the test data. This paper reports on such parasitic losses for instruments mounted on the leading edge of a stator in a low pressure turbine, with particular emphasis on understanding the impact of probe geometry on the induced loss. The instrumentation and turbine blades were modeled in a low Mach number cascade facility with an upstream turbulence grid. The cascade was designed so that leading edge probes were interchangeable in-situ, allowing for rapid testing of differing probe geometries. RANS calculations were performed to complement the experiments and improve understanding of the flow behaviour. A horseshoe vortex-like system forms where the probe body joins the blade leading edge, generating pairs of streamwise vortices which convect over the blade pressure and suction surfaces. These vortices promote mixing between freestream and boundary layer fluid, and promote the transition of the boundary layer from laminar to turbulent flow. Tests with realistic probe geometries demonstrated that the design of the shroud bleed system can impact losses and a study using idealised cylinders was performed to isolate the impact of probe diameter, aspect ratio and incidence. Beyond a probe aspect ratio of two; parasitic loss was found to scale approximately with probe frontal area. ; Other ; The authors would like to thank Rolls Royce plc. for funding, technical support and permission to publish this work, in particular Frederic Goenaga and David Lambie; thanks are also due to Mark Erland for his insights into probe design and operation. The insight and assistance of Howard Hodson and Chris Clark of the Whittle Laboratory are gratefully acknowledged. Funding was received through the SILOET II project.