A thesis submitted to the Faculty of Engineering in fulfilment of the requirements for the degree of Master of Technology in Mechanical Engineering 2005 / The objective of this project was to monitor the high temperature creep behaviour of 441 stainless steel. Two different alloys of 441 were investigated; the main difference between them being the Niobium content. Particularly imporlant to the project was how the Niobium content and grain size affected the creep resistance of the material. Creep tests were performed using purpose built constant load creep test rigs. Initially the rigs were not suitable for the testing procedures pertaining to this project. This was due to persistent problems being experienced with regards the reliability and reproducibility of the rigs. After various modifications were made the results produced from the rigs were consistent. Creep test data was used in order to determine the mechanism of creep that is operative within the material (at a predetermined temperature) under a predetermined load. Particular attention was paid to the resulting stress exponents. in order to identify the operative creep mechanism. The identification of the operative creep mechanisms was also aided by microscopical analysis. This analysis was also necessary to monitor how the grain size had altered at various annealing temperatures. Heat treatment was used as a method to alter the high temperature strength and microstructure of the material. Heat treatments were performed at various temperatures in order to determine the ideal temperature to promote optimum creep resistance of 441. All heat treatments were performed in a purpose designed and built high temperature salt bath furnace. The commissioning of the salt bath formed part of the objectives for this project. Sag testing was also conducted, using purpose built sag test rigs. It was necessary to design and manufacture a sag test rig that could be comparable to the industry accepted method of sag testing known as the two-point beam method, as this method is believed to produce inconsistent results. Conclusions have been drawn from the results of the data and from previous research on the subject matter.