The project investigates the Cavity Perturbation Method (CRM) and design of a coaxial resonator covering a wide frequency band from 2.5 to 11.5 GHz. Using the published analyses for coaxial cavities, original theoretical analysis of Cavity Perturbation Method for TEM coaxial cavity was carried out by the author. Using similar approach, the H(_011) cylindrical cavity was analyzed. The deliberate objective of the method was to produce a wide band technique since there are none available at present. The method then examines the possibility of employing the coaxial cavity as an important and necessary measuring device for dielectric measurements using CPM. It was the aim in the investigations to state clearly what assumptions were made so that the accuracy of the calculated results could be assessed. A new measurement technique, using a Sweep Generator/ Spectrum Analyzer Assembly (SG/SA) was introduced. The technique is based on very precise measurements of the changes in cavity Q and its resonant frequency, displayed on Spectrum Analyzer, when the test sample of the material is inserted. Because the main objective of the project is the method of measurements, the assessment of its validity and accuracy was fully discussed. Measurements were carried out on fifteen materials of various types of ceramics over the frequency range of interest. These materials were used mainly for the assessment of the validity and accuracy of the new Sweep Generator/Spectrum Analyzer Assembly technique, and also a test of the capability of the constructed broadband coaxial cavity to be used for dielectric measurements. The well-known Bridge and Q-meter methods, were employed initially to provide the low frequency values for the materials as reference. An attempt has also been made to justify the validity of the universal law on those materials. Additional work not directly related to the main project has been carried out on Curie temperature measurements. The object was to develop a simple and reliable method using the Gouy balance technique.