The emphasis of this thesis is on cosmology experiments using a submillimetre continuum array receiver. Divided into two parts, the thesis initially explores an astrophysical model, followed by an experiment based on results of the model. The second part deals with some of the design considerations for SCUBA, the Submillimetre Common User Bolometer Array. The thesis touches on all aspects of a problem in astronomy: the theoretical modelling, an observational experiment, and the development of instrumentation. Part I deals with the astronomical question of gravitational lenses and the microwave background. Simulations of an anisotropic submillimetre background due to dusty primeval galaxies are presented, along with the observations that would result when gravitational lensing by clusters of galaxies is considered. The experiment was tried with the single pixel bolometer system on the James Clerk Maxwell Telescope and it measured an upper limit on CBR anisotropy of ΔT/T < 4.6x10^-3 (95% confidence) towards the cluster Cl2244 at the angular scale of 5 ''. The small angular scale is due to the aid of the cluster lensing effect. The second part looks at the extensive design considerations for the feedhorn antennas and the bolometer integrating cavities in SCUBA. Single mode optics optimizes aperture efficiency and this is usually achieved by having a length of waveguide, but analysis here shows that one can still have a single moded device without a waveguide. A filter drum furthers SCUBA's capabilities by allowing the detector arrays to operate at higher frequency wavebands, as well as the primary wavebands. At the higher frequencies the feedhorns are overmoded, which is a case investigated in this thesis. The SCUBA Project will have the horns fabricated by an outside contractor according to the specifications derived in Part II.