Several methods of measuring the unsaturation remaining in natural rubber after cyclization have been compared. Approximate agreement was obtained with the reagents, perbenzoic acid, phenyl iododichloride and ozone; iodine chloride gave high values. Perbenzoic acid is considered the most satisfactory of these reagents. The unsaturation in some cyclized rubber samples was found to be below 20% of that in the original rubber. This value is inconsistent with a cyclized rubber structure of single rings each involving two isoprene units and supports instead a polycyclic structure. This structure is also supported by some degradation experiments. Infra-red spectroscopy shows that the double bonds remaining after cyclization are not of the original trialkyl-substituted type. Use has been made of measurements of total unsaturation by means of perbenzoic acid and of trialkyl-substituted double bonds by infra-red spectrometry to investigate the kinetics of the cyclization reaction in solution and catalyzed by stannic chloride. All the results were accurately fitted by theoretical equations derived for a reaction proceeding in step-wise fashion along the rubber chains. The average size of the polycyclic structures formed during cyclization was found to be independent of both rubber and catalyst concentrations but to be markedly dependent on the temperature, varying in a series of experiments from approximately one and a half condensed rings at 110 $^\circ$C to six rings at 60 $^\circ$C. The rate of the reaction was first order in the rubber and second order in the catalyst.