It is well known that when measuring electron densities in H II regions locally via emission line ratios the values observed are 100cm-3, but gobal values obtained via emission measure using surface brightness as the observed parameter are ~1cm-3. This difference is attributable to density fluctuations,and traditional models for these is the ``Filling factor (FF)'' model, of Osterbrock and Flather. Implicit in this model is the optical thinness of the fluctuations. We show here that if these fluctuations are optically thick the diagnostic diagrams which are used to find parameters, notably abundances, will be significantly affected. We present observational arguments supporting these ``clumpy'' models, and outline some initial results of their use.