Understanding the aggregation mechanism of amyloid proteins, such as Sup35NM, is essential to understanding amyloid diseases. Significant recent work has focused on using the fluorescence of thioflavin T (ThT), which undergoes a red shift when bound to amyloid aggregates, to monitor amyloid fibril formation. In the present study, the progression of the total mass of aggregates during fibril formation is monitored for initial monomer concentrations in order to infer the relevant aggregation mechanisms. This workflow was implemented using the amyloid-forming fragment Sup35NM under different agitation conditions and for initial monomer concentrations spanning 2 orders of magnitude. The analysis suggests that primary nucleation, monomeric elongation, secondary nucleation, and fragmentation might all be relevant, but their relative importance could not be determined unambiguously, despite the large set of high-quality data. Discriminating between the fibril-generating processes is shown to require additional information, such as a fibril length distribution. Using Sup35NM as a case study, a framework for fitting the parameters of arbitrary amyloid aggregation kinetics is developed based on a population balance model (PBM), which resolves not only the total aggregate mass (monitored experimentally via ThT fluorescence) but the entire fibril length distribution over time. In addition to the rich new set of ThT fluorescence data, we have reanalyzed a previously published aggregate size distribution using this method. With the size distribution, it was determined that in the reanalyzed in vitro experiment, secondary nucleation generated significantly fewer new Sup35NM fibrils than fragmentation. The proposed strategy of applying the same PBM to a combination of kinetic data from fluorescence monitoring and experimental fibril length distributions will allow the inference of aggregation mechanisms with far greater confidence than fluorescence studies alone.