Recent theoretical and critical re-evaluations of musical structure have fostered a new understanding of the concept: it has been argued that structure is best understood as a pluralistic and diachronic phenomenon emerging from structural relationships inferred individually by analysts, performers and listeners through the media of scores and performances. Contrary to conventional definitions, it has also been shown that musicians themselves – and conductors in particular – tend to conceptualise large-scale structure dynamically, that is, in terms of such properties as shape, motion, goal-directedness and climax. To date, however, the implications of this concept of structure have not been fully explored in analyses of performances of large-scale symphonic music. This thesis examines the ways in which given conductors conceptualise and project representations of structure in performance, and it utilises a range of analytical approaches suited for the study of those representations. The work centres around Wilhelm Furtwängler (1886–1954), a prominent conductor for whom musical structure and the aforementioned dynamic features of form were primary performative considerations. After examining the concept of musical structure and the conducting style of Furtwängler, the thesis explores the latter’s shaping of symphonic music through three contrasting examples: two movements commonly interpreted as being in sonata form and variation form (respectively, the first and fourth movements of Brahms’s Fourth Symphony), and a symphony in its entirety (Beethoven’s Fifth Symphony). In each case, judicious comparison with select recordings of other conductors is also undertaken for the sake of contextualisation. Alongside established quantitative techniques, the project utilises experimental qualitative methods based on the perception of intensity and motion in performance in order to develop an innovative multi-parametric approach to the representation and thus to the understanding of musical structure in performance.