Abstract. Given a finite metric CW complex X and an element α ∈ πn(X), what are the properties of a geometrically optimal representative of α? We study the optimal volume of kα as a function of k. Asymptotically, this function, whose inverse, for reasons of tradition, we call the volume distortion, turns out to be an invariant with respect to the rational homotopy of X. We provide a number of examples and techniques for studying this invariant, with a special focus on spaces with few rational homotopy groups. Our main theorem characterizes those X in which all non-torsion homotopy classes are undistorted, that is, their distortion functions are linear.The object of quantitative topology is to make more concrete various notions coming from the existential results of algebraic topology. Thus, while classical rational homotopy theory gives an exhaustive family of algebraic correlates to rational homotopy classes of simply-connected spaces, a quantitative homotopy theory seeks to give geometric examples or descriptions linked to the algebraic properties of these objects.The term "quantitative homotopy theory" seems to have first been used by Gromov in the conference paper [Gro99] and in Chapter 7 of the near-simultaneous book [Gro98], although the ideas date back as far as [Gro78]. Construed broadly, however, this program fits into a tradition of extracting metric information from topological data which includes problems from systolic geometry, geometric group theory, and other areas. In particular, geometric group theorists, as specialists in fundamental groups, have explored a host of asymptotic invariants whose higher-dimensional analogues may also be of interest. These include the Dehn function, growth of groups, and distortion of group elements and subgroups.Higher-dimensional isoperimetric functions of groups, that is, of their Eilenberg-MacLane spaces, have been studied in some detail, notably by Gromov [Gro96], Alonso-Wang-Pride [AWP], Brady-BridsonForester-Shankar [BBFS], and Young [Young]. A common theme of this body of literature is the plurality of possible definitions, many of which are equivalent in the one-dimensional case. The subject of growth of higher homotopy groups was broached in chapters 2 and 7 of [Gro98] with a number of examples and a conjecture for simply-connected spaces.Here we analyze a higher-dimensional analogue of distortion. Heuristically, the distortion of a group element α ∈ G is given byIf G is the fundamental group of a space, word length is a natural measure of size. On the other hand, if G = π n (X) for a space X, we can choose inequivalent measures of size by taking advantage of different features of a metric structure on X, leading once again to a plurality of definitions. For example, suppose that X is a CW complex with a piecewise Riemannian metric. Then we can choose to minimize the Lipschitz constant of a representative, its volume, or more generally the m-dilation for some 1 ≤ m ≤ n, that is, how much the map f : S n → X distorts m-dimensional tangent subspaces. Moreove...