A predictive theory of muscle contraction and chemical energy consumption can transform human movement science, e.g., helping us better understand movements such as walking and running and informing the design of effort-reducing assistive and prosthetic devices. Such a theory can also inform a quantitative understanding of the genetic basis of heart disease and other muscular dysfunction. An accurate theory of muscle contraction and energy consumption does not exist. While a clear picture of muscle contraction, including energy consumption, has emerged at the single molecule scale, the simplified conditions of these experiments limit their application to larger scales. We propose to produce a multi- scale mathematical theory of muscle contraction, based on molecular and cellular measurements, to understand muscle function in vivo and test such a bottom-up theory's accuracy at the whole muscle or whole body level, as will be relevant in applications. The proposed research builds on previous work, where we developed a theory, described by linear ordinary differential equations, coupled to integro-partial differential equations, that quantitatively describes experiments from single molecules to large ensembles. Because our theory is described by differential equations, unlike other models, it can be inverted to predict muscle energy consumption from muscle force. Such simulations predict a cost for the rate of muscle force production, which is thought to be critical to understanding the energetics of human walking. Despite this promising result, the theory lacks components necessary to quantitatively describe muscle at larger (i.e. cellular, organ, etc.) scales. We will therefore perform experiments, motivated by the theory, to identify and quantify the missing components. In Aim 1 we will extend the theory to the cellular scale by generating a self-consistent data set from the single molecule to cellular (muscle fiber) scale. These experiments will characterize the transient interactions (weak binding) between molecules involved in muscle contraction that are too rapid to measure at the molecular scale, and so must be characterized via multi-scale measurements interpreted with the theory. In Aim 2 we will extend the theory to conditions relevant to locomotion by performing novel experiments on muscle molecules and cells under conditions that replicate forcibly lengthened muscle (eccentric contraction), a situation that frequently occurs during locomotion. We will test hypotheses, motivated by the model, that 1) molecular bonds are forcibly broken when muscle is lengthened, and 2) this bond breaking leads to transient instabilities that cause catastrophic loss of muscle force. In Aim 3, we will collect data for muscle energy consumption from human subjects. These experiments will allow us to test and refine candidate muscle energy cost models. The theory already makes testable predictions; our measurements will allow us to test these predictions, refine the model, and improve on current muscle energy cost models.