The document describes fitting curves to data using R. It discusses intrinsically linear and linearizable relationships that can be fit with linear regression. For truly non-linear relationships, the nls method is used to fit functional forms such as power, exponential, and piecewise models. Examples are given fitting simulated data to a known cubic function using nls to estimate the power parameter, and comparing fits with and without an intercept term. Model quality is assessed using residual sum of squares.