Definitions and Rules

Partial derivatives help us track the change of multi-variable functions by dealing with one variable at the time.

DNA forms a double helix, but the curvature of this helix depends on the temperature and on the salinity (concentration of salt). If we wanted to understand the curvature, we would do experiments by varying the conditions and measuring the curvature each time. If we did our experiments well, we wouldn't try changing both the temperature and the salinity. We would first hold the salinity fixed and change the temperature. Once we understood how temperature affects curvature, we would run a second set of experiments, holding the temperature fixed and varying the salinity. Combining the results, we would understand how both temperature and salinity affect curvature.

Mathematically, the curvature is a function $f(x,y)$, where $x$ is the temperature and $y$ is the salinity. Varying the temperature means comparing $f(x,y)$ to $f(x+h,y)$, and we can ask for the rate of change. Varying the salinity means comparing $f(x,y)$ to $f(x,y+h)$. By taking limits, we can compute two kinds of derivatives:

There are many notations for partial derivatives. If $z = f(x,y)$, then $$\displaystyle f_x(x,y) = f_x = \frac{\partial f}{\partial x}= \frac{\partial}{\partial x}f(x,y) = \frac{\partial z}{\partial x}= f_1 = D_1 f= D_x f$$ and $$\displaystyle f_y(x,y) = f_y = \frac{\partial f}{\partial y}= \frac{\partial}{\partial y}f(x,y) = \frac{\partial z}{\partial y}= f_2 = D_2 f= D_y f$$

The rough and precise definitions of limits of functions of two (or more) variables work the same way:

When computing $f_x$, we treat $y$ as a constant because it is a constant. After all, we are doing today's experiments at fixed salinity. This means that we can apply all of our familiar differentiation rules, pretending that the only variable is $x$.