Let $V$ be a vector space. Let $\mathbf {F} :\mathbb {R} ^{3}\to V$, we say that $\mathbf {F}$ is a vector field; it associates a vector from $V$ with every point of $\mathbb {R} ^{3}$.

In the physical world, examples of vector fields are

(i) The electric and magnetic fields in space ${\vec {E}}(\mathbf {r} ),{\vec {B}}(\mathbf {r} )$

(ii) The velocity field in a fluid ${\vec {v}}(\mathbf {r} )$

(ii) The velocity field in a fluid ${\vec {v}}(\mathbf {r} )$

Let $C$ be a scalar field. We define the gradient as an "operator" $\nabla$ mapping the field $C$ to a vector in $\mathbb {R} ^{3}$ such that

$\nabla C=\left({\frac {\partial C}{\partial x}},{\frac {\partial C}{\partial y}},{\frac {\partial C}{\partial z}}\right)$, or as is commonly denoted $\nabla C={\frac {\partial C}{\partial x}}{\hat {x}}+{\frac {\partial C}{\partial y}}{\hat {y}}+{\frac {\partial C}{\partial z}}{\hat {z}}$

We shall encounter the physicist's notion of "operator" before defining it formally in the chapter Hilbert Spaces. It can be loosely thought of as "a function of functions"

Recall from multivariable calculus that the total derivative of a function $f:\mathbb {R} ^{3}\to \mathbb {R}$ at $\mathbf {a} \in \mathbb {R} ^{3}$ is defined as the linear transformation $A$ that satisfies

In the usual basis, we can express as the row matrix $f'(\mathbf {a} )=A=\displaystyle {\begin{pmatrix}{\tfrac {\partial f}{\partial x}}&{\tfrac {\partial f}{\partial y}}&{\tfrac {\partial f}{\partial z}}\\\end{pmatrix}}$

It is customary to denote vectors as column matrices. Thus we may write $\nabla f=\displaystyle {\begin{pmatrix}{\tfrac {\partial f}{\partial x}}\\{\tfrac {\partial f}{\partial y}}\\{\tfrac {\partial f}{\partial z}}\\\end{pmatrix}}$

The transpose of a matrix given by constituents $a_{ij}$ is the matrix with constituents $a_{ij}^{T}=a_{ji}$

Thus, the gradient is the transpose of the total derivative.

Let $\mathbf {F} :\mathbb {R} ^{3}\to \mathbb {R} ^{3}$ be a vector field and let $\mathbf {F}$ be differentiable.

We define the curl as the operator $(\nabla \times )$ mapping $\mathbf {F}$ to a linear transformation from $\mathbb {R} ^{3}$ onto itself such that the linear transformation can be expressed as the matrix

$(\nabla \times \mathbf {F} )_{ij}={\frac {\partial F_{j}}{\partial x_{i}}}-{\frac {\partial F_{i}}{\partial x_{j}}}$ written in short as $(\nabla \times \mathbf {F} )_{ij}=\partial _{i}F_{j}-\partial _{j}F_{i}$. Here, $x_{1},x_{2},x_{3}$ denote $x,y,z$ and so on.

the curl can be explicitly given by the matrix: $\nabla \times \mathbf {F} ={\begin{pmatrix}0&\partial _{1}F_{2}-\partial _{2}F_{1}&\partial _{1}F_{3}-\partial _{3}F_{1}\\\partial _{2}F_{1}-\partial _{1}F_{2}&0&\partial _{2}F_{3}-\partial _{3}F_{2}\\\partial _{2}F_{1}-\partial _{1}F_{3}&\partial _{3}F_{2}-\partial _{2}F_{3}&0\\\end{pmatrix}}$

this notation is also sometimes used to denote the vector exterior or cross product, $\nabla \times \mathbf {F} =(\partial _{2}F_{3}-\partial _{3}F_{2}){\hat {x}}+(\partial _{1}F_{3}-\partial _{3}F_{1}){\hat {y}}+(\partial _{1}F_{2}-\partial _{2}F_{1}){\hat {z}}$