A cone over a diagram $F \colon \mathsf{J} \to \mathsf{C}$ with summit or apex $c \in \mathsf{C}$ is a natural transformation $\lambda \colon c \Rightarrow F$ whose domain is the constant functor at $c$. The components $(\lambda_j \colon c \to Fj)_{j\in \mathsf{J}}$ of the natural transformation are called the legs of the cone. Explicitly:
Dually, a cone under $F$ with nadir $c$ is a natural transformation $\lambda \colon F \Rightarrow c$, whose legs are the components $(\lambda_j \colon F_j \to c)_{j \in \mathsf{J}}$. The naturality condition asserts that, for each morphism $f \colon j \to k$ of $\mathsf{J}$, the triangle \(\xymatrix@=10pt{ Fj \ar[dr]_{\lambda_j} \ar[rr]^{Ff} & & Fk\ar[dl]^{\lambda_k} \\ & c}\) commutes in $\mathsf{C}$.