# Regarding the definition of vector field flow

To make the connection to the Lie derivative, let $t \mapsto \Phi^X_t$ be the 1-parameter group of diffeomorphisms (or flow) generated by the vector field $X$. The differential $d\Phi^X_t$ of each diffeomorphism maps the vector field Y to a new vector field $\mathrm{d}\Phi^X_{t}(Y)$. (http://en.wikipedia.org/wiki/Lie_bracket_of_vector_fields)

Not sure why $d\Phi^X_t$ would map vector field $Y$ to a new vector field. Can anyone explain this? Also, what is the difference between $\mathrm{d}\Phi^X_{t}$ and $d\Phi^X_t$?

-
The Roman d is the same as the italic d. Each $\Phi^X_t$ is a diffeomorphism of $X$, which is a map $X \to X$; thus, its derivative maps $T_pX \to T_{\Phi^X_t(p)}X$ for each $p \in X$. Since this is a diffeomorphism, it's injective and surjective, and so starting with a smooth vector field on all of $X$ gives you a new vector field on all of $X$; you can check that this is smooth. Also, your title has nothing to do with your question - would you mind changing it? – Paul VanKoughnett Apr 9 '13 at 4:55
@PaulVanKoughnett But I believe $X$ is vector field here... – Zeus Apr 10 '13 at 7:52
Besides, usually, $\Phi_t^V (x)$ means flow induced by vector field $V$ that maps manifold $X$ to $X$... But above seems different from that. – Zeus Apr 10 '13 at 7:52
Oh, sorry, I thought $X$ was the manifold in question. If you replace all the $X$'s in my above comment by $M$'s, you should get a true statement. The above $\Phi^X_t$ is precisely the flow you're talking about in your second comment. – Paul VanKoughnett Apr 10 '13 at 20:04