This is not an excellent exposition of Galois theory. It is, however, a book well worth reading for the single reason that it sticks to Galois, including a full translation of Galois' 1831 memoir (13 pages).

The immediate goal for Galois was to understand solvability by radicals, in particular of the general n:th degree equation. But to understand Galois we must first study what was done before him (§§1-27). Lagrange is the most important predecessor. He presented a unified approach to the solvable cases (n4 cases. This resolvent business is historical ballast, we would say today, but it is from this tradition that Galois departs. In particular, he arrives at the "Galois group" in terms of resolvents. With the Galois group in place, things flow more smoothly. Essentially as in the modern theory, Galois shows that if an equation is solvable by radicals then its Galois group is "solvable". All this is §§28-48. Edwards the constructivist now inserts a bunch of Kronecker material on the existence of roots (§§49-61). Then it's back to Galois (§§62-71) to see how he puts his theory to use. Galois doesn't even bother to spell out that the unsolvability of the general equation of degree n>4 follows since its Galois group S_n is not solvable; instead he finds a curious criterion for solvability which is involves no group theory (even Edwards calls this result "rather strange").