You would use Bayes Theorem for this conditional probability:
Let B be the event that he takes the bus, C the event that he takes a car, and L the event that he is late.
We are looking for the probability that he takes the bus given that he is late. In probability notation it looks like:
P( B | L )
Bayes says: P( B | L ) = P (B ∩ L) / P (L)
I recommend you try work this out on your own without looking forward first.
Here is the working out:
P (B ∩ L) = .65 * .18 (intuitively it is the probability that he takes the bus and is late)
P(L) = P( B ∩ L) + P(C ∩ L) = .65 * .18 + .35 * .09
(intuitively, the probability that he takes the bus and is late, or takes a car and is late)