By what factor is Kosaraju's algorithm for finding strongly connected component slower as compared to Tarjan's algorithm. It appears to me that the factor should be 3 as there are 2 dfs passes in Kosaraju's algorithm and we also have to transpose the graph once. Am I missing something ?

Also is there any way to find the tranpose graph without declaring a new graph and storing the edges in reverse order ? (Basically reversal of graph without use of extra memory)