I am still interested to know how they know the 40th row is the one needed? What if someone else adds more rows to this table? Would the correct row still be the 40th or would it now be the 45th? It seems to me that selecting data based on the row number is a short-lived solution to getting the data needed, and I would recommend selecting the rows based on the contents of the columns in that table.
If it is about percentiles or ranking, I would use the nested "Top 1... Top 40 percent" technique. (Note: I used percent here on the idea that the table will grow/change over time, as opposed to wanting exactly the 40th row).
I also like GSquared's CTE solution.