Note that there was no MKL when Quadpack was published, and Intel was a young company. The IBM PC was yet to be conceived.
Read the included documentation and decide what type of integration is needed. Is the integrand singular at the limits of integration or within the interval? Do you wish to use an adaptive algorithm or a simple one? What precision do you require?
Software from Netlib has a more or less standard structure.
1. There will always be a set of subroutines/functions. Often, the documentation is included in the form of comments in the source code.
2. In earlier software, it was common to include the code for the driver (see 3. below) with 'C' at the beginning of such lines. You copy those lines, remove the 'C's at the beginning of each line and save as a Fortran file.
3. Sometimes, there is a short "driver", i.e., an example main program that illustrates calling the subroutine.
4. Some packages contain double-precision and single-precision versions concatenated together into one long file. You will have to split the parts apart before using.
Suppose you had composed the source code for the needed Quadpack routine yourself. What would you do to compile and run your program? You do the same, except that instead of entering the source code in an editor you download a file.
If you still don't know what to do, you really must read Fortran books or a user guide to learn how to compile, link, run and debug Fortran programs.