analyzing sequential files efficiently

I have 500 sequential files whose filename is in the form "datafile_XX.mat". Each file contains an n by 8 matrix where n varies per file. But all files have 8 columns, only the number of rows varies.

I want to get the average of the first 4 columns of all rows of all 500 files. How do I do it? Also, can you point to me a tutorial on these things?

If it helps, the files were created using 'sprintf'. Somehow there should be a reverse to this right?

Thanks

[Mohammad Abouali]

So are they MAT file or a text file? Sprintf doesn't create mat file!!! Could you upload two or three sample files?

[Geoff Hayes]

I suspect it was meant that sprintf is used to create the file name

 for k=1:500
     filename = sprintf('datafile_%d.mat',k);

     % do stuff
 end

@cgo - you could use the above as a framework for reading (or loading) your files which would be part of the do stuff. What can you tell us about the contents of each file? Was the same named variable saved to each file?