I read a short article written by Jack Elliano of Electra-Print. I haven't fully absorbed all of the article, but I think it essentially questions the usefulness of square wave response testing as a measure of output transformer performance. Here's the article:
http://electra-print.com/techblog.php?blogid=5
I think what Jack is saying makes sense, but I'm still waiting for it all to sink into my brain. Any comments? I often see people injecting square waves into their amplifiers and remarking on the duration and magnitude of ringing that is observed.