any problems with RT-PCR for large amplicon sizes? - very strange amplification curves (Jun/19/2007 )

I've recently designed a number of primer pairs for bacterial virulence genes that have amplicon sizes over 200bp (-350bp). Standard PCR reactions worked perfectly, and I could see a clear product that can be easily distinguished from any primer dimer. But my RT-PCR amplification curves were a complete mess- some rose sharply in the first 10 cycles then zig-zagged down, some didnt show any fluorescence at all, and some even dipped below the control level (no template DNA added). Someone once warned me about using larger amplicon sizes with real-time- does anyone know why?? or why this is happening at all?

did you change your extension times to match the product sizes you are expecting?

I don't think 350bp product on SYBR green should be trouble. I've seen real-time amplicons up to 500 bp.
This is more a concern of a probe based system. I wouldn't bet on a product length in this case.