PCR preferential amplification, question

[Ana Rakitin]

I would appreciate very much if anyone could help me out with a problem of 
PCR preferential amplification of smaller alleles.
I am using PCR to amplify a minisatellite region of larval fish DNA. The
allele sizes vary between 700 and 1300 bp. The PCR conditions are 
75 mM Tris-HCl (pH 9.0) 20 mM (NH4)2SO4, 2 mM MgCl2, dNTP 250 uM each, 
primers 0.5 uM each, approximately 200 ng template in a 20 ul reaction.  
The thermal cycling starts with 10 min denaturing at 95 ("hot start"),  
suspended at 85 while adding 0.5 U of Taq polymerase to each sample and 
followed by 30 cycles of 1 min at 95, 1 min at annT (57), 2 min 
extension (at 72) with a final 10 min extension time. I visualize the 
product after agarose gel electrophoresis and ethidium bromide stain. 
The problem I encountered is that for heterozygote individuals with 
alleles differing by 200 or more bp the larger allele appears as a very 
fait band or does not show up at all. When a small allele is not present 
(i.e. individuals homozygote for a large allele) the large alleles amplify 
well and appear as a strong band on the gel. I have already tried incresing 
extension time up to 4 min per cycle, changing template concentration, 
increasing Taq concentration but preferential amplification remains a 
problem.
Thank you in advance for your help.

Ana Rakitin