Dear Rachael Vanessa Wilbourn,

First you run the qRT-PCR product on a DNA-PAGE depending on your  amplicon size and also load species A sample as a control. It will give you the idea about  differences in both species. Depending on that you should proceed like decreasing or increasing primer concentration and dilution of cDNA. Sometimes cDNA of few mRNA are not very stable and vary among species, so you can try qRT-PCR immediately after cDNA synthesis. I suggest you before going to any conclusion first see your product on gel.

It looks like species A has a product that just melts more uniformly.  A shouldered peak can sometimes mean you have more than one product.  If you have just one product, then it is simply large enough or has different enough GC content across it's length to melt in a complex manner.  Totally normal and I've seen it before.

I agree with Kumar, that PAGE of amplicons can give a hint about difference between samples.

One another possibility is a presence of polymorphisms in amplified sequence in sample B, that gives several products with a slighty different melting temperature. It can be revealed by sequencing of amplicons.

I agree with all of the three answers above. I would also start by running a gel to see if you have one or multiple products. If the products from species A and the others are identical length on gel, I would suggest to sequence the amplicons (maybe use primers with tags for this purpose to be able to see all bases in the amplicon). Another thing that you could do right away is to verify that your amplicon does not have significant secondary structures (I normally use unaFold for this). If amplicon lenghts and sequences are identical according to your evaluation, secondary structure and different salt levels in the solutions could explain why some samples exhibit the secondary structure, while species A does not.