I would recommend to use quantitative RT-PCR instead of Northern blotting. It is much more sensitive and accurate. In addition to the HSP genes you will need a housekeeping gene which is used as a reference. Here you could select genes encoding tubulin or actin, for example.
Western blotting could be used to tell you the level(s) of translated protein as well; which is always a good thing to check in addition to qPCR. I agree with Christian that Northern blotting could/should be avoided here. Ultimately, it is the level of protein one is interested in, as in some (or many) cases, mRNA levels do not indicate the level or presence of the associated protein...
As mentioned qRT-PCR is probably a lot easier and if you check your amplicons on an agarose gel you can verify that your PCR products are the correct size and hence your primers are amplifying what you expect. If you want to use northern blotting this would have the advantage that you could visualise any splice variants of different mRNAs with alternative polyadenylation sites. Heat shock proteins are highly conserved so you can probably use probes against a different species if you are stuck, and you can probably get some useful pointers from a few simple sequence alignments. If decide to go for northerns you need to get very good at purifying RNA and preventing it from degrading. This means having a spotless bench, avoiding dust (which contains RNases found in human skin). Good quality water is a must and if you do not trust your water purification system you should DEPC-treat the water you use in your gel and samples. Agarose/MOPS/formaldehyde gels work well. If you are using a kit such as RNeasy, make sure you dry the column completely after the ethanol wash to remove any residual ethanol before eluting and adding sample buffer to your RNA. The exact probes you use will depend on the specific heat shock proteins you want to investigate.