well i  think u should go for separation technique u must divide it into different samples then go for extraction 

You must extract DNA from single samples of insects. You can use a leg  grinding 50 mg tissue in a 1.5 ml micro-centrifuge tube with a pestle and add 180 μl buffer ATL. Then add 20 μl of proteinase K and vortex and incubate at 56˚C to lyse the insect tissues completely for 1 - 3 h. After incubation the lysate appear viscous. In next step add a solution (200 μl Buffer AL + 200 μl 96% - 100% ethanol) vortex it to homogenize the solution. Then pipet the mixture in DNeasy Mini spin column which is placed in a 2 ml collection tube, and centrifuge for 1 min at a speed of 8000 rpm. After this step discard the collection tube and flow-through and again place the DNeasy Mini spin column in a new tube and add 500 μl Buffer AW1 and repeat the previous step of centrifugation  at 14,000 rpm to dry the DNeasy membrane.

I think You should use spin filter column multiple times with same sample.

Means Use more no. of samples in extraction and filter it multiple time after lysis with lesser volume than pull it up. I have tried it and it worked well.

Hi Brian,

You're always going to have some sort of bias, whether you sub sample the homogenised tissue or extract it all and then pool it together and then PCR.   I think the important thing ( and perhaps the only thing you can do) is to note down these biases.  As long as you are aware of things like the effect of larger organisms providing more DNA to the pool, and depending on what your end questions are, then you have done what you can.  

You could always do a set of experiments to try and ascertain just how bad the bias is- something like using a known amount of DNA from a known number of individuals and PCRing it to see if the results at the end match what you know you put in.  Try out different extraction methods, different pools etc.