The introduction of high-throughput sequencing technologies has made transcriptome analyses of plant-pathogen interactions almost routine. Nevertheless, it is still challenging to obtain RNA from populations made up of two species. An RNA extraction method that worked well on free-living Caenorhabditis elegans failed when applied to isolated Globodera pallida J2 larva. Furthermore, alternative protocols that extracted RNA from free-living J2 larva produced less satisfactory results once the animals entered their hosts' roots. We have compared several extraction procedures to ascertain whether a single protocol was capable of recovering high-quality, high-molecular-weight RNA from newly hatched J2 larva as well as from larva embedded in roots of both potatoes (Solanum tuberosum L. cv. Desiree) and a very distantly related species, Solanum sisymbriifolium. Although it was possible to recover large amounts of RNA from J2 larvae using Proteinase K treatments, this protocol failed to yield high-quality nematode RNA from infected roots. By comparison, mechanical disruption procedures yielded lower amounts of RNA from infected roots, but what was recovered was of higher quality. We conclude that different extraction protocols need to be developed to sample mixed populations of organisms.