Mycoheterotrophic plants frequently display drastic reduction in genomic and morphological features, making them attractive and charismatic subjects for systematic and evolutionary study. These same attributes have historically presented challenges for their taxonomic placement, which has complicated conservation efforts in these often rare or endangered species. Further, mycoheterotrophs may serve as useful models for testing more general hypotheses on the evolution of parasitism; specifically, of interest are potential tradeoffs between extreme fungal host specificity, morphology and reproductive biology, physiology, and the maintenance of genomic variation and stability. Combining data from morphology, genomics, physiology, and ecology provides a holistic representation of variation for evolutionary biology, integrative taxonomy, and defining conservation units of these species. Here I use two leafless, fully mycoheterotrophic species—Cephalanthera austiniae (Epidendroideae: Neottieae) and Corallorhiza striata (Epidendroideae: Epidendreae: Calypsoinae)—as models for implementing such integrative analyses. The Corallorhiza striata species complex (loosely, the striped coralroots) comprises three species: the rare Mexican and eastern North American endemics Corallorhiza involuta and C. bentleyi (respectively), and the widespread, highly variable C. striata. The latter is distributed from southern Mexico to Canada, and comprises two named varieties: striata, in the northern Rocky Mountains east to Newfoundland; and vreelandii, from Mexico to the southern Rocky Mountains, with outliers as far north as Montana, USA. Cephalanthera austiniae (the phantom orchid) is distributed through the Coast Ranges and Sierra Nevada of California north to southwestern Canada, with disjunct populations in southeastern Washington and western Idaho, USA. In the Corallorhiza striata complex, integrative analysis of complete plastid genomes, nuclear SNPs, fungal host ITS sequences, and abiotic niche data clearly differentiate C. bentleyi, C. involuta, and the widespread C. striata. Further analyses within the latter reveal divergence among four entities: 1) C. striata var. striata, 2) C. striata var. vreelandii, 3) populations from the Sierra Nevada (California, USA), and 4) populations from the Coast Ranges and Cascades (California, Oregon, USA). Each comprises a strongly supported clade based on >27,000 SNPs generated via ISSRseq. However, there is a lack of fixed differences in “extended phenotype” data (i.e. morphology, niche, fungal associates) among these four entities, as required by the “lineage + role” species concept applied here, so they are best described at the infraspecific level. Similar integrative analyses of plastid genomes, nuclear SNPs, morphology, and fungal associates are being conducted in Cephalanthera austiniae, which is considered imperiled in British Columbia, vulnerable in ID, but has no current status in CA, OR, or WA. By analyzing multiple data streams for both species, with emphasis on potential adaptive variation, I hope to deepen our understanding of the conservation status in both species, and extend this framework more broadly in other mycoheterotrophs.