Epistasis and genotype by environment interactions (GxE) are likely sources of the missing heritability for complex traits that have emerged from genome wide association studies (GWAS). In an effort to understand the role of gene by gene (GxG) and GxE interactions in the mapping of genotypes in to phenotypes, we have built a panel of 72 mitonuclear genotypes constructed from all pairwise combinations of 6 mtDNAs of D. melanogaster and D. simulans and 12 of the Drosophila Genetic Reference Panel (DGRP) strains. One prediction is that mitonuclear epistasis will scale with mtDNA sequence divergence. Another is that the importance of epistasis will vary across environments. We have quantified development time and other fitness traits on different isocaloric diets of yeast and sugar to assess the individual and joint contributions of mitochondrial, nuclear and dietary environment to these phenotypes. While certain DGRP lines support the mtDNA divergence prediction, the majority of nuclear backgrounds show no significant effect of mtDNA divergence on development time. However, a number of DGRP backgrounds show striking reversals of phenotype across different mtDNAs. The role of mitonuclear epistasis is most pronounced in the low protein diet with the slowest development. Notably, those DGRP backgrounds that demonstrate the most dramatic epistatic interactions with mtDNAs are the most responsive to dietary modification of development time. These observations raise the hypothesis that GxG and GxE interactions are not distinct phenomena but are related. This notion becomes intuitive if either epistatic partner lies in a pathway where an abiotic environment alters the genetic or cellular environment that governs the expression either gene. This joint mitochondrial and nuclear genomic approach manipulates a number of metabolic processes that can add predictability to GxG and GxE effects underlying the putative missing heritability for complex traits.