Consensus on the optimal high-throughput sequencing (HTS) approach to examine biodiversity inmixed terrestrial arthropod samples has not been reached. Metatranscriptomics could increase the pro-portion of taxonomically informative mitochondrial reads in HTS outputs but has not been investigatedfor terrestrial arthropod samples. We compared the efficiency of 16S rRNA metabarcoding, metagenom-ics and metatranscriptomics for detecting species in a mixed terrestrial arthropod sample (pooled DNA/RNA from 38 taxa). 16S rRNA metabarcoding and nuclear rRNA-depleted metatranscriptomics had thehighest detection rate with 97% of input species detected. Based on cytochrome c oxidase I, metage-nomics had the highest detection rate with 82% of input species detected, but metatranscriptomicsproduced a larger proportion of reads matching (Sanger) reference sequences. Metatranscriptomicswith nuclear rRNA depletion may offer advantages over metabarcoding through reducing the numberof spurious operational taxonomic units while retaining high detection rates, and offers natural enrich-ment of mitochondrial sequences which may enable increased species detection rates compared withmetagenomics.