After nearly a decade of field inventories in which we preserved voucher specimens of the small terrestrial mammals of Sulawesi, we combined qualitative and quantitative analyses of morphological traits with molecular phylogenetics to better understand the diversity of shrews (Soricidae: Crocidura) on the island. We examined the morphology of 1368 specimens and obtained extensive molecular data from many of them, including mitochondrial DNA sequences from 851 specimens, up to five nuclear exons from 657 specimens, and thousands of ultraconserved elements from 90 specimens. By iteratively testing species limits using distinct character datasets and appropriate taxon sampling, we found clear, mostly consistent evidence for the existence of 21 species of shrews on Sulawesi, only seven of which were previously recognized. We divide these 21 species into five morphogroups, provide emended diagnoses of the seven previously named species, and describe 14 new species. The Long-Tailed Group contains Crocidura caudipilosa, C. elongata, C. microelongata, new species, and C. quasielongata, new species; the Rhoditis Group contains C. rhoditis, C. pseudorhoditis, new species, C. australis, new species, and C. pallida, new species; the Small-Bodied Group contains C. lea, C. levicula, C. baletei, new species, C. mediocris, new species, C. parva, new species, and C. tenebrosa, new species; the Thick-Tailed Group contains C. brevicauda, new species and C. caudicrassa, new species; and the Ordinary Group contains C. musseri, C. nigripes, C. normalis, new species, C. ordinaria, new species, and C. solita, new species. Documenting these endemic species reveals a local radiation (20 of the 21 species are members of an endemic clade) in which elevational gradients played a prominent role in either promoting speciation, or at a minimum, fostering the cooccurrence of phenotypically similar species. As now understood, the species-level diversity of Crocidura on Sulawesi is nearly three times the known diversity of any other insular shrew fauna. This study highlights the fact that if we wish to understand the true extent of biodiversity on Earth, large-scale, vouchered organismal inventories followed up with thorough examinations of genetic, morphological, and geographic traits are sorely needed in montane tropical regions, even for purportedly well-studied groups such as mammals.