200 words) 1 In monocots other than maize and rice, the repertoire and diversity of microRNAs (miRNAs) and 2 the populations of phased, secondary, small interfering RNAs (phasiRNAs) are poorly 3 characterized. To remedy this, we sequenced small RNAs from vegetative and dissected 4 inflorescence tissue in 28 phylogenetically diverse monocots and from several early-diverging 5 angiosperm lineages, as well as publicly available data from 10 additional monocot species. We 6 annotated miRNAs, siRNAs and phasiRNAs across the monocot phylogeny, identifying miRNAs 7 apparently lost or gained in the grasses relative to other monocot families, as well as a number 8 of tRNA fragments misannotated as miRNAs. Using our miRNA database cleaned of these 9 misannotations, we identified conservation at the 8 th , 9 th , 19 th and 3' end positions that we 10 hypothesize are signatures of selection for processing, targeting, or Argonaute sorting. We show 11 that 21-nt reproductive phasiRNAs are far more numerous in grass genomes than other 12 monocots. Based on sequenced monocot genomes and transcriptomes, DICER-LIKE5 (DCL5), 13 important to 24-nt phasiRNA biogenesis, likely originated via gene duplication before the 14 diversification of the grasses. This curated database of phylogenetically diverse monocot miRNAs, 15 siRNAs, and phasiRNAs represents a large collection of data that should facilitate continued 16 exploration of small RNA diversification in flowering plants. 17 18 19 20 55 2018). Rice, Brachypodium, and maize are the most studied of the grasses, with miRNAs 56 characterized using varying genotypes, tissue types, growth and stress conditions (Jeong et al., 57 2011; Zhang et al., 2009). With the major goal of assessing the diversity and origins of miRNAs 58 4in monocots, we analyzed sRNA data from 38 phylogenetically diverse monocots, spanning 59 orders from the Acorales to the Zingiberales. We described sRNA size classes, miRNA 60 conservation, divergence, sequence variability, 5' and 3' end nucleotide preferences, and single 61 nucleotide sequence profile characterizing positional biases and providing novel insights within 62 the miRNA sequences. We performed comparative analysis of miR2118 and miR2275 and their 63 long non-coding RNA (lncRNA) targets in monocots relative to other flowering plants, 64 demonstrating their presence and absence in these species. We found that both miR2118 and 65 miR2275 are conserved across diverse monocot species, and are present in vegetative tissues but 66 are found at high abundances predominantly in inflorescence tissues. The 21-and 24-nt PHAS 67 loci are most numerous in the genomes of grasses, relative to other monocots, and are similarly 68 most abundant in inflorescence tissues. Fewer PHAS loci were identified in non-grass monocots. 69 Overall, our study provides a deep comparative analysis of sRNAs in monocots, including a refined 70 database of monocot miRNAs. 71 72 Results 73 74Sequencing from diverse monocots demonstrates atypically abundant 22-nt siRNAs 75 We c...