A lfalfa, a perennial forage crop, experiences seasonal changes in growth patterns and morphology in the temperate regions of the world. Fall dormancy, referring to the characteristic growth reduction and decumbent shoot orientation of certain genotypes in autumn, typically occurs in late summer and early autumn as temperature declines and photoperiod shortens (Castonguay et al., 2006;McKenzie et al., 1988). For practical purposes, the dormancy level of alfalfa cultivars is ABSTRACT Alfalfa (Medicago sativa L.) is a widely planted perennial forage crop. Fall dormancy is generally negatively correlated with winter injury in alfalfa. To understand the genetic basis of the two traits, we identified quantitative trait loci (QTL) controlling autumn growth and winter injury using a tetraploid alfalfa F 1 population. In total, 601 marker alleles were scored from 78 restriction fragment length polymorphism (rFLp), 123 simple-sequence repeat (SSr), and 48 single nucleotide polymorphism (SNp) markers. Linkage maps were constructed for each parent separately. Both maps contained eight linkage groups (LGs), with a length of 898 cM for WISFAL-6 and 845 cM for ABI408. Using interval mapping, we identified 15 QTL from an across-environment analysis and 71 QTL within individual environments for autumn plant height; winter injury; and autumn shoot, crown, and root biomass across four Iowa environments. of the 71 QTL, 42 were identified at 18 chromosomal locations that were identified in multiple environments for the same trait. possible pleiotropic QTL that contributed to dry weight of shoot, crown, and taproot were found, which partially explained the observed genetic correlations between those traits. However, few QTL were related to both autumn plant height and winter injury, supporting the observation of no genetic correlation between the two traits in this study. These results indicated that the two traits could be manipulated independently and, possibly, efficiently improved using marker-assisted selection. Because most QTL identified in this study were mapped to intervals of at least 10 cM, validation and localization in additional populations is needed to facilitate application of marker-assisted selection.