Switchgrass is cross-pollinated by wind-dispersed pollen and has significant genetic variation within and among populations. Switchgrass is a polyploid with a basic chromosome number of x=9 (Gould 1975). Switchgrass has two genetically and morphologically distinct ecotypes, generally referred to as upland and lowland ecotypes due to the landscape position they historically occupied (Vogel et al. 2011). Upland ecotypes originated in upland areas that were not subject to flooding and often prone to drought, whereas lowland ecotypes occurred in flood plains and riparian zones subject to occasional flooding (Casler et al. 2012). Forage type switchgrass cultivars historically have been upland ecotypes with fine stems and reduced plant height, more stems per plant, and more decumbent leaves (Fig. 1). Lowland ecotypes

have upright leaves with a bluish tint and a later heading date than uplands (Casler et al. 2012). Lowland ecotypes are tetraploids (2n=4x=36), whereas the upland ecotypes can be either tetraploids or octoploids (2n=8x=72; Vogel et al. 2011). Upland and lowland tetraploids can be crossed (Martinez-Reyna et al. 2001) and can produce high-yielding F1 hybrids (Vogel and Mitchell

2008) . Upland x lowland hybrids averaged 30 to 38% high-parent heterosis for aboveground biomass (Vogel and Mitchell 2008). Chapters 5, 6, and 7 address switchgrass breeding and genetics in more detail.

Canopy architectural traits such as morphologic development, phenology, tiller density, and leaf area index (LAI) are in a continual state of flux (Moore and Moser 1995; Redfearn et al. 1997, Mitchell and Moser

2000) . Switchgrass is photoperiod sensitive and requires shortening day length for floral induction, resulting in switchgrass morphology being strongly correlated to day of the year (DOY) and growing degree days (GDD) (Mitchell et al. 1997). Grassini et al. (2009) developed a model for predicting switchgrass growth and development. Switchgrass has a determinate growth habit where most vegetative growth terminates with inflorescence development (Mitchell et al. 1997). After flowering, tillers advance to the seed ripening stages, growth stops, and tiller senescence

occurs. In switchgrass swards in eastern Nebraska, there were no vegetative tillers present by DOY 196 and 100% of the tillers had elevated apical meristems (Mitchell et al. 1998). As phenology advanced, tiller density declined by an average of 9.4 tillers m-2 d1 and an average tiller density of 1525 tillers m-2 (Mitchell et al. 1998). In Texas, tiller density and mass increased as row width increased and tiller mass increased as N fertility increased (Muir et al. 2001). Switchgrass LAI increased as phenology advanced and varied across years with maximum LAI ranging from 4.9 to 7.7, with at least 95% of the variation in LAI explained by dOy (Mitchell et al. 1998). The predictability of switchgrass development in response to DOY and GDD indicates switchgrass management recommendations for adapted cultivars may be made based on DOY within a region (Mitchell et al. 1997). It is likely that bioenergy specific switchgrass strains will have similar responses. One concern with feedstock delivered to the biorefinery is product consistency (Schmer et al. 2012). Targeting harvest date based on phenologic stage is one mechanism by which feedstock uniformity could be managed. It is likely that harvesting after senescence will minimize the variability in feedstock composition and provide a more uniform product to the biorefinery.