Environmental Engineering Reference
In-Depth Information
remnant, populations derived from seed or plants collected at a single site may serve as breeding
populations. In this regard, Cave-in-Rock, an accession deriving from one collection site, served
as the breeding population for 'Shawnee' (Vogel et al. 1996). Other cultivars, such as Pathfinder,
were developed by selection within broader germplasm pools created as composites or synthetics
from seed collected at multiple sites that define a target region of interest (Newell 1968b). Multisite
composites or synthetics are often thought to have a broader genetic base capable of sustaining
genetic gains for more generations of selection and to buffer new cultivars against the possibility
of pests or abiotic stresses. DNA marker diversity studies in switchgrass (Casler et al. 2007b;
Narasimhamoorthy et al. 2008) and selection experiments in another grass (Burton 1974) suggest
that this perception by breeders may not be accurate. Germplasm pools perceived to be either narrow
or broad in their genetic diversity are both capable of allowing sustained genetic gains. The polyploid
nature of switchgrass serves as a reservoir, storing large amounts of genetic variability.
Genetic variability exists within both broad and narrow gene pools of switchgrass, related to a
wide range of traits, including seedling tiller number (Smart et al. 2003a, 2003b, 2004); cell wall
composition and forage quality (Vogel et al. 1981; Godshalk et al. 1988); plant height, vigor, and
biomass production (Newell and Eberhart 1961; Talbert et al. 1983; Hopkins et al. 1993; Das et al. 2004;
Missaoui et al. 2005b; Rose et al. 2007); photoperiod-related traits such as earliness and phytomer
number (Van Esbroeck et al. 1998; Boe and Casler 2005; Casler 2005); and biotic or abiotic stress
tolerances (Hopkins and Taliaferro 1997; Vogel et al. 2002b; Gustafson et al. 2003). The risk associated
with selection for a trait for which little or no genetic variation exists within a population is illustrated
by P concentration in Alamo (Missaoui et al. 2005c). P uptake by switchgrass could be improved only
by selection for increased biomass yield because of lack of genetic variation for P concentration.
Vogel and Pedersen (1993) and Vogel and Burson (2004) reviewed breeding procedures used to
improve switchgrass. Most breeding programs heavily utilize spaced plants as the selection units, in
which seeds are germinated in the glasshouse and plants with 3-6 tillers are transplanted to the field
on spacings that range from 0.3 to 1.1 m. In most cases, random plants are selected from a bag of
seed representing the population to be improved. This is phenotypic selection, sometimes referred
to as restricted recurrent phenotypic selection (RRPS) which refers to a number of modifications
designed to improve efficiency and the rate of genetic gains (Vogel and Pedersen 1993). Some of
these restrictions or modifications include selection for seedling vigor in the glasshouse, removal of
some environmental variation in the field using a grid system, and intercrossing selected plants as
early and rapidly as possible using excised tillers in the glasshouse (Burton 1974).
One particular aspect of Burton's RRPS, planting extra seedlings in the glasshouse and conducting
some form of seedling selection before establishment of field nurseries, may be particularly useful
in switchgrass recurrent selection programs. Smart et al. (2003a) demonstrated that seedling tiller
number is moderately heritable, creating Cycle-2 populations with mean tiller numbers of 1.2 and 2.0
compared to the base population of 1.6 tillers per plant. Although seedling tiller number did not affect
establishment of switchgrass in the field (Smart et al. 2003b), differences in seedling tiller number
translated directly to differences in adult-plant tiller numbers (80 vs. 119 tillers per plant). Furthermore,
the single-tiller population had a 28% greater leaf elongation rate, 60% more mass per tiller, and 24%
higher biomass yield per plant (Smart et al. 2004). Selection of switchgrass seedlings with a single
tiller at a defined length of time postgermination appears to be an effective mechanism to improve the
efficiency of a field-based recurrent selection program. This is supported by additional studies of adult
plants that indicated cultivars with fewer tillers, but more phytomers per tiller and a higher proportion
of reproductive tillers, have the highest biomass yield potential (Boe and Casler 2005).
Family selection or genotypic selection methods have also been used for switchgrass improvement.
Family selection may involve the use of spaced plants and similar or identical selection protocols
as phenotypic selection. In this case, spaced plants are generally arranged in rows where each row
represents a family and plants within a row are generally half-sibs of each other, i.e., they each derive
from one maternal parent (Vogel and Pedersen 1993). The efficiency of family selection is greatly
enhanced by two-stage selection in which the best families are selected based on row means and the
