The backcross breeding method is often used to transfer recessive trai
ts controlled by one or a few genes from one pure line to another. Gui
delines are needed to help plant breeders to use backcrossing cost-eff
ectively. In a backcrossing program, a plant breeder may choose betwee
n methods with two or three stages per cycle. The two-stage method uti
lizes alternate backcrossing and selfing to identify BCiSj plants with
the recessive trait while the three-stage method uses two sequential
crosses followed by selfing. In the three-stage method, n(min), the mi
nimum number of plants required to recover at Least one or two BCiSj p
lants in the i(th) cycle with a given probability (1 - alpha), is obta
ined by growing only one BCiSij plant for each of n(min), BCiSo. Alter
native values for the number of BCiSo plants (n(b)) and BCiSj plants p
er BCiSo (n(s)) are present for alpha = 0.05 and alpha = 0.01 for trai
ts controlled by one, two, three, or four recessive genetic loci. A me
thod to compare the cost-effectiveness of alternative values is presen
ted for cases where the relative costs of crossing, selfing, and evalu
ation of S-j progeny are known. Unless time is the paramount concern,
it is only in cases where the ressive trait is controlled by a single
locus that it can be more cost-effective to make two sequential crosse
s to the recurrent parent before selfing than to cross and self and th
en only if the cost of evaluating BCiSj plants is high relative to the
cost of producing BCiSo plants. When using the three-stage backcross
method, the breeder can reduce the n(b) and increase n(s) if the cost
of evaluating BCiSj plants is low relative to the cost of producing BC
iSo plants.