C. Revellin et al., SAMPLING TO DETERMINE NODULE OCCUPANCY OF SOYBEAN (GLYCINE-MAX (L) MERR.) BY BRADYRHIZOBIUM-JAPONICUM STRAINS, Canadian journal of microbiology, 44(8), 1998, pp. 795-799
The confidence interval of the percentage of soybean (Glycine max. (L.
) Merr.) nodule occupancy by two strains of Bradyrhizobium japonicum w
as studied using samples of nodules per plant varying by multiples of
8 (8-56). These samples were recovered from all of the nodules of four
sets of 21-24 plants grown in soil colonized only by these strains an
d typed using strain-specific sera. Approximate confidence intervals (
ACI) were estimated (from the ANOVA model) using the formula, length o
f ACI = 4 [(0.040/P) + (0.278/PN)](1/2), where P is the number of samp
led plants and N is the number of nodules characterised for each plant
. The estimated ACI decreased when the number of plants sampled increa
sed and decreased to a lesser extent when the number of nodules typed
per plant increased. It was possible to define a plant and nodule samp
ling strategy to minimize the labor and time required to get acceptabl
e ACI values. Under our experimental conditions, when 192 nodules were
typed, combinations of 24 plants x 8 nodules per plant and 32 plants
x 6 nodules per plant were the most satisfactory, corresponding to cal
culated ACI of 22.3 and 20.8%, respectively.