Skip to main content
. 1998 Feb;64(2):659–664. doi: 10.1128/aem.64.2.659-664.1998

TABLE 2.

Coefficients determined for response models

Parameter Response (y)a
EPS concn (mg/ml) CFU/ml Time to endpoint (h)b
Constant (β0) 235.82 1.8 × 108 12.8
Temp (β1) −4.64c 1.8 × 107c −1.7d
pH (β2)  NIe 1.75 × 108d −14.5d
Casitone (β3) 2.34c NI −0.3c
Temp × pH (β4) NI 3.4 × 107f 0.9f
Temp × Casitone (β5) −0.57c NI NI
pH × Casitone (β6) NI 6.3 × 106 −0.2
Temp27) −2.17f NI 0.2f
pH28) −20.91 7.5 × 107 9.9d
Casitone (β9) 0.24 9.4 × 105 0.05c
r2g 0.838 0.923 0.983
Ph 0.011 0.0005 0.0000
a

Y = β0 + β1(T − 40) + β2(pH − 5) + β3(casitone − 20) + β4(T − 40) (pH − 5) + β5(T − 40)(casitone − 20) + β6(pH − 5)(casitone − 20) + β7(T − 40)2 + β8(pH − 5)2 + β9(casitone − 20)2, where T is the temperature of fermentation (°C), pH is the pH of the medium during fermentation, and casitone is grams of Bacto-casitone per liter of growth medium. 

b

Time at which 90% of the glucose in the growth medium was utilized by the organism. Time to endpoint had a significant lack of fit (P < 0.05) with the response surface model. 

c

Significantly different at P < 0.05. 

d

Significantly different at P < 0.001. 

e

NI, not included. Inclusion of this coefficient in the model did not improve the fit. 

f

Significantly different at P < 0.01. 

g

Regression coefficient of the model to predict the particular response variable. 

h

P value for the model to predict the particular response variable.