how do I eliminate excess levels in lattice contourplot key / legend?

I'm working on a contourplot( ) graph, the subject of several previous
posts to this list. The contours express probabilities from 0 - 1, but the
key that is automatically generated by contourplot pads the probability
scale with values that are impossible (i.e, range goes from -0.2 to 1.2),
which will be confusing to those who view the plot. How can I get the
scale to just run from 0 - 1?

I have toiled mightily in the depths of the Murrell & Sarkar graphics
books. Any pointers will be gratefully received.

--Seth

Minimal, commented code:

library(lattice)

model <- function(a,b,c,d,e, f, X1,X2)          # model function for
contour plot
{J <- a + (b*X1) + (c*X2) + (d*X1*X2) + e*(X1^2) + f*(X2^2)
 pp <- exp(J)/(1+exp(J))
 return(pp)}

g <- expand.grid(X1= seq(0.3,0.9,0.01), X2 = seq(0.3,1, 0.01))
## create x and y data matrix

g$z <- model(-29, -14, 52, 80, -3, -56, g$X1, g$X2)
## Create variable z using model and g data frame

contourplot(z ~ X1 * X2,
data = g,
region = TRUE,         # Add color to background
cuts = 5            # Number of contour/color intervals.
)