Question 1152710
Assuming I can use normality, since otherwise the problem can't be done without knowing the skew. 
z=(xbar-35)/9/sqrt(42)
at least 39 for sample mean  
z>(4*sqrt(42)/9)= 2.88
that probability is 0.0020

between 33 and 39 
z is -2*sqrt42/9 for $33, or -1.44
probability z is between -1.44 and 2.88 is 0.9231

half-interval for 98%CI is z=2.33 and the interval is z sigma/sqrt(n)=3.24
98% sample means will be ($31.76, $38.24)