Procedures for various sources of heteroskedasticity in RCA(1) process are discussed , asymptotic distribution is considered and empirical powers of test statistics are compared numerically