Estimator performance under constrained history of asymmetrical soft vicariance (True Z = 0; True E (( θ
) < 0.05; stage 1 model). Each panel depicts a frequency histogram of 250 mode estimates of Z (A and B) and E((θ
)2) (C and D) under a constrained history of asymmetrical soft vicariance where Z = 0 and E((θ
)2) < 0.05. For all mode estimates, tolerance was 0.001 (2,000 accepted draws). Mode estimates for panels (A) and (C) are obtained using the local linear regression algorithm, whereas the mode estimates for panels (B) and (D) are obtained using the raw 2,000 accepted values.