“…A common approach to handling such data has been to regress the log of the number recaptured (usually after adding one to eliminate problems with a 0 recapture) as a function of time in days after release (Gillies 1961, Gillies and Wilkes 1965, Lindquist et al 1967, Wada et al 1969, Dow 1971, McDonald 1977, Seawright et al 1977, Nelson et al 1978, Reisen et al 1978, Reisen and Aslamkhan 1979, Reisen et al 1980, Nayar 1981, Rawlings et al 1981, Nayar 1982, Haramis and Foster 1983, Linthicum et al 1985, Tripis & Hausermann 1986, Walker et al 1987, Rodriguez et al 1992, Day et al 1994, Constantini et al 1996, Morrison et al 1999. The survival rate purportedly is estimated by exponentiating the resulting slope, and survival rates for different cohorts are compared using a t-test for equal slopes for two linear regressions (Haramis andFoster 1983, Morrison et al 1999).…”