Survival rates of hatchery-reared and wild juveniles in field releases
did not differ significantly, although survival rates varied among release
trials (Fig. 3). For each paired release, AICc indicated that the best
exponential decline model was the fully-reduced model in which model
parameters did not differ between hatchery-reared and wild juveniles.
Further, the fully-reduced models were strongly supported by the data
in each case with Akaike weights of 0.65, 0.83, and 0.89 for releases in
Aberdeen Creek, Cadle Creek and Sheepshead Cove, respectively. Model
averaged MLEs of survival rates of juvenile cohorts averaged 8% among
all releases ( y0 = 0.08), but were variable among release sites:
Aberdeen Cove (y0=0.06), Cadle Creek (y0=0.00) and Sheepshead
Cove (y0=0.18; Fig. 3). Only 1.5% (3 out of 206) of all recaptured crabs
from the release in Sheepshead Cove were captured in the adjacent
control area, indicating a high degree of site fidelity for both cohorts at
this location. Each of the three individuals recaptured in the control area
were hatchery-reared; we captured no released wild crabs outside of
the release area.