Stewart; Defining migration rates of Porophrys vetulus 



475 



Table 3 



Specific structure of the movement parameter matrices (P) for the 14 models explored in this analysis. One-season models (1-5) 

 allowed movement to occur only once per year in the month shown, two-season models (6-9) allowed movement in two months 

 per year, and monthly movement models (10-14) allowed movement in >2 months per year, as specified. 



Model 



Movement parameter matrices 



One-season: two movement parameters (north and south in the month indicated) 



1 October 



2 November 



3 January 



4 May 



5 June 



Two-seasons: four movement parameters (north and south in the months indicated) 



6 October, May 



7 October, June 



8 November, May 



9 November, June 



Monthly movement 



10 One parameter, all months, north=south 



11 Two parameters, all months, north and south 



12 One parameter, movement from May through November only 



13 Two parameters, movement from May through November only, north and south 



14 Four parameters, movement from May through October, north and south, and movement in November, north 

 and south 



zero and all tags were recaptured immediately, then 

 the minimum possible reporting rate would be the ratio 

 of tags returned to total releases; this value, 0.21, is 

 therefore a logical lower bound on reporting rate. These 

 bounds are comparable to the range of reporting rates 

 observed for other species (Gaertner and Hallier, 2004). 

 Long-term tag-induced mortality (included in tag loss 

 in this model) may be twice that of natural mortality 

 (Manzer, 1952). Tag loss was also reported to be high 



because of corrosion of the tag pins for English sole 

 tagged in a similar fashion over roughly the same period 

 (Forrester and Ketchen, 1955). The proportion of tags 

 lost could not have been greater than 0.78 because 0.22 

 of the tags were recovered; this value (0.22) is greater 

 than that used to calculate the minimum bound for 

 reporting rate because it includes those tags recovered 

 without full reporting of location information. The value 

 of 0.78 for the proportion of tags lost corresponds to a 



