variances were calculated for each vessel type 

 over the entire survey area. The log mean and 

 log variance for each vessel type was trans- 

 formed, and the 959^ confidence interval about 

 the retransformed mean was calculated accord- 

 ing to the following formulae lAitchison and 

 Brown 1957): 



and the 959^ confidence interval about the esti- 

 mate were calculated as follows: 



c = exp 



(,.,(.1^).,,,) 



where c = mean catch per vessel 



L = mean natural log of catch per vessel 

 V (L) = estimated variance of natural logs 

 to catch per vessel 

 n = number of vessels interviewed 



The variance of c, v (c), is approximated by: 



c 

 n 



{v(L) + y2(v(L))^} 



and 



c±1.96\ v(c) 



is a 957ir confidence interval about c. 



Estimated Total Catches 



The mean catch per inlet per day, its variance 

 assuming s and c were independent, and 95'7c 

 confidence intervals were calculated for each ves- 

 sel type using the following formulae: 



sc = s X c 



where 'sc = mean catch per inlet per day 

 s — mean trips per inlet per day 

 c = mean catch per vessel 



V (sc) = (s)2 V (c) + (c)2 V is) + V (c) V (s) 



where v isic) = estimated variance of catch per 

 inlet per day 



V (s) = estimated variance of trips per 



inlet per day 



V (c) = estimated variance of catch per 



vessel 



CI = sc ±1. 96 Vv{sc) 

 where CI = 95% confidence interval about sc 



The total estimated catch iTSC) per vessel type 

 802 



where 



TSC = [sc ± 1.96 Vv(sc)] x 33NI 



TSC = total estimated catch 

 NI = number of inlets where vessel type 



occurred 

 33 = number of days in fishing season 



The total estimated catch and confidence in- 

 terval for the total survey area and all vessel 

 types were determined by summing the estimated 

 catches and extracting the square root of the sum 

 of the squares of the variances of all four vessel 

 types. 



Lengths, Weights, and Age Compositicjn of Catches 



A total of 2,778 Atlantic mackerel were mea- 

 sured to the nearest centimeter fork length to 

 determine the length frequencies of the catch. 

 Each length was converted to a weight using the 

 formula logio weight = -5.2314 + 3.0796 logio 

 length (Wilk et al. 1978), and a mean weight was 

 calculated for all vessel types. The mean weight 

 was multiplied by the total estimated number 

 caught to determine the total weight of the 

 catch. 



For age composition analysis, Atlantic mack- 

 erel were obtained in April from recreational and 

 commercial fishermen fishing primarily along the 

 New Jersey coast and transported to the NEFC 

 Sandy Hook Laboratory where they were mea- 

 sured to the nearest centimeter fork length and 

 sexed. The heads were removed, frozen, and sent 

 to the NEFC Woods Hole Laboratory, NMFS, for 

 otolith removal and aging. Aging was accom- 

 plished by placing intact otoliths in black trays, 

 imbedding them in clear epoxy resin, and count- 

 ing annular rings using reflected light at 25-75 x 

 magnification under a binocular microscope. The 

 number of fish from the length-frequency sample 

 of 2,778 measured at each centimeter length was 

 multiplied by the percentage age composition at 

 that length increment to determine the number of 

 fish caught in each age-group at each centimeter 

 increment. The numbers at each age were summed 

 from all length increments and divided by the 

 total number of fish measured to determine the 

 percentage composition of each age in the recrea- 

 tional catch. The percentage composition at each 

 age was multiplied by the total estimated Middle 



