278 
Fishery Bulletin 96(2), 1998 
o: 
n, 2-2 
= O 
s ^ r,s * 
( 8 ) 
The variable k t s represents ob- 
served kill for the i th observed 
day in stratum s, and a l s is the 
sample variance of the ob- 
served kill. Variables a? and D , 
are observed and total number 
of days of effort in the stratum, 
respectively. Estimates of over- 
all kill rate, r , and total inciden- 
tal kill, rh , across all strata, and 
variances, are then weighted 
averages: 
^s D sK 
D 
y T~\ 2 ~ 2 
-2 .9 -s ° r ,s , 
O = 5 ’ 
r D 2 
rh = Dr , 
-2 n 2-2 
- D <j r , 
(9) 
( 10 ) 
(ID 
( 12 ) 
where D is total number of days 
of effort during the year. 
Results 
Driftnet fishery 
Figure 1 
Approximate location of observed sets in the driftnet fishery during the period July 
1990-December 1995. 
During the observed period, the 
driftnet fishery was subject to 
area and time restrictions in- 
side 370.4 km (200 nautical 
miles [n mi] ) for effort that took 
place largely (80% of seasonal effort) from 15 August 
to 31 January. Between 1 February and 30 April, fish- 
ing effort was restricted to locations farther than 
370.4 km offshore. From 1 May through 14 August 
(15 July for 1990-92) effort was required to be lo- 
cated farther than 138.9 km (75 n mi) offshore. For 
the remainder of the year, fishing was allowed within 
138.9 km of shore. Since the beginning of the NMFS 
observer program in 1990, estimated total effort-days 
for the fourth quarter increased each year except for 
1994. Combined effort for third and fourth quarters 
was consistently over 3,500 sets each year (Table 2). 
Even though there was effort along the entire Cali- 
fornia coast, the highest concentration of observed 
sets was south of 33°N (Fig. 1). Observer coverage in 
this fishery typically ranged between 10% and 15% 
of estimated total yearly effort for most areas and 
quarters. (Exceptions as low as 2% and as high as 
25% were noted for some quarters and areas.) Bias 
of a ratio estimate is typically a problem in the case 
of small sample sizes such as may occur in a per- 
fishing block situation. Because sample size was gen- 
erally quite large and stratification was not used, 
estimates were not corrected for bias. During the 
study period, July 1990-December 1995, nine (2%) 
of the 472 marine mammals observed entangled were 
released alive (one unidentified sea lion, one com- 
mon dolphin, Delphinus spp., two California sea li- 
ons, one humpback whale, Megaptera novaeangliae, 
one Cuvier’s beaked whale, Ziphius cavirostris, and 
, 
