Lo et al.: Application of the continuous egg sampler to estimation of the daily egg production of Sardinops sogax 
555 
Table 1 
Total number of samples and positive samples of sardine eggs sampled in CUFES and CalVET surveys. The range and mean of 
duration (minutes) for CUFES collections for 1994, 1996, and 1997 surveys. 
1994 
(18 Apr— 11 May) 
1996 pilot CUFES survey 
1997 DEPM survey 
leg 1 
(15-21 Mar) 
leg 2 
(21 Mar-6 Apr) 
leg 1 
(11 Mar-27 Mar) 
leg 2 
(28 Mar-7 Apr) 
CUFES 7 (positive) 

1396 
905 
896 
331 
(889) 
(568) 
(550) 
(137) 
Duration mean range (min) 
— 
3.5 
18.13 
30 
26.9 
2-5 
3-35 
1-54 
1-34 
CalVET 1 2 (positive) 
684 
91 
— 
141 
— 
(74) 
(66) 
— 
(102) 
— 
Survey area (km 2 ) 
380,175 
157,000 
174,196 
(253,850 in USA) 
1 Total collection in leg 1, 1996, was 1437. The first 41 tows were experimental. 
2 CalVET surveys in 1996 and 1997 were taken in the high-density stratum; in 1994 they were taken in a fixed grid over the whole survey area. 
analysis of the 1996 cruise was used to develop an adap- 
tive allocation survey design, similar to that proposed by 
Thompson et al. ( 1992), to estimate the biomass of sardine 
that incorporates a CUFES into the DEPM. Using the 
adaptive allocation survey design, we conducted a DEPM 
survey for Pacific sardine with a CUFES in 1997, and the 
precision of our estimate of daily production of eggs was 
compared with the precision of the estimate provided by 
a DEPM survey carried out in 1994 under standard meth- 
ods (Lo et al., 1996). We also considered the accuracy of 
shipboard counts of eggs collected in a CUFES in relation 
to counts of preserved samples made after cruise end. We 
present our results in chronological order, beginning with 
those from the 1996 pilot survey which provided the ratio- 
nal for the new survey design in 1997; we next describe 
the new survey design, and end with a comparison of the 
1997 CUFES and DEPM survey with the conventional 
DEPM survey in 1994. 
Survey data and spatial models 
Survey data 
The data used in our study were taken from three ichthyo- 
plankton surveys (Table 1); 
1 A pilot CUFES cruise in 1996. This cruise consisted of 
leg 1 (during which both a CUFES was used and full- 
water-column tows were taken ) and leg 2 (during which 
only a CUFES was used to survey the large geographic 
area of spawning sardine (Fig. 1); 
2 A DEPM survey in 1997. This survey for Pacific sardine 
employed a new survey design with the CUFES and the 
California vertical tow (CalVET, see below) (Smith et 
al., 1985) (Fig. 2). The allocation of CalVETs was deter- 
mined by the egg density observed from the CUFES. 
3 Results of the 1994 DEPM survey off California, U.S., 
and Baja California, Mexico (Lo et al., 1996). This was 
a conventional fixed-grid DEPM survey employing only 
CalVETs. This cruise was used as a standard for com- 
paring DEPM surveys with and without the CUFES. 
The CalVET net consisted of 150-pm nylon netting. The 
diameter of the CalVET net frame was 25 cm; the tow was 
lowered to a depth of 70 m and was retrieved vertically. 
The CUFES was installed midship on the NOAA vessel 
Dcwid Starr Jordan onto the intake pipe over the side of 
the vessel; it extended 3 m below the water surface (see il- 
lustration in Checkley et al., 1997). Eggs were sieved from 
the water flow with the 500-pm nylon mesh of the CUFES 
concentrator. 
The density of eggs taken in the CalVET net was ex- 
pressed as the number of eggs/0.05 m 2 of sea surface wa- 
ter, a standard procedure in the DEPM, where 0.05 m 2 is 
the area of the mouth opening of the CalVET net. All eggs 
taken in the CalVET samples, regardless of survey, were 
counted and staged in the laboratory. The density of eggs 
taken in the CUFES was expressed as the number of eggs 
taken per minute. The interval over which eggs accumu- 
lated in CUFES samples varied depending on their abun- 
dance. When abundance was low, samples were collected 
over intervals of 0.5 h equivalent to 7.4 km or 4 nmi on the 
transect line; when abundance was high, they were col- 
lected over intervals of at least 1 minute (0.24 km or 0.13 
nmi). All CUFES samples were counted at sea, preserved, 
and recounted in the laboratory. All eggs taken by the 
CUFES in the pilot survey, but not the subsequent DEPM 
survey, were staged in the laboratory. We used the system 
detailed in Lo et al. ( 1996) and grouped eggs by their ages 
into half-day age classes for egg mortality computation 
and one-day age classes for spatial statistical analysis. 
In our study, we considered the estimates of only the 
daily production of eggs, P 0 , one of the key parameters in 
