Lo et al.: Application of the continuous egg sampler to estimation of the daily egg production of Sardinops sagax 
569 
(12 nmi) apart, would ensure uncorrelated samples and 
provide the unbiased estimates of abundance of eggs in 
each development stage of sardine eggs. 
Use of yolksac larvae 
Although yolksac larvae have been used to estimate P 0 
with the DEPM method (Lo, 1985; 1986; Lo et al., 1996; 
Hunter and Lo, 1997), but are not a requirement for using 
the CUFES, we included yolksac larvae to estimate P 0 
because the mortality rate of eggs and yolksac larvae are 
similar for northern anchovy (Lo, 1985, 1986) and because 
the development of early stages of anchovy and sardine is 
similar (Ahlstrom, 1943; Zweifel and Lasker, 1976; Moser 
and Alhstrom, 1985; Lo et al., 1996). Our threshold for 
taking CalVET samples, 2 eggs per minute, generated 102 
positive CalVETs (72% were positive). This number is far 
fewer positive tows than the number needed to be assured 
a significant slope for the regression of numbers of eggs on 
their ages. In anchovy, where the eggs are less patchy than 
sardine, about 500 positive tows were required to assure a 
CV of the estimate of mortality rate close to 0.6 (Lo, 1997). 
To obtain a significant slope in the 1997 survey, we used 
the number and average age of yolksac larvae (adjusted 
for observed mean temperature) as well as staged eggs. 
This introduced a potential bias, because by doing this we 
assumed that yolksac larvae have the same mortality rate 
as eggs. To avoid this bias, the number of staged egg sam- 
ples should be increased. One approach would be to stage 
the eggs taken with a CUFES in the high-density stra- 
tum and combine them with the CalVET samples from the 
same stratum. This would greatly increase the number of 
staged eggs available but requires the assumption that 
CUFES staged eggs are an unbiased sample of the full- 
water-column, a condition not met in our 1996 pilot study. 
Another approach would be to increase the number of 
CalVET samples per mile in the high-density stratum 
from the present one sample per 4 nmi to a higher fre- 
quency. A doubling of the CalVET sampling rate would 
significantly increase survey costs and may not increase 
the number of positive samples sufficiently to be able to 
use only eggs for estimation of P 0 . Considering the relative 
risks and costs of these approaches, we feel that the use 
of the yolksac larvae in the estimation of P 0 was prefer- 
able. In other species, the distribution of eggs may be less 
concentrated than they are for sardine and sampling at 
the 4 nmi sampling rate may be adequate. Clearly, other 
solutions may exist, and we recommend considering these 
issues with each new application. 
Conclusions 
We conclude that the CUFES is a useful tool with the 
DEPM when it is used adaptively to establish sampling 
strata for CalVETs. In this mode, the CUFES increases 
precision and reduces cost per transect mile. Perhaps one 
of the major benefits of using a CUFES in the DEPM is 
that one can better afford to expand survey boundaries 
and thereby reduce the potential bias of not enclosing the 
entire population, a very common bias in field surveys in 
general (Gunderson, 1993). Clearly, the more contiguous 
the distribution of spawned eggs within the surveyed hab- 
itat, the greater the benefits in using the CUFES. 
Acknowledgments 
We thank three unanimous reviewers for their comments; 
David Griffith, Ron Dotson, and Amy Hays for operating 
the CUFES for 1996 and 1997; David Ambrose, William 
Watson, and Elaine Acuna for counting and staging eggs of 
CUFES and CalVET samples; Geoffrey Moser for provid- 
ing the information on misidentification of sardine eggs; 
and crew members of NOAA RV David Starr Jordan for 
their cooperation. 
Literature cited 
Alheit, J. 
1993. Use of the daily egg production method for estimating 
biomass of clupeoid fishes, a review and evaluation. Bull. 
Mar. Sci. 53(2)750-767. 
Ahlstrom, E. H. 
1943. Studies on the Pacific pilchard or sardine ( Sardinpos 
caerulea ). 4. Influence of temperature on the rate of devel- 
opment of pilchard eggs in nature. U.S. Fish. Wild], Serv., 
Spec. Sci. Rep. 23, 26 p. 
All 1 strom, E. H., and H. G. Moser. 
1980. Characters useful in identification of pelagic marine 
fish eggs. Calif. Coop. Oceanic Fish. Investig. Rep. 21:121- 
131. 
Armstrong, M., P. Shelton, I. Hampton, G. Jolly, and Y. Melo. 
1988. Egg production estimates of anchovy biomass in the 
southern Benguela. Calif. Coop. Oceanic. Fish. Investig. 
Rep. 29:137-157. 
Barange, M., and I. Hampton. 
1997. Spatial structure of co-occuring anchovy and sardine pop- 
ulations from acoustic data: implications for survey design. 
Fish. Oceangr. 6:294-108. 
Bentley. P. J., R. L. Emmett, N. C. H. Lo, and H. G. Moser. 
1996. Egg production of Pacific sardine ( Sardinops sagax ) 
off Oregon in 1994. Calif. Coop. Oceanic. Fish. Investig. 
Rep. 37:193-200. 
Chambers, J. M., and T. J. Hastie (ed). 
1992. Statistical models in S. Wadsworth and Brooks/Cole 
Advanced Books and Software, Pacific Grove, CA, 608 p. 
Checkley, D. M., Jr, R. C. Dotson, and D. A. Griffith. 
1999. Continuous, Underway sampling of eggs of northern 
anchovy (Engraulis mordax ) and Pacific sardin e (Sardinops 
sagax ) in spring 1996 and 1997 off southern and central 
California. Deep-Sea Res. (special rep.) II 47(2000) 1139- 
1155. 
Checkley, D. M., Jr., P. B. Ortner, L. R. Settle, L.R., and 
S. R. Cummings. 
1997. A continuous, underway fish egg sampler. Fish. 
Oceanogr. 6:58-73. 
Cressie, N. 
1991. Statistics for spatial data. John Wiley and Sons, Inc. 
New York, Chichester, Toronto, Brisbane, Singapore, 900 p. 
Cressie, N., and D. M. Howkins. 
1980. Robust estimation of the variogram. I. J. Int. Assoc. 
Math. Geology 12:115-125. 
