Schobernd et al.: A comparison of numbers of fish larvae extruded from plankton nets of different mesh size 
251 
opening of the net and longer tow durations used by 
Comyns (1997) resulted in higher volumes of water fil¬ 
tered and subsequent greater numbers of larvae cap¬ 
tured than in the Hernandez et al. (2011) study. Simi¬ 
larly, mean volume filtered in our study was less than 
that in the Comyns (1997) study, 170 m 3 and 256 m 3 , 
respectively. Additionally, of the different gear used in 
the 3 studies, only the SEAMAP bongo frame allowed 
side-by-side towing and paired sample collection. These 
‘true’ paired tows ensured that the nets sampled at the 
same location in the water column, at the same tow 
speed and ambient light level, and would, presumably, 
encounter the same assemblage of larvae. The arrange¬ 
ment of the nets used to gather plankton samples by 
Comyns (1997) and Hernandez et al. (2011) did not al¬ 
low simultaneous sampling with different mesh sizes. 
Although the opportunistic nature of this study al¬ 
lowed the sampling of various regions and species in 
the Gulf of Mexico, directed sampling in months and 
locations of peak spawning for species of interest would 
have improved the probabilities of capturing greater 
numbers of smaller larvae, and increased the number 
of taxa within the target families that would have been 
‘available’ for evaluating mesh retention. The impor¬ 
tance of sampling with nets of different size during 
times of peak spawning when earliest stage larvae are 
most abundant was exemplified by the observed size 
distributions among sciaenid and scombrid larvae. 
Sampling with different mesh sizes coincided with re¬ 
ported months of peak spawning for 4 species of Sci- 
aenidae and 5 species of Scombridae. A clear relation¬ 
ship between abundance ratios of the smallest sampled 
larvae and the mesh size of sampling nets was evident 
for those 2 families. This was the case even though 
scombrid larvae were the second least abundant of the 
target families. Although retention of small clupeid 
and lutjanid larvae was observed to differ between the 
paired samples collected with the nets with both mesh 
sizes and was supported by the significant difference 
in the length-frequency distributions of the 2 taxa, 
the modeled results for those taxa failed to indicate 
a substantial difference in abundance-at-size between 
the samples collected with the finer- and coarser-mesh 
nets. The lack of apparent difference between the nets 
with the 2 mesh sizes in size-related retention for those 
2 families most likely resulted from the smallest larvae 
being unavailable to our plankton samplers because of 
a mismatch between sampling with these nets and the 
times, locations, and seasons of clupeid and lutjanid 
spawning (Fitzhugh et al., 6 ; Ditty et al., 2005; Hanisko 
et al., 2007). 
Species-specific correction factors for larval abun¬ 
dance by body length in coarser mesh nets, generated 
from comparison studies of net meshes have been used 
6 Fitzhugh, G. R., M. S. Duncan, L. A. Collins, W. T. Walling, 
and D. W. Oliver. 2004. Characterization of red snapper 
(Lutjanus campechanus) reproduction: for the 2004 Gulf of 
Mexico SEDAR. Southeast Data, Assessment and Review 
SEDAR7-DW-35, 27 p. [Available from website.] 
to adjust larval abundances in order to mitigate bias 
caused by extrusion of the smallest size category of 
larvae in sampled assemblages (Lo, 1983; Houde and 
Lovdal, 1984; Somerton and Kobayashi, 1989). How¬ 
ever, our study did not result in species-level compari¬ 
sons of retention between the nets with different mesh 
sizes because of the problematic nature of morpholo¬ 
gy-based identification of fish larvae in waters of the 
U.S. Southeast (Richards, 2006; Fahay, 2007). Large 
abundance ratios for unidentified and percoid larvae 
in our samples highlight the effect of extrusion at the 
smallest sizes. Those 2 categories can represent a large 
portion of ichthyoplankton survey catches. In our study 
alone, the unidentified and percoid categories account¬ 
ed for 9.1% of all specimens. Improved identification 
of smaller larvae with genetic procedures (Marancik 
et al., 2010) would provide more accurate estimates of 
total abundances of both eggs and larvae and aid in 
identifying true larval retention patterns by species. 
Although not at species-level, these models represent 
the first empirically derived approach to evaluating 
the degree of extrusion in SEAMAP ichthyoplankton 
samples. 
Currently, SEAMAP larval indices are used as in¬ 
dicators of spawning stock biomass, not as direct esti¬ 
mates of biomass. Furthermore, these indices are cal¬ 
culated from the abundance of larger larvae that can 
be reliably identified to species by using established 
morphological features and that are of a size indicat¬ 
ing full recruitment to the sampling gear. Including 
the size fraction of larvae that are underrepresented 
in SEAMAP samples collected in bongo nets with a 
0.333-mm mesh could eventually lead to more realistic 
estimates of larval mortality and therefore more pre¬ 
cise larval indices than those currently in use. Such an 
improvement in the reliability of a SEAMAP larval in¬ 
dex was recently demonstrated when data on the abun¬ 
dance of small, genetically identified, early stage red 
snapper larvae were, for the first time, incorporated 
into the SEAMAP index (Pollack 7 ). 
Corrections for larval extrusion will also aid in bet¬ 
ter estimation of larval fish injuries and mortalities 
in future Gulf of Mexico damage assessments. Previ¬ 
ously larval fish mortalities and subsequent production 
have been estimated in preparation of offshore lique¬ 
fied natural gas developments (Gallaway et al., 2007) 
and as a result of the 2010 Deep Water Horizon oil 
spill (Muhling et al., 2012). Such estimates have been 
based on SEAMAP data for constructing baseline lar¬ 
val conditions for injury calculations (French McCay et 
al. 8 ), however net efficiency issues with larval reten¬ 
tion in the standard SEAMAP nets were simply noted. 
7 Pollack, A. G. 2015. Personal commun. Riverside Tech¬ 
nology, Inc. Southeast Fish. Sci. Cent., Natl. Mar. Fish. Serv., 
NOAA, 3209 Frederic St., Pascagoula, MS 39568. 
8 French McCay, D., M. C. McManus, R. Balouskus, J. J. Rowe, 
M. Schroeder, A. Morandi, E. Bohaboy, and E. Graham. 
2015. Technical Reports for Deepwater Horizon Water Col¬ 
umn Injury Assessment—WC_TR.10: Evaluation of baseline 
densities for calculating direct injuries of aquatic biota dur- 
