154 
Fishery Bulletin 1 14(2) 
Several of these features remain relatively constant 
throughout embryonic development; however, many 
characteristics are stage specific. 
Overall, there are 4 significant limitations involved 
with morphological identifications. First, natural vari- 
ation in morphological features of fish eggs, within a 
species, is common because of the effects of both mater- 
nal (Chambers and Leggett, 1996; Marteinsdottir and 
Steinarsson, 1998; Marteinsdottir and Begg, 2002) and 
environmental (Chambers and Leggett, 1996; Kucera 
et al., 2002) factors. Second, overlaps in morphological 
stages and similarities among the eggs of related and 
unrelated cryptic taxa can lead to incorrect identifica- 
tions (Hyde et al., 2005; Berrien and Sibunka 2 ; Gleason 
and Burton, 2012). Third, egg fixation in preservatives 
causes the loss of natural pigmentation and can obscure 
the developmental stage of an embryo (Valdez-Moreno 
et al., 2010). Finally, fixation can cause substantial egg 
shrinkage, a phenomenon that varies with preservative 
type and length of exposure (Hiemstra, 1962). In the 
Northwest Atlantic specifically, descriptions of eggs are 
lacking for more than 50% of fish species, and for those 
species for which there are descriptions, characteristics 
used to confidently identify field-collected specimens 
may be lacking for some or all stages of egg develop- 
ment (Kendall and Matarese, 1994). 
Over the past decade, the use of molecular identi- 
fication techniques has increased in response to the 
limitations associated with morphological identifica- 
tions. For ichthyoplankton, these methods most fre- 
quently have involved the use of polymerase chain re- 
action (PCR) with species-specific primers and probes 
(Shao et al., 2002; Fox et al., 2005; Hyde et al., 2005; 
Carreon-Martinez et al., 2010) or the use of multiplex 
suspension bead arrays (Gleason and Burton, 2012); 
the former technique is even implemented onboard a 
research vessel (Hyde et al., 2005) and is used with 
formalin-preserved samples (Goodsir et al., 2008). Al- 
though these methods have been used successfully, 
the primary focus in these studies was to discriminate 
among a limited number of species. As a result, these 
methods rarely have proved versatile enough for use 
in fisheries monitoring programs or in large-scale egg 
surveys for which the identification of a wide diversity 
of species is necessary. 
Around 2005, a new means of molecular identifica- 
tion was proposed with the goal of providing a uni- 
versal approach to species identification. This concept, 
termed DNA barcoding , is based on the premise that 
the sequence diversity of a single mitochondrial gene, 
cytochrome c oxidase subunit 1 (COI), is suitable for 
identifying most animal species (Hebert et al., 2003a, 
2003b). Concerns over the use of a single locus to dis- 
criminate between closely allied species have been 
stressed since the DNA barcoding movement began; 
however, multiple studies have concluded that COI 
analysis is sufficient for species-level diagnoses. For ex- 
ample, this approach has been used successfully to bar- 
code many Australian fish species (Ward et al., 2005) 
and to reveal overlooked marine species in the Indian 
Ocean (Zemlak et al., 2009). The method of DNA bar- 
coding has also been used to connect the egg and larval 
stages of marine species along the coastline of Yucatan, 
Mexico, to their adult counterparts (Valdez-Moreno et 
al., 2010) and for a large-scale larval fish study in the 
Straits of Florida (Richardson et al., 2007; Richardson 
et al., 2010b). To date, the DNA barcoding database for 
fishes (Barcode of Life Data System [BOLD], website, 
accessed March 2015) contains sequences of more than 
175,000 specimens, representing more than 15,000 spe- 
cies (Ratnasingham and Hebert, 2007). These advances, 
coupled with a simultaneous decrease in cost (Richard- 
son et al., 2007), have made it possible to consider the 
use of DNA barcoding for identification of fish eggs and 
to incorporate this approach into ecosystem monitoring 
programs. 
In our study, we used DNA barcoding for the large- 
scale identification of fish eggs. We sequenced DNA 
from eggs that were selected from a 10-year, multisea- 
sonal archive of 456 ethanol-preserved samples that 
exists within the Northeast Fisheries Science Center 
(NEFSC), National Marine Fisheries Service, NOAA. 
Before 2000, fish eggs were identified by using morpho- 
logical criteria. Since 2000, fish eggs collected through 
the NEFSC Ecosystem Monitoring (EcoMon) program 
have been counted but not identified because of a lack 
of personnel. The goals of this study are 1) to compare 
the species identifications, those produced from mor- 
phological versus molecular analyses of eggs, in terms 
of species diversity and taxonomic resolution and 2) to 
evaluate the feasibility of incorporation of DNA bar- 
coding into long-term, regional-scale ichthyoplankton 
monitoring programs. 
Materials and methods 
Sample collection and processing 
The NEFSC since 1971 has conducted ichthyoplank- 
ton surveys multiple times annually along the north- 
eastern U.S. continental shelf between Cape Hatteras, 
North Carolina, and Nova Scotia, Canada (Richardson 
et al., 2010a). During these surveys, ichthyoplankton 
was sampled throughout the water column (to a depth 
within 5 m of the seafloor or to a maximum depth of 
200 m) with paired bongo samplers that had diameters 
of 61 cm and that were equipped with 333-pm mesh 
nets. These samples had been preserved in formalin 
and — because formalin fixation causes significant dam- 
age to DNA, inhibiting amplification of the target COI 
gene — therefore were not considered suitable for this 
project (Hajibabaei et al., 2005). However, an addition- 
al bongo net, with a 20-cm diameter, was included on 
the towing cable, just above the 61-cm bongo net at 
about 15-20% of the sampling stations. Samples col- 
lected in the smaller net had been fixed and preserved 
in 95% ethanol (EtOH). 
For our study, we used an archive of 456 EtOH- 
preserved ichthyoplankton samples that were collected 
