Lewis et al.: Integrating DNA barcoding of fish eggs into ichthyoplankton programs 
157 
were queried against the web-based BOLD database. 
We used a threshold of a 99% match to assign a match 
to species. For specimens for which the BOLD engine 
failed to find a match, sequences were applied directly 
to the Standard Nucleotide BLAST on GenBank (Na- 
tional Center for Biotechnology Information, website). 
Sequence data, electropherograms, and primer de- 
tails for specimens are available within the completed 
NIFEB project file on BOLD. Sequence data was also 
submitted to GenBank (accession numbers: KP1 10771- 
KP112146). Low-quality sequences can often provide 
sufficient information to identify a sample to species; 
however, we chose to exclude them from subsequent 
analyses to ensure the reliability of our database. 
Data analysis 
Standard box plots were used to present the SST range 
and size range of each species of egg that we collected 
and barcoded. In addition, for species that were identi- 
fied from at least 15 stations, we further sought to de- 
termine the correlation between egg measurement and 
the SST at sampling stations. The graphical and statis- 
tical tools available through RStudio (RStudio, Boston, 
MA) were used to perform linear regression analysis on 
the effects of SST on egg diameter. Notably, the process 
of measuring and photographing eggs did not occur for 
the first 2 plates (190 eggs) of samples; therefore, these 
eggs were excluded from the analyses where egg diam- 
eter was used. 
For SST in our analyses, it was assumed that the 
eggs collected at a station were located above the ther- 
mocline or that the waters were well mixed. During 
the late fall (November-December) and winter (Janu- 
ary-February), the water column in our region is well 
mixed and, therefore, the vertical distribution of eggs 
has little effect on the temperature they experience. 
The same is true year-round in the shallow areas on 
Georges Bank. However, during the late spring (May- 
June) and summer (August), the water column in our 
region is stratified. Typically, fish eggs collected in 
ichthyoplankton sampling are positively buoyant and 
concentrate in the upper water column, but there are 
exceptions (Conway et al., 1997). Because our sampling 
was not vertically stratified, we relied on SST as the 
best estimate of the temperature experienced by the 
eggs. 
Results 
In total, DNA was sequenced from 1603 unidentified 
fish eggs collected at 456 stations and that were pro- 
cessed as ethanol-preserved ichthyoplankton samples. 
Of these eggs, 93.26% (1495 eggs) were sequenced suc- 
cessfully, providing medium- or high-quality barcodes 
suitable for species-level identifications. Of the 108 un- 
identified eggs, 60 eggs (3.74%) failed both first- and 
second-round sequencing attempts, and 8 eggs (0.50%) 
were flagged because of contamination that occurred 
at an unknown point during the barcoding process. 
The remaining 40 eggs (2.50%) produced low-quality 
sequences. In many cases, identifications could be as- 
signed to these low-quality sequences; however, we 
chose to be conservative and classify these eggs as un- 
identified. Sequence analysis revealed that the 1495 
successfully sequenced fish eggs represented 50 identi- 
fied species, 49 of which could be definitively matched 
to species-level barcodes, and 1 taxa with 12 eggs pro- 
duced a match to a specimen identified previously only 
at the family level (Engraulidae) on GenBank (Table 
1 ). 
The number of identified eggs per species ranged 
from 1 to 196. The 4 most frequently identified eggs 
were those of silver hake ( Merluccius bilinearis), 
fourspot flounder ( Hippoglossina oblonga ), Gulf Stream 
flounder (Citharichthys arctifrons), and red hake ( Uro - 
phycis chuss ); together, these 4 species accounted for 
more than 45% of all successful barcode identifications. 
In comparison with previous morphological attempts 
at fish egg identification in the northeastern United 
States (Colton and Marak 1 ; Berrien and Sibunka 2 ), 
the eggs of our barcoded species fit into 3 general cat- 
egories: category I (6 of our 50 species) are eggs that 
have never before been identified to the species-level as 
eggs; category II (33 species) describes eggs that have 
historically been identifiable only at higher taxonomic 
levels (genus or family) or at the species level during 
specific stages of development (i.e., eggs at mid to late 
stages); and category III (11 species) contains eggs that 
are well described and can be identified confidently to 
species level through the use of morphological criteria 
at all stages of development (Table 1). 
Overall, egg abundances varied significantly with 
sampling season (Fig. 2, A-D). In aggregate, the great- 
est abundances and diversity of eggs were found during 
the sampling periods of late summer (August) and late 
spring (May-June). The lowest abundance of eggs was 
encountered in late autumn (November-December), 
and the lowest diversity was found in the winter ( Janu- 
ary-February). Of the 50 species with eggs identified, 
25 species had eggs collected during multiple seasons 
(indicative of cross-seasonal or elongated spawning). 
The most frequently observed cross-seasonal collec- 
tion pattern was between late spring and late summer; 
however, the eggs of Atlantic cod (Gadus morhua) and 
pollock ( Pollachius virens ) were collected throughout 
late autumn and winter. The eggs of only 1 species, 
offshore hake (Merluccius albidus), were collected in 
all 4 sampling seasons. The 3 most abundant species 
of eggs identified, with all seasons combined, were red 
hake, Gulf Stream flounder, and silver hake (Table 1). 
The successfully barcoded fish eggs were collected 
at a wide range of temperatures; recorded SSTs ranged 
from 3.11°C to 27.02°C (Table 1). In general, the SST 
range of an individual species was much narrower 
(Fig. 3). Eggs of haddock (Melanogrammus aeglefinus), 
American plaice ( Hippoglossoides platessoides), and 
spotted pikeconger (Hoplunnis tenuis ) were collected at 
the lowest average SSTs (4.63°C, 7.15°C, and 7.29°C, 
