8 
Fishery Bulletin 1 1 1 (1) 
Table I 
(Upper): Comparison of taxonomic resolution between bongo and In Situ Ichthyoplankton Imaging System 
(ISIIS) samples collected south of Woods Hole, Massachusetts, in October 2008 as part of this study. Data are 
presented as “total,” which is the combined lowest level of identification across all taxa; “family,” which is a 
comparison just at the family level (where all taxa are subsumed into relevant family taxa), and “species,” 
where only identifications to species level are presented. (Lower): Summary comparison between the bongo 
sampler and ISIIS gears for number and proportion of identifications at family, genus, and species levels, as 
well as number and proportion of unknowns. 
Identification level 
Taxa 
Total (lowest) 
Family level 
Species level 
Bongo 
ISIIS 
Bongo 
ISIIS 
Bongo ISIIS 
Clupeidae 
1 
3 
2 
3 
Brevoortia tyrannus 
1 
0 
1 0 
Gadidae 
3 
13 
3 
13 
Merlucciidae 
0 
0 
48 
44 
Merluccius bilinearis 
48 
44 
48 44 
Phycidae 
0 
83 
48 
104 
Urophycis spp. 
48 
21 
48 21 
Ophidiidae 
0 
29 
7 
34 
Lepophidium profundorum 
7 
5 
7 5 
Gobiidae 
3 
6 
3 
6 
Paralichthyidae 
1 
62 
217 
135 
Citharichthys arctifrons 
14 
0 
14 0 
Etropus spp. 
10 
8 
10 8 
Paralichthys oblongus 
4 
0 
4 0 
Paralichthys dentatus 
188 
65 
188 65 
Scopthalmidae 
31 
10 
31 
10 
Unknown 
0 
60 
0 
60 
Total larvae 
359 
409 
Numbers 
Proportion 
Bongo 
ISIIS 
Bongo 
ISIIS 
Family 
39 
206 
0.11 
0.50 
Genus 
58 
29 
0.16 
0.07 
Species 
262 
114 
0.73 
0.28 
Unknown 
0 
60 
0.00 
0.15 
Total 
359 
409 
1 
1 
the technology. The version of ISIIS used in this study 
was an early prototype (Cowen and Guigand, 2008); 
considerable advancements have been made in the 
image sharpness and depth of field since the field 
work reported here, and these changes should improve 
identification of individual fishes, especially of smaller 
taxa. 
Larval lengths were different for ISIIS and the bon- 
go sampler. The bongo sampler collected smaller larvae, 
indicating limitations with our ISIIS image-processing 
procedures for recording larval fishes <5 mm (and obvi- 
ous diagnostic morphological features on small larvae). 
On the other hand, ISIIS imaged larger larvae, indicat- 
ing that avoidance of the ISIIS by larger larvae was 
reduced. With the potential of an increase in image 
resolution to advance identification of smaller larvae 
(e.g. the improved image of a chaetognath in Fig. 3, 
upper right), the overall size range sampled by ISIIS 
could be a significant improvement over the range of 
the bongo sampler that has been used by the NEFSC 
for the past 30-plus years. If there is an effort to merge 
abundance time series between the bongo and ISIIS, 
careful calibration studies would be required to account 
for variances, including length-based, diel, and regional 
differences in detectability. These types of calibration 
studies also are necessary to combine data across dif- 
ferent mesh sizes of the bongo sampler (see Johnson 
and Morse, 1994; Richardson et ah, 2010). 
Our results indicate that ISIIS could be a valuable 
addition to the survey sampling toolbox because it sue- 
