Johansson et al.: Seasonal distribution and environmental associations of iarval and juvenile Sebastes spp. 
267 
micronekton surveys in the eastern North Pacific. Lar¬ 
val and juvenile rockfishes may represent a great deal 
of unrecorded diversity because they are often difficult 
or impossible to identify to species based on body shape 
or pigmentation patterns. This highly diverse genus is 
represented by at least 65 species in the Pacific Ocean 
off the coast of North America, of which 36 to 40 oc¬ 
cur in waters off southern Washington and northern 
Oregon (Love et al., 2002). 
After parturition, the life history of all rockfish spe¬ 
cies includes a pelagic phase, consisting of both lar¬ 
val and juvenile stages (Boehlert, 1977; Moser and 
Boehlert, 1991). Settlement to the bottom occurs after 
a period of a few months to a year, and varies by spe¬ 
cies (Love et al., 2002). Larval and pelagic-juvenile- 
stage rockfish represent a significant proportion of 
the ichthyoplankton and micronekton community off 
Washington, Oregon, and California (Auth, 2009; Phil¬ 
lips et al., 2009; Thompson et al., 2016; Thompson et 
al., 2017). 
The early-stage larvae of a number of species of 
rockfish occurring off the west coast of North Ameri¬ 
ca have been described morphologically (Sakuma and 
Laidig, 1995; Moser, 1996). However, misidentification 
of larval and juvenile fish often occurs even at the 
family level (Ko et al., 2013). Visual identification of 
field-caught late-larval and early-juvenile rockfishes 
is especially difficult because different but closely re¬ 
lated species of Sebastes occur together and there are 
few distinguishing morphological characters (Seeb and 
Kendall, 1991; Sakuma and Laidig, 1995; Rocha-Oliva- 
res et al., 2000). This difficulty of identifying late lar¬ 
val and juvenile rockfishes is particularly problematic 
given their importance in the fisheries they support 
in the northeast Pacific Ocean (Love et al., 2002) and 
their history of being heavily fished. 
Surveys of pelagic larval and juvenile fish are a 
widely used fishery-independent means of tracking 
abundances of marine fishes (Lasker, 1985; He et al., 
2015). These surveys are both easier and cheaper than 
sampling adult fishes, and are likely to have a smaller 
impact on populations. In order to provide early indica¬ 
tions of the potential recruitment of fishes to the north¬ 
ern California Current, the NOAA Northwest Fisher¬ 
ies Science Center (NWFSC) Fish Ecology Division has 
conducted surveys as part of the Stock Assessment Im¬ 
provement Program (SAIP) during summer from 2004 
through 2012 to collect late-larval and juvenile fish 
from southern Washington to central Oregon (Phillips 
et al., 2009). 
The ability to efficiently identify large numbers of 
individual late-larval and pelagic juvenile rockfish 
would make it possible to characterize the distribu¬ 
tions and abundances of the early life stages of species 
(Love et al., 2002; Ralston and Stewart, 2013), as well 
as conduct associated recruitment studies and estimate 
juvenile biomass (Sakuma and Laidig, 1995; Laidig et 
al., 2007). Molecular methods based on allozyme, mic¬ 
rosatellite, or DNA sequence variation offer an alterna¬ 
tive to morphological approaches to identify larvae to 
species quickly and efficiently (Seeb and Kendall, 1991; 
Rocha-Olivares et al., 2000; Gharrett et al., 2001; Tay¬ 
lor et al., 2004; Gray et al., 2006). 
Our objectives were to identify to species late-stage 
larvae and juveniles of the genus Sebastes collected in 
SAIP tows of midwater trawls made during 2005-2008 
by using visual meristic or pigmentation patterns. We 
used the cytochrome b sequence-based method of Tay¬ 
lor et al. (2004) to identify to species the remaining 
visually unidentifiable fraction. Using both of these 
techniques, we also sought to assess spatial and tempo¬ 
ral variations in composition, concentration, and other 
characteristics of the late-larval and juvenile rockfish 
community off the coast of Oregon and Washington, and 
to relate their distributions to environmental condi¬ 
tions during the main upwelling season (May-October). 
Materials and methods 
Sampling procedure 
Midwater trawls were used to sample micronekton, in¬ 
cluding juvenile fishes, from spring to fall during 2005- 
2008. With the exception of May 2005 and July 2006, 
monthly cruises were conducted from May through 
September of each year, and an additional cruise was 
included from October 2005. In each sampling cruise, 
the same 5 stations were sampled along each of 4 
transects: Heceta Head (HH) (44.00°N), Newport (NH) 
(44.65°N), and the Columbia River (CR) (44.16°N) off 
Oregon and Willapa Bay (WB) (46.67°N) off Washing¬ 
ton (Fig. 1). Stations ranged from approximately 20 
to 100 km offshore along each transect. A single sta¬ 
tion, located 185 km (100 nautical miles) off the NH 
transect (the NH-100 station) was sampled just once in 
June 2008. Owing to weather or mechanical problems, 
some stations or transects (or both) were not sampled 
during some cruises. 
To collect the samples used in this project, a Nordic 
264 rope trawl was towed for 15 min with the head 
rope at a target depth of 30 m. The effective mouth 
area of the trawl was 12 m high and 28 m wide (336 
m 2 ; Emmett et al., 2004). The main body of the net 
had variable mesh sizes (from 162.6 cm at the mouth 
to 8.9 cm at the codend), and a 6.1-m long, 3-mm 
stretched knotless web liner in the codend. All tows of 
the trawl were conducted at night. After all fish and 
invertebrates >10 cm in total length were removed, the 
remaining catch was subsampled as follows: samples 
with an unsorted volume of <0.25 m 3 were frozen in 
their entirety, whereas samples with an unsorted vol¬ 
ume of >0.25 m 3 were subsampled in the amount of 
0.25 m 3 or 20% of the entire sample, whichever was 
larger. The retained unsorted catches were frozen at 
sea and later thawed and sorted in the laboratory. For 
additional details on the collection method, see Phillips 
et al. (2007). 
In the laboratory, individual rockfish were identified 
to the lowest possible taxonomic level by using a va- 
