Kupchik and Shaw: Effects of recruitment through a coastal boundary layer on growth of larval Brevoortia patronus 
215 
■D 
D) 
•D 
■5 
D) 
30 
25 
20 
15 
10 
5 
0 
1.0 
0.8 
0.6 
0.4 
0.2 
0.0 
n 
10 
20 
30 
~r 
40 
50 
60 
70 
B 
• • % • • 
I * •• • •• 
10 
20 
30 40 
~r 
50 
-r~ 
60 
n 
70 
Age (das) 
Figure S 
Data from analysis of otolith microstructures of larval Gulf menhaden 
{Brevoortia patronus) collected at Bayou Tartellan, Louisiana, from Octo¬ 
ber 2006 to April 2007 (year 1) and from September 2007 to April 2008 
(year 2). (A) Otolith ring mean distance from the core by sample year. The 
vertical black dashed line represents the modeled developmental shift, at 
33 days after spawning (das), from the larval stage into the beginning of 
the juvenile stage. The thin dashed lines represent 95% confidence inter¬ 
vals for the measured distances. (B) Mean ring width for individual daily 
rings observed in otoliths. Decreases in ring width correspond with the 
modeled developmental shift that occurred at 33 das. 
these ages is likely to be due to the variability of food 
sources or other resources that result in differences in 
the instantaneous larval growth rate (Warlen, 1988; 
Lyczkowski-Shultz et ah, 1990; Warlen, 1992) during 
the initial larval stage, when Gulf menhaden larvae 
are selective particulate feeders (Stoecker and Govoni, 
1984; Deegan, 1990; Lozano et ah, 2012). 
Differences in water temperature between sampling 
year 1 and 2 are correlated with the slight differences 
in growth rate and otolith microstructure between both 
years. In year 1 the highest growth rate was 0.63 mm/ 
day and was much lower than the maximum growth 
rate in year 2 (0.77 mm/day), when overall warmer wa¬ 
ter temperatures likely aided increased somatic growth 
(Houde, 1974; Heimbuch et al., 2007). However, differ¬ 
ences in mean ring distance from the otolith core, al¬ 
though small, showed the opposite trend, with slight¬ 
ly faster otolith ring growth during year 1. Previous 
research has revealed a similar decoupling between 
otolith growth and somatic growth (Mosegaard et al., 
1988; Fey, 2006). Despite this partial decoupling, oto¬ 
lith microstructure for both years showed an increase 
in otolith ring width directly before the beginning of 
the ontogenetic shift from selective 
particulate feeding to omnivorous fil¬ 
ter feeding—an increase that suggests 
a direct relationship between the on¬ 
togenetic shift and otolith deposition. 
The high growth rate before the shift 
in feeding strategy may be the result 
of a cumulative experience of adjust¬ 
ing to food source or environmental 
resources, before learning new feed¬ 
ing skills to accommodate the ongoing 
development of new feeding structures 
(Stoecker and Govoni, 1984; Deegan, 
1990; Maillet and Checkley, 1990; Lo¬ 
zano et al., 2012). 
There was overwhelming agree¬ 
ment with the pooled 2-cycle Laird- 
Gompertz model, with the fit of the 
2-cycle Laird-Gompertz models for 
each individual year, and with the 
otolith microstructure analysis, for 
the mean age for the shift in growth. 
This growth stanza is a reflection of 1) 
transitioning from offshore spawning 
grounds through the coastal boundary 
layer and to an estuarine water mass 
with differing primary productivities 
and 2) the onset at which the onto¬ 
genetic shift from the larval stage to 
the juvenile stage begins. The modeled 
shift in growth (33 das) coincides with 
the previously reported shift in growth 
rate at 33.6 das for Gulf menhaden in 
Fourleague Bay, Louisiana (Raynie 
and Shaw, 1994). Mean distance from 
the otolith core and mean ring width 
decreased after 33 das, which prob¬ 
ably reflects limited feeding at this point owing to the 
ontogenetic shift (Deegan, 1990; Raynie, 1991; Warlen, 
1992; Lozano and Houde, 2013). Transforming larvae at 
this age, which were between approximately 19 and 21 
mm SL, may be expending more energy in the develop¬ 
ment of a feeding apparatus as they change food source 
for the juvenile and adult stage (Deegan, 1990; Mail- 
let and Checkley, 1990), and in deepening their body 
rather than increasing length (Deegan, 1990; Raynie 
and Shaw, 1994). 
The slower growth rates during the period when 
feeding structures and feeding strategy begin to change 
were similar across sampling years and modeling tech¬ 
niques. The growth rates for the pooled model (0.11 
mm/day) and yearly two-cycle models (year 1, 0.12 mm/ 
day; year 2, 0.10 mm/day) showed strong agreement 
with the rates for the age-grouped Laird-Gompertz 
models (SL grouping: 0.10 mm/day; age grouping: 0.08 
mm/day). Although useful in comparisons with previous 
work, the age-grouped Laird-Gompertz models are less 
parsimonious and did not capture the ontogenetic shift 
in feeding strategy from a selective particulate filter 
feeder to an omnivorous filter feeder, except through 
