250 
Fishery Bulletin 119(4) 
Discussion 
Recruitment 
This study is the first to directly estimate annual recruit- 
ment of any population of Gulf sturgeon. Across the 6 years 
of this study, age-1 recruits were observed every year, 
and mean annual recruitment was relatively stable. The 
Apalachicola River produced approximately 50 age-1 Gulf 
sturgeon in most years, although recruitment was substan- 
tially greater in 2014, when we estimated that there were 
210 age-1 sturgeon. 
Although there are no direct recruitment estimates 
for Gulf sturgeon to which our results can be compared, 
Pine and Martell® used an age-structured, mark-recapture 
model to back-calculate recruitment on the basis of 
the adult population, and they reported annual recruit- 
ment in the Apalachicola River to be 100-300 individuals/ 
year, slightly greater than our estimates. However, direct 
recruitment estimates do exist for Atlantic sturgeon in sev- 
eral southern rivers. The largest populations of Atlantic 
sturgeon in the southeastern United States have annual 
recruitment that is an order of magnitude greater than 
what we generally observed in the Apalachicola River: 
the Altamaha River, in Georgia, produces 500—2500 age-1 
juveniles every year (Schueller and Peterson, 2010), and 
the Savannah River, in Georgia and South Carolina, pro- 
duces 500-600 juveniles annually (Bahr and Peterson, 
2016). Recruitment of Gulf sturgeon in the Apalachicola 
River appears more similar to that observed in small riv- 
ers in Georgia’s coastal plain, rivers where annual recruit- 
ment is <100 age-1 juveniles per year, if any, such as in the 
Ogeechee River (Farrae et al., 2009), Satilla River (Fritts 
et al., 2015), and St. Marys River (Fox et al., 2018). These 
populations are considered particularly small and imper- 
iled (ASSRT, 2007). 
Although recruitment of Gulf sturgeon in the Apalachicola 
River was substantially lower than recruitment of Atlantic 
sturgeon in similar large, Piedmont river systems (e.g., the 
Savannah and Altamaha Rivers), all 3 river systems did 
produce new recruits every year, indicating that spawn- 
ing occurred every year. In contrast, small populations of 
Atlantic sturgeon (e.g., those in the Satilla and St. Marys 
Rivers) did not produce age-1 recruits every year. This 
comparison to recruitment of Atlantic sturgeon indicates 
that, although Gulf sturgeon in the Apalachicola River 
spawn successfully every year, recruitment may be lim- 
ited. Analysis of recruitment across other populations of 
Gulf sturgeon will be necessary to determine how the pop- 
ulation in the Apalachicola River compares to others of the 
same subspecies. 
We tagged fish largely within the Brothers River; 
therefore, our recruitment estimates pertain primarily 
to the population of juveniles that reside in this part of 
the river system. Over many years ‘of historical sam- 
pling by the U.S. Fish and Wildlife Service (A. Kaeser, 
unpubl. data), juvenile Gulf sturgeon have been captured 
in only one other location outside the Brothers River— 
below the JWLD. Moreover, in this study, we set nets in 
the Apalachicola River, including in many reaches with 
characteristics (e.g., depth, salinity, and temperature) 
similar to those of the Brothers River and, in most years, 
caught no age-1 juveniles outside of the Brothers River. 
Our capture of several age-1 juveniles below the JWLD 
indicates that juveniles can (sometimes) be found at this 
location. Because our estimates are specific to the geog- 
raphy we sampled, we acknowledge that the presence of 
age-1 fish in undiscovered aggregation sites, should they 
exist, would mean that we have underestimated true, 
population-level recruitment in the Apalachicola River 
system. However, unless there are unknown aggregation 
sites that contain dozens or hundreds of unsampled age-1 
fish, it seems likely that our estimates either accurately 
represent or provide a robust indicator of recruitment 
to age 1 in the river system. Annual recruitment in the 
Apalachicola River appears to be measured in dozens of 
fish, not hundreds or thousands. 
The validity of the abundance estimates produced in 
this study relies heavily on the assumption of population 
closure (demographic and geographic)—no births, deaths, 
immigration, or emigration can occur during the sampling 
period (Huggins, 1989). Under this assumption, capture 
rates could fluctuate throughout the sampling season 
(as they did during this study), resulting in more accu- 
rate estimates than values produced with an open model 
(Stanley and Richards, 2005). Although the assumption of 
closure can never be proven per se, there are several rea- 
sons why we believe the population was essentially closed 
during our sampling. For numerous studies (e.g., Wooley 
and Crateau, 1985; Hightower et al., 2002) in which the 
life history of Gulf sturgeon was examined, results indi- 
cate that all juveniles returned to aggregation sites in 
their natal river during summer months. We conducted 
our sampling mainly in the summer at the aggregation 
site in the Brothers River. CloseTest results (Table 2) indi- 
cate that age-1 cohorts likely met the closure assumption 
in most years of the study—the lack of closure in 2016 
could have been a result of elevated mortality due to 
exceptionally warm water temperatures (see discussion of 
survival in the next section). 
The telemetry data collected during this study largely 
support the assumption of closure for the Brothers 
River. No tagged age-1 juveniles were detected outside 
the aggregation site in that river for more than a short 
period (<4 h), and no individuals marked (i.e., acoustically 
tagged) near the JWLD were observed in the Brothers 
River during the summer in which they had been tagged. 
The JWLD appears to have a separate aggregation site 
that is occupied by only a few, if any, age-1 fish. However, 
if the population was not closed, our estimates of annual 
recruitment would be lower than the true number of 
recruits. 
Mark-recapture models also include the assumption 
that marks (i.e., tags) were not lost or overlooked. In lake 
sturgeon (Acipenser fulvescens), PIT tag retention is 99% 
(Briggs et al., 2019), and we would expect similar results 
for other species in Acipenseridae. Each Gulf sturgeon we 
captured was thoroughly scanned for PIT tags, and many 
