154 
Fishery Bulletin 117(3) 
approximately 4.5 m. This winter trawl survey provided 
an opportunity to sample maturing rainbow smelt prior 
to their spawning migration into coastal rivers and, there¬ 
fore, to avoid bias associated with analyzing fecundity 
of fish intercepted during spawning runs when partially 
spawned fish are present and oocytes can be extruded 
during handling. 
In the laboratory, frozen rainbow smelt were thawed, 
excess water was removed with absorbent paper, and 
total weight (TW, to the nearest 0.01 g), total length 
(TL, to the nearest 1 mm), and sex were recorded. Sub¬ 
samples of female rainbow smelt for fecundity analysis 
were selected with a target sample size of 10 fish for 
each length bin of 10 mm. Scales were removed from the 
female subsamples between the lateral line and dorsal 
fin for age determination (Elzey et al. 6 ). Female rain¬ 
bow smelt have an asymmetrical ovarian pair with a left 
ovary larger than the right one. Ovaries were excised 
from females, excess tissue and water were removed, and 
ovarian weight (OW, to the nearest 0.01 g) and matura¬ 
tion stage (Grim and Glebe, 1990) were recorded. Any 
remaining ovarian tissue and membrane that could not 
be removed was assumed to represent a negligible weight 
relative to the total ovary weight. Body weight (BW) was 
recorded as TW-OW. 
Two subsamples of oocytes were removed for fecundity 
analysis from each paired ovary: one subsample anterior 
and one subsample posterior of the midpoint of each ovary. 
The sampling protocol targeted 500 oocytes/subsample to 
estimate egg density. Subsamples were placed on absor¬ 
bent paper, weighed (to the nearest 0.01 g), and fixed in 
3% formalin in small jars; once in the jars, samples were 
shaken immediately to prevent oocytes from clumping 
together. Additionally, to confirm previous findings of syn¬ 
chrony in oocyte development (Chen, 1970; Clayton, 1976), 
4 oocyte subsamples were collected from 4 rainbow smelt 
before samples were fixed to measure oocyte diameters. 
A finding that all oocytes developed at the same rate would 
support the use of determinate methods of fecundity esti¬ 
mation. Again, one subsample was collected anterior and 
one subsample posterior of the midpoint of each ovary. The 
diameter and maturity stage of 20 oocytes were recorded 
from each of these subsamples with an ocular micrometer 
(to the nearest 0.01 mm). The Kruskal-Wallis rank sum 
test was used to determine if there were differences in 
oocyte diameter among the locations in the ovary where 
oocyte samples were removed. 
Fixed subsamples of oocytes were processed for fecundity 
analysis by counting all oocytes with a compound micro¬ 
scope at lOx magnification. We calculated total fecundity, 
the number of mature oocytes prior to spawning, for an 
individual rainbow smelt as the product of the mean num¬ 
ber of oocytes per gram for the 4 oocyte subsamples and 
the TW of the paired ovaries (Nielsen and Johnson, 1983). 
6 Elzey, S. P., K. J. Trull, and K. A. Rogers. 2015. Massachusetts 
Division of Marine Fisheries Age and Growth Laboratory: fish 
aging protocols. Mass. Div. Mar. Fish. Tech. Rep. TR-58, 43 p. 
[Available from website.] 
Oocyte density (total estimated number of oocytes divided 
by OW) and relative fecundity (total fecundity divided by 
BW in grams) also were calculated as reproductive indices 
to compare among age classes. 
Total fecundity estimates for each individual were 
regressed against TL, TW, OW, and age to determine 
the relationships of fecundity to size and age of rainbow 
smelt. Regressions and the Shapiro-Wilk normality test 
were run on nominal data by using the MASS package, 
vers. 7.3-45 (Venables and Ripley, 2002), in the statistical 
software R, vers. 3.3.0 (R Core Team, 2016). The response 
variable and some predictor variables were not normally 
distributed, and all nominal regressions had patterns of 
variance inconsistency. As a result, both dependent vari¬ 
ables and independent variables were log transformed to 
improve normality and the constancy of variance (Sokal 
and Rohlf, 1995). All analyses were performed in R, vers. 
3.3.0. The level of statistical significance for all analyses 
was 0.05. 
Maturity Maturity and sex of marine samples of rainbow 
smelt were determined by macroscopic gonad examination. 
Immature rainbow smelt were readily identified because 
no development of gonadal tissue was visible; in contrast, 
the well-developed testis and ovaries of mature samples 
occupied a majority of the abdominal cavity space. Matu¬ 
rity ogives are important biological references used in 
stock assessments (ICES') and are critical for description 
of the reproductive potential of a short-lived, anadromous 
fish, such as the rainbow smelt. Logistic models are com¬ 
monly used to determine the relation between body size 
and sexual maturity (Roa et al., 1999). In our study, we 
applied a generalized linear model (GLM) with binomial 
error structure and logit link function by using the MASS 
package in R to model maturity of rainbow smelt (ICES'). 
All marine samples of rainbow smelt were assigned a 
maturity status (mature or immature) that was used as 
the response variable in the model, and TL was used as 
the explanatory variable. The GLM assumes that all sam¬ 
ples are independent and generates a multiyear estimate 
of the maturity ogive for sexes combined. The GLM esti¬ 
mates maturity at length when 50% of the population is 
mature (L 50 ). A maturity ogive was not prepared for the 
samples collected with a fyke net during spawning runs 
because immature rainbow smelt were rare or absent at 
the 4 stations at coastal rivers where this fyke net sam¬ 
pling occurred. 
Maturity and condition indices (Crim and Glebe, 1990) 
were generated to determine if maturation patterns were 
evident among age, size, and area subsets and to com¬ 
pare with earlier data sets (Chen, 1970; Clayton, 1976). 
Values of Fulton’s condition factor, calculated as K=( TW/ 
TL 3 )x 100,000, increase with increasing weight at a given 
length, an indication of better physical “condition” for a 
' ICES (International Council for the Exploration of the Sea). 
2008. Report of the workshop on maturity ogive estimation for 
stock assessment (WKMOG); Lisbon, Portugal, 3-6 June. ICES 
CM Documents 2008/ACOM:33, 72 p. [Available from website.] 
