166 
Fishery Bulletin 117(3) 
Implications for populations of rainbow smelt 
A synopsis of information on stocks of rainbow smelt in 
the Gulf of Maine led to the conclusion that sharp declines 
have occurred in fisheries for and abundance of rainbow 
smelt in recent decades (Enterline et al., 2012). In Mas¬ 
sachusetts, participation in rainbow smelt fisheries has 
declined to very low levels, and fisheries have ceased in 
most estuarine rivers where winter ice is no longer suffi¬ 
cient to support fishing. However, empirical measures of 
changes in population data related to this declining trend 
are absent. The fecundity data presented here indicate 
that reproductive potential has been lost with the decline 
in the proportion of rainbow smelt in spawning runs 
that are age 3 or older. This change is most pronounced 
in the spawning runs in the Jones River, where our data 
indicate that large increases in age-1 rainbow smelt and 
decreases in rainbow smelt age 3 or older have occurred. 
The results of analysis indicate declines of 20-40% in 
EPR from previous studies to this study for populations 
sampled in the Jones and Parker Rivers, and these esti¬ 
mates are most influenced by changes in Z, which does 
not include partially recruited age-1 rainbow smelt. Given 
these age structure changes and that rainbow smelt age 
3 or older are nearly 5 times as fecund as age-1 rainbow 
smelt, spawning runs today may be recruitment limited. 
Despite this implication, the high proportion of age-1 rain¬ 
bow smelt in spawning runs observed in this study may 
be important for population survival: current spawning 
runs would experience even lower EPR with lower levels 
of maturation among age-1 fish. Further investigations on 
recruitment relationships and fecundity of rainbow smelt 
with sampling from a wider geographic range and longer 
temporal scale are needed to assist management and res¬ 
toration efforts for rainbow smelt. 
Size at age has been reported to decline with increasing 
latitude for spawning populations of anadromous rainbow 
smelt (McKenzie, 1964; Pouliot, 2002; Enterline et al., 
2012; O’Malley et ah, 2017). A large majority of landlocked 
and anadromous rainbow smelt in populations from east¬ 
ern Maine northward historically matured at age 2 (Rupp, 
1959; McKenzie, 1964; Chen, 1970). Age-1 rainbow smelt 
are known to spawn at variable rates in U.S. waters of 
the Gulf of Maine, with higher percentages found in Mas¬ 
sachusetts, lower percentages in rivers of western Maine, 
and near absence in Canadian Maritime provinces (Warfel 
et al., 1943; McKenzie, 1964; Murawski and Cole, 1978; 
Pouliot, 2002; Enterline et ah, 2012). In Massachusetts, 
age-2 rainbow smelt have been found to form the pre¬ 
dominant cohort in spawning runs, with low, fluctuating 
percentages observed for age-1 and age-3 cohorts, age-4 
fish being uncommon, and fish of ages 5-6 rarely observed 
(Murawski and Cole, 1978; Lawton et al. 8 ). 
Temperature-related differences in growth rates may 
be a cause of the decline in proportion of age-1 rainbow 
smelt in spawning runs with increasing latitude. Murawski 
and Cole (1978) back-calculated length at age from scales 
of rainbow smelt from the Parker River and found close 
agreement with the actual mean length at age, except for 
age-1 fish. The lower back-calculated size at age-1 indi¬ 
cates that only larger age-1 rainbow smelt were recruited 
to the spawning run. The proportion of age-1 females was 
also lower than that of age-1 males in the spawning run, 
and Z was lower in females than in males (Murawski and 
Cole, 1978). These findings suggest a trade-off for matura¬ 
tion of rainbow smelt in which fast-growing rainbow smelt 
that mature in their first year may realize greater lifetime 
reproductive potential but are exposed to higher mortality 
than fish that mature later. Both our study and the pre¬ 
vious studies of rainbow smelt in Massachusetts found 
a higher proportion of males than females among age-1 
fish in spawning runs (Murawski and Cole, 1978; Sutter, 
1980; Lawton et al. 8 ). The difference between sexes may be 
related to the longer schedule and larger energy allocation 
required for oogenesis than for spermatogenesis that has 
been found in landlocked rainbow smelt (Chen, 1970; Foltz 
and Norden, 1977). 
The maturation schedules of species of Pacific salmon 
(Oncorhynchus spp.) are likely the best studied of all diadro- 
mous species. Natural selection influences maturation strat¬ 
egies in Pacific salmon species as reflected by oocyte size and 
fecundity (Campbell et al., 2006). Phenotypic differences in 
egg size and maturation in species of Pacific salmon have 
been related to water temperature and growth rates, with 
fast growth, early maturation, high fecundity, and small egg 
size thought to be a response to lower survival rates (Mur¬ 
ray and McPhail, 1988; Kinnison et al., 1998). Further work 
is needed to determine if the selective pressures on Pacific 
salmon species are similar to those on anadromous rainbow 
smelt, given their phylogenetic proximity. 
Luey and Ad el man (1984) suggested that landlocked pop¬ 
ulations of rainbow smelt are prone to selective pressures 
on their reproductive life history because they are small, 
fast growing, and short-lived. The variability of size and 
age in spawning populations of anadromous rainbow smelt 
in the broader geographic region of New England (Chase 
et al., 2009; Enterline et al., 2012; O’Malley et al., 2017) sup¬ 
ports the expectation of a link between growth rates and 
maturation schedule. Results from genetic analyses of rain¬ 
bow smelt in our study area by Kovach et al. (2013) provide 
evidence of gene flow among the 4 locations where fyke nets 
were deployed and indicate that larval dispersal is most 
influenced by circulation patterns in the Gulf of Maine and 
that the circulation patterns were more influential to pop¬ 
ulation genetic structure than natal homing. More infor¬ 
mation is needed to determine environmental influences, 
maturation thresholds, and population recruitment conse¬ 
quences across a wider geographic range for the species. 
Our data provide evidence of differences in length and 
suggest that there are differences in growth among popula¬ 
tions of rainbow smelt over a small spatial scale. However, 
the spatial and temporal scales of our data do not allow 
inferences on interpopulation differences in reproductive 
life history traits. Further work is needed to determine the 
nature of phenotypic divergence in reproductive life history 
of anadromous rainbow smelt across the geographic range 
of this species and on the population-level consequences of 
phenotypic plasticity in age and size at maturity. 
