ORIGINAL PAPER

The role of fecundity regulation and abortive maturationin the reproductive strategy of Norwegian spring-spawningherring (Clupea harengus)

James Kennedy • Richard D. M. Nash •

Aril Slotte • Olav S. Kjesbu

Received: 27 August 2010 / Accepted: 14 February 2011 / Published online: 1 March 2011

� Springer-Verlag 2011

Abstract This study investigates the reproductive strat-

egy, an important component in the estimation of stock

reproductive potential, in Norwegian spring-spawning

(NSS) herring (Clupea harengus), an iteroparous, extreme

capital spawner, through the estimation of fecundity over a

period of 3 years including two complete maturation cycles

and three spawning seasons. NSS herring have an ‘opti-

mistic’ strategy, with almost all adult herring caught in

August being in the vitellogenic stage of ovary develop-

ment, despite overwintering energy levels not being

determined at this time. Fecundity in the summer, i.e.,

more than half a year before spawning in spring (February–

April), was also much higher than could be supported by an

individual’s concurrent energy levels. Consequently,

fecundity was later reduced through atresia with the

majority of this occurring before overwintering. The total

reduction and the length of the time period in which the

reduction took place appeared to vary between years.

During the spawning season, atresia was mostly prevalent

in small first-time spawners\180 g and several individuals

aborted ovary development at this time. Final fecundity

varied between years with a difference of up to 18% and

was linked to annual variations in condition. In conclusion,

this extensive field study has demonstrated that each indi-

vidual herring can display a suite of size-specific repro-

ductive tactics to fine-tune oocyte production in response to

fluctuating levels of planktonic prey.

Introduction

There is a vast array of reproductive strategies among

fishes, and the particular strategy for a particular species

has important consequences at the population level in terms

of stock reproductive potential and sustainable fishing

mortality (Murua and Saborido-Rey 2003). Species that

partake in many breeding events must balance their energy

investment between the current year’s reproduction and

future reproduction to maximise their lifetime reproductive

output. However, when food is seasonal and unpredictable,

the decision of how much energy to invest in reproduction

can have large consequences, especially when development

of the gonads is a lengthy process. To cope with such

variations, animals must have the ability to alter their

reproductive investment as energy dictates. There are

several different mechanisms that can be employed to do

this including the reduction in fecundity in the current

reproductive cycle (Kurita et al. 2003; Kennedy et al.

2007), adjustment of the length of the spawning migration

(Slotte 1999a) or even the skipping of a reproductive event

entirely, either by not restarting ovary development or

through the complete reabsorption of the developing

oocytes (Rideout et al. 2005). These strategies can improve

lifetime fecundity by allowing for increased growth and

higher fecundity in future reproductive events (Jørgensen

et al. 2006) or simply by avoiding death due to low energy

reserves (Trippel and Harvey 1989). There are, thus,

Communicated by D. Righton.

J. Kennedy

Department of Biology, University of Bergen,

P.O. Box 7800, 5020 Bergen, Norway

R. D. M. Nash � A. Slotte � O. S. Kjesbu

Institute of Marine Research, P.O. Box 1870,

5817 Nordnes, Bergen, Norway

Present Address:J. Kennedy (&)

Møreforsking Marin, P.O. Box 5075, 6021 Alesund, Norway

e-mail: james@mfaa.no

123

Mar Biol (2011) 158:1287–1299

DOI 10.1007/s00227-011-1648-0

several factors that must be weighed against each other to

produce the best strategy.

Atlantic herring (Clupea harengus) are an example of a

teleost fish that must balance resources between current

and future reproduction. Based upon data collected by the

Institute of Marine Research (Norway), Norwegian spring-

spawning (NSS) herring have an estimated maximum age

of about 25 years and may spawn up to 20 times. The

biomass, Fulton’s condition factor (K) (see background

information in Nash et al. 2006) and migration routes have

fluctuated greatly in the past 100 years (Holst et al. 2002;

Ndjaula et al. 2010).

Herring are capital determinate spawners. This means

that gonad development is funded from stored reserves

built up during the feeding season. NSS herring feed in the

Norwegian Sea during the summer months with size-

dependent migration where larger individuals migrate to

more northerly and westerly feeding grounds (Misund et al.

1998). In late autumn, NSS herring currently migrate to

overwintering grounds in Northern Norway where they rely

on their stored reserves for energy and gonad growth

(Slotte 1999b). Selection of new overwintering grounds is

likely to be a stochastic process but may be influenced by

the ratio of recruit and repeat spawners (Huse et al. 2010).

In late winter to early spring, they migrate southwards for a

distance of up to 1,500 km from the wintering grounds to

spawn along the Norwegian coast. Unlike many herring

populations (see McQuinn 1997), NSS herring do not show

a strict spawning ground fidelity, rather larger better con-

dition fish migrate to more southerly spawning grounds

(Slotte 1999a). This state-dependent migration is probably

related to larval survival, as the southerly spawning

grounds may provide the resulting larvae with higher

temperatures; hence, the possibility for higher growth and

survival rates (Slotte and Fiksen 2000). To complete the

life cycle, the larvae drift northward from the spawning

grounds and spend the first 3–5 years as immature fish in

the Barents Sea. Thereafter, they migrate to join the adult

populations on the feeding grounds in the Norwegian Sea

(Dragesund 1970).

NSS herring is a suitable candidate for the study of

fecundity regulation as it is relatively long lived and

individuals [320 mm have a low growth rate, growing

\10 mm per year (Dragesund et al. 1980). Slow growth

excludes any confounding effects that growth in body size

can have on fecundity estimation (Kjesbu and Witthames

2007). In addition, herring have a short feeding season,

thus do not feed for a significant period during vitello-

genesis (Slotte 1999b; Kurita et al. 2003). This means that

the majority of the material used for gonad development

comes from stored reserves, thus eliminating any influence

that unaccounted feeding activity can have on fecundity.

During this study, we sampled extensively over 3 years,

NSS herring on the summer feeding grounds, the over-

wintering grounds and on the spawning grounds. This study

investigates the dynamics in maturity and fecundity at

different periods in the year to provide a comprehensive

understanding of the reproductive strategy in an extreme

capital spawner.

Materials and methods

Collection of samples

Herring samples were collected by pelagic trawl in 2006,

2007 and 2008 during the IMR research cruises and from

commercial fishing vessels in the herring summer feeding

area, overwintering area and spawning grounds (Table 1;

Fig. 1). Females were sampled randomly from the catch

and each individual was measured (total length, nearest

5 mm), weighed (total, gonad and intestine weight, nearest

1 g) and scales taken for age reading. A sub-sample of the

ovary was preserved in 10% buffered formalin and trans-

ported back to the laboratory for analysis. To confirm that

growth in length of NSS herring above 320 mm is low, as

growth rate is known to vary in other herring stocks

(Cardinale and Arrhenius 2000; Oskarsson 2008), the mean

length for year classes 1998–2002 were taken from the

IMR central database (CDB) and tracked for the period of

the study. The herring caught for our study were considered

to be a representative sample of the entire mature popu-

lation of NSS herring. The sampling during the feeding

season covered the major areas that herring migrate to

during the summer. During winter, in the period of our

study, the entire adult herring population was thought to be

concentrated into a relatively small area off Northern

Norway. In all years, samples at the spawning grounds

were taken from an extensive area covering most of the

known spawning grounds.

Some autumn spawners were caught during the sam-

pling; these were identified as herring with a gonadoso-

matic index (GSI) (calculation below) above 10% in the

summer (Husebø et al. 2005). Autumn spawners were

excluded from all analyses; therefore, all results pertain

only to NSS herring.

Estimation of fecundity

Fecundity was estimated using the Auto-diametric method

(Thorsen and Kjesbu 2001), which works on the principle

that the number of oocytes per gram of ovary is inversely

proportional to the average size of the oocytes. This rela-

tionship is based upon physical principles, as opposed to

biological, so is unlikely to be affected fish length, weight

or K. For ovaries containing vitellogenic oocytes, the

1288 Mar Biol (2011) 158:1287–1299

123

diameter of 200 vitellogenic follicles were measured using

a binocular microscope at 79 magnification, a camera

displaying a live image and computer-aided automatic

particle analysis program (Image-J, http://www.rsbweb.

nih.gov/ij/) that analyses images captured from the camera.

For ovaries that did not contain vitellogenic oocytes, i.e.

oocytes with an insufficient level of contrast to the back-

ground, 50 randomly selected pre-vitellogenic oocytes

were measured manually using a dissecting microscope

and the manual measurement tool within the image ana-

lysis software. As pre-vitellogenic oocytes have a smaller

size range than vitellogenic oocytes, 50 oocytes were suf-

ficient to get an accurate average oocyte diameter.

The stages in the maturation of herring oocytes can be

related to their diameter. Oocytes \240 lm are in the pre-

vitellogenic stage, 250–375 lm in the cortical alveoli stage

and[375 lm are vitellogenic (Ma et al. 1998; Kurita et al.

2003). Fish with a leading cohort (LC) oocyte diameter

(average diameter of the largest 10% of oocytes)[250 lm

were considered to have commenced ovary development.

Fecundity (F) was calculated for fish with an average

(formalin fixed) oocyte diameter (OD) [400 lm using the

equation from Oskarsson et al. (2002):

F ¼ 1:708� 1010 � 1:04� OW0:936� �

� OD�2:301

OW ¼ Fresh ovary weight(g)

The original equation was based upon fixed ovary weight,

whereas the current study used fresh ovary weight; there-

fore, a correction factor was added to the equation to

correct for the approximate 4% increase in ovary weight

due to fixation (Oskarsson et al. 2002).

When using the Auto-diametric method, the estimation

of fecundity becomes inaccurate for gonads with high

atretic intensity. This is because the weight of the atretic

oocytes is taken into account, however, due to a greater

transparency, they are less likely to be detected, during the

image analysis, than normal oocytes. Atretic oocytes are

smaller than normal oocytes; thus, the average oocyte

diameter will be over-estimated and so fecundity will be

under-estimated. Therefore, ovaries with an atretic inten-

sity greater than 15% were excluded from fecundity ana-

lysis but were noted. Also, ovaries that contained hydrated

oocytes, which indicate spawning may have begun, were

excluded from the fecundity analysis.

K and GSI were calculated for each fish using the fol-

lowing equations.

K ¼ 100; 000� Total weight (gÞ=Length (mm)3

GSI ¼ 100� Gonad weight (g)=Total weight (g)

Table 1 The location, time of sampling, total number of NSS herring sampled (n), length range (mm) and percentage of herring sampled during

the study in pre-vitellogenic (PV) gonadal stage, cortical alveoli (CA), vitellogenic (V) and hydrating oocyte stage (H)

Location Month and year n Length range PV CA V H

Feeding area July–August 2006 408 275–400 1.0 2.9 96.1 0.0

Feeding area July–August 2007 204 250–375 3.0 8.3 88.7 0.0

Overwintering area October–November 2006 124 255–385 3.3 0.0 96.7 0.0

Spawning grounds February 2006 215 260–380 0.4 0.0 77.3 22.4

Spawning grounds February 2007 232 265–380 0.0 0.0 53.1 46.9

Spawning grounds February–March 2008 314 270–400 0.5 0.0 99.5 0.0

60°

65°

70°

75°(a)

66°

68°

70°

72°

74°

Lat

itu

de

(b)

15° 10° 5° 0° 5° 10° 15° 20°

10° 15° 20°

15° 10° 5° 0° 5° 10° 15° 20°

Longitude

60°

65°

70°

75°(c)

Fig. 1 Sampling location off the coast of Norway for NSS herring

during July–August (feeding season) 2006 (open diamond) and 2007

(solid circle) (a), October–November (overwintering) 2006 (opencircle) (b) and February (spawning season) 2006 (open square) 2007

(upward pointing triangle) and February–March 2008 (downwardpointing triangle) (c). Dashed rectangle (c) indicates location of b

Mar Biol (2011) 158:1287–1299 1289

123

A common criticism of K is that this variable shows a

length dependency with larger fish tending to have a higher

value of K. This was examined using common Pearson

correlation analysis in the current data sets for all years and

seasons independently. There was no significant correlation

between length and K in either of the feeding seasons.

However, there was a weak trend in the overwintering and

in the spawning season data in 2006, 2007 and 2008 (see

‘‘Results’’). Slotte (1999b) shows that the relative weight

loss during overwintering and migration is inversely pro-

portional to length in NSS herring. This indicates that the

weak positive correlation between length and K is a result

of greater weight loss, relative to body size, from feeding to

spawning rather than a statistical artefact and so is appro-

priate as a measure for weight at length for NSS herring.

Fecundity is determined by condition close to spawning

within a year (see ‘‘Results’’). In order to estimate whether

annual variations in condition can explain the annual

variations in fecundity, condition before spawning for the

population was quantified using the average weight of all

340-mm herring in the IMR CDB caught in January and

above 67�N for the years in question. Fecundity was cal-

culated using the fecundity-length equation from each year.

Weight at length decreases as herring migrate, thus, using

only fish from January and occupying north of 67�N gives

a representative estimate of condition between years

without the effect of including fish which had already

migrated large distances. The fecundity and weight of

340-mm herring were used for this comparison as it rep-

resents a length class that was about midway in the length

distributions of the repeat spawners ([320 mm) and thus

was considered representative of the whole repeat-spawner

part of the stock. To examine the differences in fecundity

within the two maturation cycles in the study, the mean

fecundity of a 340-mm fish was estimated using the

fecundity-length equations for each sampling period

(Table 2).

Estimation of atresia

The intensity of atresia was estimated in herring caught

during the spawning season that had a K \ 0.75 or with an

average oocyte diameter \1,200 lm. These criteria were

chosen because atresia rises sharply in herring with a

K \ 0.70 and is absent when the average oocyte diameter

is [1,200 lm (Oskarsson et al. 2002). In ovaries where

the average oocyte diameter was [800 lm, atresia was

estimated using whole mount preparations. Atretic oocytes

are distinguishable from normal oocytes as they are

irregular in shape, relatively smaller than normal oocytes

and have an uneven transparency (Oskarsson et al. 2002). Ta

ble

2T

he

fecu

nd

ity

–w

eig

ht

(F–W

),fe

cun

dit

y–

len

gth

(F–

L)

and

fecu

nd

ity

–le

ng

th–

Fu

lto

n’s

con

dit

ion

(F–

LK

)re

lati

on

ship

sfo

rN

SS

her

rin

gg

rou

ped

by

yea

ran

dse

aso

n,

wit

hth

en

um

ber

of

sam

ple

sin

the

rela

tio

nsh

ip(n

),th

er2

of

the

rela

tio

nsh

ipan

dth

eco

ntr

ibu

tio

no

fK

toth

er2

val

ue

(Kr2

)

Yea

rS

easo

nn

F–

Wre

lati

on

ship

r2F

–L

rela

tio

nsh

ipr2

F–L

Kre

lati

on

ship

r2K

r2

20

06

Fee

din

g3

98

F=

27

99

W-7

,96

80

.46

F=

77

09

L-1

.74

91

05

0.4

7F

=8

21

L?

61

,88

29

K-2

.45

91

05

0.4

8(0

.02

)

20

07

Fee

din

g1

87

F=

25

19

W-8

,35

60

.17

F=

56

19

L-9

.68

91

04

0.1

4F

=5

67

L?

66

,83

39

K-1

.66

91

05

0.1

6(0

.01

)

20

06

Ov

erw

inte

rin

g1

16

F=

26

29

W-1

6,7

17

0.8

6F

=7

04

9L

-1.7

09

10

50

.78

F=

65

7L

?1

04

,03

69

K-2

.40

91

05

0.8

5(0

.07

)

20

06

Sp

awn

ing

20

4F

=2

74

9W

-15

,85

10

.85

F=

70

49

L-1

.75

91

05

0.7

7F

=5

70

L?

98

,87

09

K-2

.03

91

05

0.8

5(0

.08

)

20

07

Sp

awn

ing

11

6F

=2

82

9W

-16

,09

80

.86

F=

76

49

L-1

.92

91

05

0.8

1F

=6

59

L?

71

,56

49

K-2

.10

91

05

0.8

5(0

.04

)

20

08

Sp

awn

ing

25

3F

=2

49

9W

-7,4

92

0.7

5F

=6

72

9L

-1.6

19

10

50

.72

F=

61

7L

?8

0,0

10

9K

-2.0

29

10

50

.78

(0.0

6)

Len

gth

(L)

=to

tal

len

gth

(mm

)an

dw

eig

ht

(W)

=to

tal

wei

gh

t(g

)

1290 Mar Biol (2011) 158:1287–1299

123

A minimum of 200 oocytes were counted, and the relative

intensity of atresia i.e. percentage of oocytes which were

atretic; 100 9 number of atretic oocytes/(number of nor-

mal and atretic oocytes) was calculated. Atresia was

estimated in ovaries containing oocytes \800 lm using

the profile counting method (Andersen 2003). A section of

the ovary was dehydrated in ethanol and embedded in

Technovit� resin. Several 4-lm sections were then cut and

stained with toluidine blue. To prevent the same oocytes

being present on different slides, the minimum distance

between each section was equal to the LC oocyte diame-

ter. The sections were examined under the microscope

with the total number of normal and atretic oocytes

counted (atretic oocytes were distinguished from normal

oocytes using the criteria from Hunter and Macewicz

(1985)), and the relative intensity was calculated. How-

ever, this intensity is an underestimation of the true

intensity of atresia as atretic oocytes are smaller than

normal oocytes so are less likely to be encountered in

histological sections. The relative intensity was then

adjusted using the equation from Gonzalez-Vasallo (2006)

who compared the intensity of atresia in herring ovaries

estimated using the profile counting method and the

unbiased physical disector method (Andersen 2003;

Kjesbu et al. 2010):

RIA ¼ 3:0881� A0:75ðr2 ¼ 0:93Þ

RIA, unbiased relative intensity of atresia; A, relative

intensity of atresia from profile counting. We considered

herring with an atretic intensity greater than 15% at the

time of sampling to be high because this level is outside the

range found in herring that are down regulating fecundity

during the autumn (Kurita et al. 2003).

Statistics

ln-transformed data were used for fitting linear regression

models between length and weight. Weight-, length- and

length-K-based models were fitted to un-transformed

fecundity data, as the assumption of linearity, normal dis-

tribution and homogeneity of variance between years was

met, for each sampling period and year using linear

regression. This was also done for the combined data from

the three spawning seasons. Variations in fecundity

between years were assessed using ANCOVA with either

length, weight or K as a covariate. A two-factor ANCOVA

was used to test the effect of ovary development stage on

fecundity in the summer season with length and LC oocyte

diameter as covariables.

The ability of the weight, length and length-K models

from the spawning season to accurately predict fecundity

was tested. A regression model was established using only

half the samples from each year (selected randomly using

the random function in Microsoft Excel), these were then

used to predict the fecundity of the remaining samples. The

observed fecundity was divided by the predicted fecundity.

The resultant values were plotted against their respective

predictor variable. To test the reliability of using previously

established relationships to predict fecundity for other years

where no year specific relationship is available, the length-

K model from the combined data was used to predict the

fecundity for the fish in Oskarsson et al. (2002) and Kurita

et al. (2003). The predicted values were then compared with

the actual fecundity estimated in the respective studies.

To examine the general effect of ovary development on

the strength of the fecundity-length/weight relationships,

all fish from the study were pooled and divided among ten

Table 3 Coefficient of determination (r2) values for the fecundity–

weight (W-r2), fecundity–length (L-r2) and fecundity–length ? Ful-

tons’s K (LK-r2) relationship and fecundity of a 340-mm (including

95% confidence intervals) NSS herring at different stages of ovary

development indicated by leading cohort (LC) oocyte diameter

LC oocyte diameter (lm) n W-r2 L-r2 LK-r2 K-r2 Fecundity

(thousands)

400–499 365 0.41 0.39 90 ± 2.2

500–599 148 0.38 0.33 0.41 0.08 76 ± 3.6

600–699 26 0.45 0.31 85 ± 13.0

700–799 22 0.80 0.57 0.81 0.24 74 ± 10.4

800–899 43 0.75 0.62 0.71 0.09 68 ± 4.9

900–999 53 0.88 0.81 0.85 0.04 66 ± 3.1

1,000–1,099 38 0.89 0.83 0.88 0.05 61 ± 4.2

1,100–1,199 168 0.83 0.76 0.83 0.07 68 ± 2.6

1,200–1,299 247 0.84 0.77 0.84 0.07 65 ± 1.3

1,300? 115 0.80 0.61 0.82 0.21 61 ± 2.1

Relationships are from pooled data for all sampling dates. A blank value for the LK-r2 indicates the addition of K to the regression was not

significant. K-r2 indicates the contribution of K in the LK relationship

Mar Biol (2011) 158:1287–1299 1291

123

groups according to LC oocyte diameter (Table 3) and the

fecundity of an average 340-mm herring was calculated for

each group. Differences in fecundity between groups were

tested using ANCOVA with length as covariate together

with a Scheffe post hoc test. The Scheffe post hoc test was

used due to unequal n between groups. Annual variations in

the average potential fecundity in the spawning season was

tested against the average weight of a 340-mm herring

from the IMR CDB using linear regression.

Results

Weight and age at length

There was no difference in weight at length, compared

using ln-transformed data, between the 3 years in the

feeding season or the 6 years in the spawning season

(ANCOVA; p \ 0.05), and these were best described using

linear regression (Table 4). The average length at age of

herring from age 5 (average length 310 mm (±10 mm))

increased by approximately 10 mm per year throughout the

period of the study.

All the herring caught during the summer had a

K greater than 0.70. There was no significant correlation

between length and K in either of the feeding seasons

(p [ 0.05). However, there was a weak trend in the over-

wintering (p = 0.006, r = 0.25) and in the spawning sea-

son data in 2006 (p \ 0.001, r = 0.50), 2007 (p = 0.006,

r = 0.48) and 2008 (p = 0.006, r = 0.28). These rela-

tionships were significantly different between years

(ANCOVA; p \ 0.0001).

Ovary development

The majority of the fish caught during the summer sam-

pling had clearly begun ovary development (Table 1) and

had an LC oocyte diameter between 350 and 600 lm

(Fig. 2a, b). However, several fish in the cortical alveoli

stage were also caught during the summer. The majority of

the fish caught in the overwintering grounds were in the

vitellogenic stage (Table 1) and had an oocyte diameter in

the range of 500–1,000 lm (Fig. 2c). On the spawning

grounds, the majority of the fish were either in the

vitellogenic stage or had began oocyte hydration (Table 1)

of the vitellogenic fish the oocytes were generally in the

range of 800–1,400 lm (Fig. 2d, e, f).

Ovary stage had a significant effect on fecundity with

highest fecundity at the beginning of vitellogenesis

(ANCOVA; p \ 0.0001). The general pattern showed that

fecundity decreased as ovary development progressed with

the greatest decreases occurring during early vitellogenesis

(Table 3; Fig. 3). There were no further significant

decreases in fecundity after the LC oocyte diameter

reached 600–699 lm (Table 5). There was, however, a

small but significant increase in fecundity between

900–999 and 1,000–1,099, followed by a decrease from

1,100–1,199 to 1,200–1,299 lm, but the fecundity at these

stages (1,000–1,099 and 1,100–1,199 lm) still remained

significantly less than the fecundity at the beginning of

ovary development (Table 5).

The explanatory power of weight and length on fecundity

was lowest at the beginning of ovary development and gen-

erally increased until the oocytes were around 900–1,099 lm,

it then showed a small decrease during the final part of ovary

development (Table 3). The effect of K on fecundity varied

throughout ovary development with no clear pattern.

Variation in fecundity

The linear regression models of fecundity-weight (Table 2)

had a tendency to systematically overestimate the fecundity

of herring in the lower weight range (Fig. 4a). However,

for the fecundity-length or the fecundity-length-K regres-

sions (Table 2), there was no systematic underestimation of

fecundity for herring in the lower length range (Fig. 4b, c).

There was a positive correlation between K and the

observed/predicted values from the fecundity-length

regression (p \ 0.001, r = 0.44) but not for the observed/

predicted values from the fecundity-weight or fecundity-

length-K regressions (p [ 0.05).

Fecundity in the feeding season was significantly higher

in 2007 than 2006, both when length and weight was a

covariate (ANCOVA; p \ 0.0001; Fig. 5). The LC oocyte

diameter was significantly lower in summer 2007 (mean =

431 lm, (SD = 61), n = 408) compared to summer 2006

(mean = 441 lm, (SD = 75), n = 206) (ANOVA; p \0.0001, Fig. 2), indicating a difference in ovary develop-

ment stage between years. As fecundity differences can be a

result of the development stage of the ovary the LC oocyte

diameter was included as a covariate in the ANCOVA

together with length/weight. This resulted in the year effect

on fecundity being insignificant (p [ 0.05).

There were significant differences in fecundity in the

spawning season between years when length, weight or

length together with K were used a covariates, both,

between years in the current study (Fig. 5) and between

Table 4 The weight (W; g)—length (total length, L; mm) relation-

ships for NSS herring grouped by season

Season n Relationship r2

Feeding 612 ln(W) = 2.59 9 ln(L) - 9.28 0.85

Overwintering 124 ln(W) = 3.23 9 ln(L) - 13.05 0.93

Spawning 761 ln(W) = 3.59 9 ln(L) - 15.26 0.92

1292 Mar Biol (2011) 158:1287–1299

123

years in the current study and years from previously pub-

lished studies (Oskarsson et al. 2002; Kurita et al. 2003)

(ANCOVA; p = 0.0001). The average fecundity of a

340-mm herring, calculated from the length fecundity

equations from the spawning seasons, varied between years

(ANOVA; p \ 0.0001), was lowest in 1999 and differences

between years were up to 18%. This variation was related

to the variation in the average weight of the 340-mm her-

ring taken from the IMR CDB i.e. the herring caught in

January and north of 67�N (linear regression; r2 = 0.73,

p = 0.031, Fig. 6), but was not correlated with the average

weight of the 340-mm herring used in the fecundity esti-

mates (linear regression; p [ 0.05).

The average fecundity of a 340-mm herring, calculated

from the length fecundity equations from the feeding and

spawning seasons, decreased significantly from 88 thou-

sand to 68 thousand over the maturation cycle starting in

summer 2006 and ending in spring 2007, a decrease of

23%. In the subsequent cycle, fecundity decreased from 93

thousand in summer 2007 to 68 thousand in spring 2008, a

decrease of 27%. The largest value of fecundity for a

340-mm herring in August 2006 and February 2007 was

129 thousand and 69 thousand, respectively, which equated

to a difference of 47%. For all length groups, weight-spe-

cific fecundity increased by 9.1% between October 2006

and February 2007 (ANCOVA; p \ 0.001) while length-

0

200

400

600

800

1000

1200

1400

1600

0

200

400

600

800

1000

1200

1400

1600L

C o

ocy

te d

iam

eter

(µm

)

(a) (b)

(c)

240 260 280 300 320 340 360 380 4000

200

400

600

800

1000

1200

1400

1600

240 260 280 300 320 340 360 380 400

Length (mm)

(d)

(f)(e)

Fig. 2 Leading cohort (LC)

oocyte diameter plotted against

length for NSS herring caught

on the feeding grounds in July–

August 2006 (a) and 2007 (b),

on the overwintering grounds in

October 2006 (c) and on the

spawning grounds in 2006 (d),

2007 (e) and 2008 (f). Points

marked with an x indicate fish

with an atretic intensity greater

than 15%. Dashed line at

250 lm indicates the threshold

for a fish to be considered to

have started ovary development

Mar Biol (2011) 158:1287–1299 1293

123

240 280 320 360 400

1300+

240 280 320 360 400

Length (mm)

1200-1299

240 280 320 360 4000

40

80

120

160

2001100-1199

1000-1099900-999

0

40

80

120

160

200

Po

ten

tial

fec

un

dit

y (t

ho

usa

nd

s)

800-899

700-799600-699

0

40

80

120

160

200500-599

0

40

80

120

160

200400-499

Fig. 3 Relation between length

and fecundity for NSS herring at

different stages of oocyte

development. Legend indicates

oocyte size (lm) class in the

graph. Linear regression lines

are shown

Table 5 Post hoc comparisons of fecundity at length (as covariate) between oocyte diameter groups

1 2 3 4 5 6 7 8 9

400–499 (1)

500–599 (2) ***

600–699 (3) *** 0.16

700–799 (4) *** *** 0.86

800–899 (5) *** *** 0.36 1.00

900–999 (6) *** *** 0.54 1.00 1.00

1,000–1,099 (7) *** *** *** 0.28 0.33 0.10

1,100–1,199 (8) *** *** *** 0.20 0.17 * 1.00

1,200–1,299 (9) *** *** 0.32 1.00 1.00 1.00 ** ***

1,300? (10) *** *** 0.84 1.00 0.97 1.00 *** *** 0.98

Values are p values with * p \ 0.05; ** p \ 0.001; *** p \ 0.0001

1294 Mar Biol (2011) 158:1287–1299

123

specific fecundity did not change (ANCOVA; p [ 0.05).

Weight at length decreased by 10% during the same period

(ANCOVA; p \ 0.001).

The resultant model from the combined fecundity data

from all three spawning seasons (2006–2008) was

F ¼ 611� Lengthþ 8:69� 104 � K � 2:07� 105

r2 ¼ 0:84; df ¼ 576; p\0:001

Using this equation to predict the fecundity in December

1996, January 1998 (Oskarsson et al. 2002) and February

1999 (Kurita et al. 2003) using the original length and weight

data gave an average underestimation of 7, 11 and 26%,

respectively, compared to the observed fecundity estimate

based upon gravimetric/Auto-diametric methodology.

Atresia

In 2006, 2007 and 2008, atresia was present in 16, 37 and

16% of the fish examined for atresia in the spawning

260 280 300 320 340 360 380 400

Length (mm)

0.00

1.00

2.00

3.00

Ob

serv

ed/p

red

icte

d

100 200 300 400 500

Weight (g)

0.00

1.00

2.00

3.00

260 280 300 320 340 360 380 400

Length (mm)

0.00

1.00

2.00

3.00

(a)

(b)

(c)

Fig. 4 The deviation from linear prediction of the fecundity and the

observed fecundity for fecundity versus weight (a) and length (b) and

length-K (c)

240 260 280 300 320 340 360 380 400

Length (mm)

0

40

80

120

160

200

Po

ten

tial

fec

un

dit

y (t

ho

usa

nd

s)

0

40

80

120

160

200 (a)

(b)

Fig. 5 Relation between length and fecundity, with regression lines

shown, for NSS herring caught in the feeding area during summer

1998 (open circle solid line), 2006 (open diamond dashed line) and

2007 (open square dash dot line) (a), and spring 2006 (open diamonddash dot line), 2007 (open square solid line) and 2008 (opendownward pointing triangle dashed line) (b)

Mar Biol (2011) 158:1287–1299 1295

123

seasons, respectively. Note that this result does not repre-

sent the prevalence in the whole population as these fish

were not selected randomly but due to a higher likelihood

that atretic oocytes would be present at that time of the

year. All fish, which had an atretic relative intensity greater

than 15%, which is considered to be high, had a

weight \180 g (Fig. 7), a length \ 305 mm and a K \0.72. The percentage of fish that had atresia accounted for

4, 2 and 4% of all the fish analysed for fecundity in the

2006, 2007 and 2008 spawning seasons, respectively.

These consisted of twenty 4-year-old fish and two 5-year-

old fish so are likely to be recruit spawners (ICES 2010).

Nine of these fish (eight 4-year-old and one 5-year-old) had

an atresia [95%.

Discussion

Skipped spawning is a common phenomena amongst tel-

eosts with the most common sub-group categories in

female wild fish being ‘resting’ (failure to begin ovary

development) and ‘reabsorbing’ (reabsorption of oocytes

through atresia) (Rideout et al. 2005). The resting category

is not prevalent in the NSS herring, as can be seen from the

fact that almost all herring caught in the feeding grounds

had begun ovarian development. This is probably because

all individuals were above the threshold value of K (0.70),

which guarantees the beginning of ovary development

(Kennedy et al. 2010). Even though some herring would be

above the threshold value of K when ovary development

begins, they may still not have accumulated sufficient

energy reserves to fully support ovarian development,

survive the winter and carry out extensive migrations.

However, after they have begun ovary development, her-

ring will feed for approximately another 5 months, which

may give them enough time to accumulate sufficient

reserves (Slotte 1999b; Kurita et al. 2003).

During early vitellogenesis, herring had a high fecun-

dity, which was poorly related with size or condition

(Table 3). As ovary development begins approximately

5 months before the end of the feeding period (Kurita et al.

2003), fecundity at that time cannot reflect the level of

energy reserves at the over wintering and spawning time,

because these have not yet been determined. Herring have

a determinate fecundity which means that after spawning

has commenced no more oocytes are recruited to the

developing cohort. For many determinate spawners,

including herring, there is a gap of several months between

the end of oocyte recruitment and spawning. This tempo-

rally restricted pulse of recruitment of a large number of

pre-vitellogenic oocytes to the developing pool allows

subsequent reductions of fecundity in order for it to reflect

the level of energy reserves present closer to the breeding

season. This allows the possibility of a higher fecundity

than would be achieved if fecundity was determined by the

energy reserves when the oocytes were initially recruited.

280 300 320 340 360

Weight (g)

45

50

55

60

65

70

75P

ote

nti

al f

ecu

nd

ity

(th

ou

san

ds)

1999

1997

1998

2006

20072008

Fig. 6 Average potential fecundity versus average weight for a

340-mm NSS herring. Numbers on the graph indicates year. Errorbars show 95% confidence limits. Line shows linear regression fit.

Data for 1997 and 1998 are from Oskarsson et al. (2002) and data for

1999 are from Kurita et al. (2003). Data for 1997 are from December

1996 but are applicable to the 1997 spawning season

100 200 300 400 500

Weight (g)

0

20

40

60

80

100

Rel

ativ

e in

ten

sity

of

atre

sia

(%)

Fig. 7 Relative intensity of atresia in relation to weight for NSS

herring caught on the spawning grounds between 2006 and 2008.

Dashed line at 15% indicates threshold above which intensity is

considered high

1296 Mar Biol (2011) 158:1287–1299

123

This down-regulation of fecundity occurs in other herring

populations (van Damme et al. 2009) and also seems to be

a common mechanism across other determinate spawners

too as it has been documented in, cod (Gadus morhua)

(Thorsen et al. 2006), plaice (Pleuronectes platessa)

(Kennedy et al. 2007) and Greenland halibut (Reinhardtius

hippoglossoides) (Kennedy et al. 2009).

As ovary development progresses, body size and con-

dition explain an increasing amount of the variation in

fecundity and fecundity down-regulation ceased during

overwintering. During overwintering, the rate of food

consumption by herring declines and they start to rely on

stored energy reserves (Slotte 1999b). After this point, it is

unlikely that the herring will gain any additional energy,

therefore fecundity is reduced to closely match body con-

dition. These results indicate that condition many months

before spawning plays an important role in the determi-

nation of fecundity in herring.

Our results showed a lower degree of down-regulation

compared to that found by Kurita et al. (2003), who

reported that fecundity decreased by 57% between July and

February, whereas our study showed a decrease of 23 and

27% between August and February in the years studied.

Kurita et al. (2003) estimated a higher potential fecundity

in July and a lower potential fecundity in February com-

pared to the present study. The range in the summer

fecundity from Kurita et al. (2003) is equivalent to the

variation in the fecundity seen in the current study. How-

ever, due to this high variation between individuals, a

larger number of samples are required to get a reliable

average for fecundity than was sampled in Kurita et al.

(2003), so Kurita et al. (2003) may have overestimated the

average fecundity which may explain some of the differ-

ence in the level of down-regulation. The lower fecundity

in February 1999 found by Kurita et al. (2003) is related to

the low condition of herring in that year and will also

account for some of the difference in down-regulation

between studies. This shows that fecundity differences

between years are brought about mainly by variation in the

degree of down-regulation of oocytes as opposed to vari-

ation in the number of oocytes beginning vitellogenesis.

The highest fecundity estimates for a 340-mm fish in

August 2006 (129,000) and February 2007 (69,000),

showed a difference of 47%, which is comparable to Kurita

et al. (2003)’s value of 57%. This leads us to conclude that

high levels of down-regulation are possible for individual

fish; however, these represent more of an extreme value in

individuals rather than the norm.

The largest part of the down-regulation in fecundity

occurred during early vitellogenesis. Between overwinter-

ing and spawning fecundity did not decrease despite further

decreases in body weight (Fig. 3). Kurita et al. (2003)

found that the largest decrease in fecundity occurred

between July and October, which is in agreement with the

current study. However, Kurita et al. (2003) state that the

fecundity continues to decreases between October and

January which is in contrast to the present study. The

reason for this difference in the two studies is not clear but

may represent flexibility in the fecundity regulation pro-

cess. In years of lower energy reserves or higher energy

depletion during overwintering, down-regulation may

occur for a greater part of the ovary development leading to

a lower fecundity; the weight of herring caught North of

67� in 1999 was the lowest of the 6 years studied (Fig. 6).

During the spawning season, there were many small fish

(\180 g) which exhibited high levels of atresia ([15%),

including some with relative intensities of [95% (Fig. 7).

It is likely that the all of these fish were recruit spawners as

deduced from their age. The herring with an atretic inten-

sity [95% would not have spawned in that season, prob-

ably due to insufficient energy reserves to complete the

maturation of their oocytes, thus abandoned ovary devel-

opment and spawning. These herring were probably

attempting to spawn for the first time as deduced from their

age, thus are not skipping spawning in the true sense of the

term (Rideout et al. 2005). Atretic intensities up to 100%

have previously been documented in herring caught close

to the spawning season (Oskarsson et al. 2002). These

consisted of herring with K less than 0.70, similar to the

current study. However, Oskarsson et al. (2002) did not

give details on the weight or age of these herring. Irre-

spective of that, it appears that this aborted maturation may

commonly occur in a small number of first-time spawners

each year. The reasons for this may be due to smaller

herring having insufficient energy reserves for spawning.

This could be a result of lower energy storage capacity due

to their small body size, smaller fish also have higher

weight-specific metabolic rates and swim less economi-

cally (Schmidt-Nielsen 1984). It is interesting to note that

the aborting of ovary development occurred very close to

the spawning season and on the spawning grounds indi-

cating the ‘decision’ not to spawn was made as far into

ovary development as was possible. It is currently unclear

whether the herring with atretic intensities between 15 and

60% would spawn or would go on to absorb the remaining

oocytes. The majority of these individuals had oocytes

between 400 and 1,000 lm, the stage in which the major

part of down-regulation occurs. However, the intensities

are higher than expected at that development stage (Kurita

et al. 2003), and their oocytes are less developed than the

other herring caught at the same time. It therefore seems

likely that these herring will not spawn and will go on to

reabsorb the remainder of their oocytes.

There were significant variations in fecundity, in the

spawning season, between years with a difference of up to

18% between years with this difference being significantly

Mar Biol (2011) 158:1287–1299 1297

123

correlated with change in body weight between years

(Fig. 6). Variations in the average potential fecundity of a

340-mm herring was better correlated with the average

body weight of 340-mm herring taken from the IMR CDB

than the average weight of the herring that were actually

involved in fecundity estimation. The weight of the herring

taken from the IMR CDB was limited to herring caught

from latitudes above 67�N. The herring above 67�N will be

mostly overwintering/early migrating herring and so give a

good reflection of condition before the start of the migra-

tion. The timing of sampling for fecundity during the

spawning season was not standardised between years and

the herring within a sampling period were also caught at

different stages of migration. This leads to a reduced ability

to accurately quantify the variability in condition between

years, and its effects on the fecundity, as herring which had

began migration, assuming they started with a similar

condition, will have a lower condition than those that had

not.

When estimated at a similar point in ovary development,

weight is generally considered to be the best predictor of

fecundity at the individual level (Koops et al. 2004;

Skjæraasen et al. 2006; Kennedy et al. 2007), provided

none of the fish have fallen below a critical energy point

where relative fecundity starts to fall (Kjesbu 2009).

However, weight-based fecundity linear models overesti-

mated the fecundity of herring in the lower weight range

(Fig. 4a). This is most likely due to the greater prevalence

and intensity of atresia seen in herring of this size leading

to a much lower fecundity. The best predictor of fecundity

at the individual level was a combination of length and

K. However, this equation did not prove useful for the

prediction of fecundity in other years with considerable

underestimations of fecundity when tested against observed

fecundity data from previous years. The reason for this is

likely due to variation between years in the migratory

distance the herring had undertaken before capture. This

would affect the energy reserves of the herring at the time

of sampling and so affect the predicted fecundity. Thus,

when using established fecundity models to backcalculate

total egg production for species that undertake spawning

migrations, it would be advisable to use the length-weight

relationships for individuals which have not began their

spawning migration.

The addition of K increased the explained variance in

fecundity in all sampling periods but the amount of vari-

ance explained varied through ovary development with no

distinct pattern (Table 3). Many previous studies have

found that measurements of energy reserves generally add

little to the explained variance in fecundity (see Rideout

and Morgan 2010 and references therein). Rideout and

Morgan (2010) described this as a problem of sampling

time which results in the masking of the effects of

condition i.e. sampling too early in development can mask

the effect of indices of condition if the effects of condition

on fecundity have not been fully expressed. On the other

hand, the effect of condition at the beginning of ovary

development on fecundity may not be expressed until close

to spawning, so sampling too close to spawning may mask

the effect as condition, both absolute and relative to other

fish, may have changed. As most of the down-regulation of

fecundity occurred during early development of the ovary

and there was no change in fecundity from overwintering

to the spawning season, it is clear, as mentioned earlier,

that energy reserves at the beginning of overwintering are

important for the determination of fecundity. However,

K may not be an accurate measure of energy levels in

herring which may be why it explains only a small and

varying amount of variance in the fecundity linear models

(Davidson and Marshall 2010).

Acknowledgments The study was funded by the Research Council

of Norway (NFR) project ‘‘Timing and determination of fecundity

and skipped spawning: implications for stock—recruitment theory of

determinate spawners’’ (173341/S40) and encouraged by discussion

with and the terms of references of the NAFO Working Group on

Reproductive Potential and COST Action Fish Reproduction and

Fisheries (FA0601). We would like to thank the crew and personnel

of the M/S Libas, M/S Gardar and R/S G.O. Sars for the collection of

samples. A warm thank you also to Anders Thorsen, Merete Fonn and

Bente Njøs Strand for help and advice in the analyses of samples and

to Cindy van Damme (IMARES, The Netherlands) and the anony-

mous readers and referees who gave valuable comments on earlier

drafts of the manuscript.

References

Andersen TE (2003) Unbiased stereological estimation of cell

numbers and volume fractions: the disector and the principles

of points counting. In: Kjesbu OS, Hunter J, Witthames PR (eds)

Modern approaches to assess maturity and fecundity of warm-

and cold-water fish and squids. Institute of Marine Research,

Norway, pp 11–18

Cardinale M, Arrhenius F (2000) Decreasing weight-at-age of

Atlantic herring (Clupea harengus) from the Baltic Sea

between 1986 and 1996: a statistical analysis. ICES J Mar

Sci 57:882–893

Davidson D, Marshall CT (2010) Are morphometric indices accurate

indicators of stored energy in herring Clupea harengus? J Fish

Biol 76:913–929

Dragesund O (1970) Distribution, abundance and mortality of young

and adolescent Norwegian spring spawning herring (Clupeaharengus, Linne) in relation to subsequent year class strength.

Fiskdir skr Ser havunders 15:451–556

Dragesund O, Hamre J, Ulltang Ø (1980) Biology and population

dynamics of the Norwegian spring-spawning herring. Rapp P-V

Reun Cons Int Explor Mer 177:43–71

Gonzalez-Vasallo BD (2006) Regulatory reproductive mechanisms

and somatic growth in Atlantic herring (Clupea harengus L.).

MPhil Thesis, University of Bergen, Bergen

Holst JC, Dragesund O, Hamre J, Misund OA, Østvedt OJ (2002)

Fifty years of herring migration in the Norwegian sea. ICES J

Mar Sci 215:352–360

1298 Mar Biol (2011) 158:1287–1299

123

Hunter JR, Macewicz BJ (1985) Rates of atresia in the ovary of

captive and wild northern anchovy, Engraulis mordax. Fish Bull

83:119–136

Huse G, Fernø A, Holst JC (2010) Establishment of new wintering

areas in herring co-occurs with peaks in the ‘first time/repeat

spawner’ ratio. Mar Ecol Prog Ser 409:189–198

Husebø A, Slotte A, Clausen LAW, Mosegaard H (2005) Mixing of

populations or year class twinning in Norwegian spring spawn-

ing herring? Mar Freshw Res 56:763–772

ICES (2010) Report of the working group on widely distributed

stocks (WGWIDE). ICES CM 2010/ACOM: 15. 482 pp

Jørgensen C, Ernande B, Fiksen Ø, Dieckmann U (2006) The logic of

skipped spawning in fish. Can J Fish Aquat Sci 63:200–211

Kennedy J, Witthames PR, Nash RDM (2007) The concept of

fecundity regulation in plaice (Pleuronectes platessa) tested on

three Irish Sea spawning populations. Can J Fish Aquat Sci

64:587–601

Kennedy J, Gundersen AC, Boje J (2009) When to count your eggs: is

fecundity in Greenland halibut (Reinhardtius hippoglossoidesW.) down-regulated? Fish Res 100:260–265

Kennedy J, Skjæraasen JE, Nash RDM, Thorsen A, Slotte A, Hansen

T, Kjesbu OS (2010) Do capital breeders like Atlantic herring

(Clupea harengus L.) exhibit sensitive periods of nutritional

control on ovary development and fecundity regulation? Can J

Fish Aquat Sci 67:16–27

Kjesbu OS (2009) Applied fish reproductive biology: contribution of

individual reproductive potential to recruitment and fisheries

management. In: Jakobsen T, Fogarty MJ, Megrey BA, Moks-

ness E (eds) Fish reproductive biology: implications for assess-

ment and management. Wiley-Blackwell, Chichester,

pp 293–332

Kjesbu OS, Witthames PR (2007) Evolutionary pressure on repro-

ductive strategies in flatfish and groundfish: relevant concepts

and methodological advancements. J Sea Res 58:23–34

Kjesbu OS, Fonn M, Gonzales BD, Nilsen T (2010) Stereological

calibration of the profile method to quickly estimate atresia

levels in fish. Fish Res 104:8–18

Koops MA, Hutchings JA, McIntyre TM (2004) Testing hypotheses

about fecundity, body size and maternal condition in fishes. Fish

Fish 5:120–130

Kurita Y, Meier S, Kjesbu OS (2003) Oocyte growth and fecundity

regulation by atresia of Atlantic herring (Clupea harengus) in

relation to body condition throughout the maturation cycle. J Sea

Res 49:203–219

Ma Y, Kjesbu OS, Jørgensen T (1998) Effects of ration on the

maturation and fecundity in captive Atlantic herring (Clupeaharengus). Can J Fish Aquat Sci 55:900–908

McQuinn IH (1997) Metapopulations and the Atlantic herring. Rev

Fish Biol Fish 7:297–329

Misund OA, Vilhjalmsson H, Jakupsstovu SHI, Røttingen I, Belikov

S, Asthorsson O, Blindheim J, Jonsson J, Krysov A, Malmberg

SA, Sveinbjørnsson S (1998) Distribution, migration and

abundance of Norwegian spring spawning herring in relation to

the temperature and zooplankton biomass in the Norwegian Sea

as recorded by coordinated surveys in spring and summer 1996.

Sarsia 83:117–127

Murua H, Saborido-Rey F (2003) Female reproductive strategies of

marine fish species of the North Atlantic. J North Atl Fish Sci

33:23–31

Nash RDM, Valencia AH, Geffen AJ (2006) The origin of Fulton’s

condition factor—setting the record straight. Fisheries

31:236–238

Ndjaula HO, Nash RDM, Slotte A, Johannessen A, Kjesbu OS (2010)

Long-term changes in the total egg production of Norwegian

spring-spawning herring Clupea harengus (L.)—implications of

variations in population structure and condition factor. Fish Res

104:19–26

Oskarsson GJ (2008) Variation in body condition, fat content and

growth rate of Icelandic summer-spawning herring Clupeaharengus L. J Fish Biol 72:2655–2676

Oskarsson GJ, Kjesbu OS, Slotte A (2002) Predictions of realised

fecundity and spawning time in Norwegian spring-spawning

herring (Clupea harengus). J Sea Res 48:59–79

Rideout RM, Morgan MJ (2010) Relationships between maternal

body size, condition and potential fecundity of four north-west

Atlantic demersal fishes. J Fish Biol 76:1379–1395

Rideout RM, Rose GA, Burton MPM (2005) Skipped spawning in

female iteroparous fishes. Fish Fish 6:50–72

Schmidt-Nielsen K (1984) Scaling. Why is animal size so important?

Cambridge University Press, Cambridge

Skjæraasen JE, Nilsen T, Kjesbu OS (2006) Timing and determina-

tion of potential fecundity in Atlantic cod (Gadus morhua). CanJ Fish Aquat Sci 63:310–320

Slotte A (1999a) Effects of fish length and condition on spawning

migration in Norwegian spring spawning herring (Clupeaharengus L.). Sarsia 84:111–127

Slotte A (1999b) Differential utilization of energy during wintering

and spawning migration in Norwegian spring-spawning herring.

J Fish Biol 54:338–355

Slotte A, Fiksen Ø (2000) State-dependent spawning migration in

Norwegian spring-spawning herring. J Fish Biol 56:138–162

Thorsen A, Kjesbu OS (2001) A rapid method for estimation of

oocyte size and potential fecundity in Atlantic cod using a

computer-aided particle analysis system. J Sea Res 46:295–308

Thorsen A, Marshall CT, Kjesbu OS (2006) Comparison of various

potential fecundity models for north-east Arctic cod Gadusmorhua, L. using oocyte diameter as a standardizing factor.

J Fish Biol 69:1709–1730

Trippel EA, Harvey HH (1989) Missing opportunities to reproduce—

an energy-dependent or fecundity gaining strategy in white

sucker (Catostomus commersoni). Can J Zool 67:2180–2188

van Damme CJG, Dickey-Collas M, Rijnsdorp AD, Kjesbu OS (2009)

Fecundity, atresia and spawning strategies of Atlantic herring.

Can J Fish Aquat Sci 66:2130–2141

Mar Biol (2011) 158:1287–1299 1299

123