Sinusoidal swimming in fishes: the role of season, density of large zooplankton, fish length, time of the day, weather condition and solar radiation

PRIMARY RESEARCH PAPER

Sinusoidal swimming in fishes: the role of season,density of large zooplankton, fish length, timeof the day, weather condition and solar radiation

Oldrich Jarolım • Jan Kubecka • Martin Cech •

Mojmır Vasek • Jirı Peterka • Josef Matena

Received: 18 February 2010 / Revised: 7 July 2010 / Accepted: 27 July 2010 / Published online: 12 August 2010

� Springer Science+Business Media B.V. 2010

Abstract The sinusoidal swimming of fish, previ-

ously interpreted as foraging behaviour, was studied

with respect to season, density of large zooplankton,

fish length, time of the day, weather condition and

solar radiation in Rımov Reservoir, Czech Republic,

using a bottom-mounted, split-beam transducer (7�,

nominal angle; frequency 120 kHz). The proportion

of sinusoidally swimming fish increased from April

to August while this behaviour was absent in

October. The occurrence of sinusoidal swimming

showed an apparent pattern throughout the day; it

increased sharply around sunrise, was highest within

5–6 h around solar noon, and sharply decreased

around sunset. Significantly less frequent occurrence

of sinusoidal swimming was recorded during cloudy

days compared to sunny days. The vast majority of

records came from fish of standard length ranging

from 100 to 400 mm, which represents the typical

size range of common bream Abramis brama and

roach Rutilus rutilus of age [1?, the main zoo-

planktivores in the reservoir. The presence of these

larger fish in the open water of the reservoir, as well

as the presence of sinusoidal swimming, apparently

correlates with the presence of large zooplankton

(Daphnia, Leptodora and Cyclops vicinus) in the

epilimnion. The increase of sinusoidal swimming

between April, June and finally August resulted in an

increase of zooplankton component in fish guts. It

appears that high values of solar radiation, and stable

calm weather during high pressure periods, result in

optimal optical conditions for sinusoidal swimming,

making this foraging behaviour more efficient and

widely used in fishes exploiting the zooplankton

production in the reservoir.

Keywords Common bream Abramis brama �Daphnia � Echosounder � Leptodora � Roach Rutilus

rutilus � Rımov Reservoir � Sonar5

Introduction

The mode of swimming of a fish has consequences

for its ability to escape from predators or unfavour-

able environmental conditions and its reproductive

behaviour, as well as its foraging (Wootton, 1998).

With respect to swimming patterns and locomotion

of fish, it is generally assumed that there are indi-

vidual modes of straight swimming (Lindsey, 1978;

Webb, 1984a, b; Videler, 1993), i.e. swimming in a

Handling editor: Luiz Carlos Gomes

O. Jarolım � J. Kubecka � M. Cech (&) �M. Vasek � J. Peterka � J. Matena

Biology Centre, Academy of Sciences of the Czech

Republic, Institute of Hydrobiology, Na Sadkach 7,

370 05 Ceske Budejovice, Czech Republic

e-mail: [email protected]

123

Hydrobiologia (2010) 654:253–265

DOI 10.1007/s10750-010-0398-1

horizontal plane. Much less attention has been paid to

fish swimming in a vertical plane represented by, e.g.

gliding swimming patterns (Weihs, 1973, 1974) or

sinusoidal swimming patterns (Cech & Kubecka,

2002).

Haberlehner (1988) was the first to notice roach

Rutilus rutilus swimming up-and-down and consum-

ing plankton in a backwater of the Danube (SCUBA

diving observation). Cech & Kubecka (2002) first

described the term ‘sinusoidal swimming’ on the

basis of stationary uplooking acoustic data from the

open water of Rımov Reservoir, Czech Republic

(North Sea-drainage area). This term was used

because the trajectories of ‘sinusoidally swimming’

pelagic fishes resembled a regular sinusoidal curve

when displayed on the echogram (Fig. 1). The

patterns of change in target strength (TS; Simmonds

& MacLennan, 2005) revealed that the sinusoidal

swimming is an active swimming mechanism, rather

than a mechanism based on swim-bladder volume

changes, comprising tilting of the fish’s body during

the ascending and descending phase of a sinusoidal

cycle. The average amplitude of a sinusoidal cycle

(distance between the uppermost and the lowermost

positions of the sinusoidal movement curve) was ca.

1 m and the average frequency of cycling was nearly

4 cycles min-1.

Sinusoidal swimming was detected during all

observations from June to August and 83% of the

fish [100 mm standard length (LS) exhibited this

behaviour. This movement pattern was not observed

in November. A smaller data set from May showed

less intense occurrence of sinusoidal swimming in the

open water, mainly because of coincidence with

the spawning of the dominant cyprinid species in the

reservoir (Cech & Kubecka, 2002). Sinusoidal swim-

ming was also closely dependent on time of day and

weather conditions. It started after sunrise and was

replaced by straight swimming before sunset. Sinu-

soidal swimming was never observed during the night

(Cech & Kubecka, 2002) or during extremely bad

weather (heavy rain, storms, strong winds; Cech,

pers. observation).

Following the work of Janssen (1981, 1982) and

Thetmayer & Kils (1995) dealing with visual fish

feeding, and considering the above findings, it has

been argued that sinusoidal swimming is an efficient

method for fish visually searching for prey, mainly

large zooplankton (Daphnia, Leptodora), whose spa-

tial distribution in Rımov Reservoir is patchy in

summer (foraging behaviour; Cech & Kubecka,

2002). Daphnia and Leptodora are the most important

prey items of adult cyprinid species occupying the

open water of the reservoir in summer (Vasek et al.,

2003; Vasek & Kubecka, 2004; Vasek et al., 2008).

These two cladocerans are highly preferred in the diet

of the larger common bream Abramis brama and of

roach, the main zooplanktivores in Rımov Reservoir,

having Ivlev electivity indices of 0.68–0.95 (Cech &

Kubecka, 2002).

The present study is focused on sinusoidal swim-

ming with respect to the time of year, weather

conditions, solar radiation and the presence of juvenile

and adult fish in the open water of Rımov Reservoir,

from which this unique behaviour was first reported in

detail. The main questions were: (1) Is the sinusoidal

swimming widely used in fishes only during summer or

could it be observed similarly through the spring and in

early fall? (2) Does the proportion of sinusoidally

swimming fish in the open water fish stock as well as

the total abundance and biomass of the open water fish

stock increased during late spring and summer com-

pared to early spring? (3) Is this increase to any extent

correlated to the abundance of large zooplankton and

its species composition? (4) Is the sinusoidal swim-

ming really performed only by a specific size cohort of

fish (136–359 mm LS; according to Cech & Kubecka,

2002) or is it also widely used by both smaller and

larger fish? (5) Is the sinusoidal swimming influenced

by weather conditions and intensity of solar radiation

(i.e. by optical conditions), or is the number of fish

performing this behaviour independent of those abiotic

factors?

1

3

5

79

11

13

15

17

19

Dep

th (m

)

sinusoidal

'uncertain'

straight

Fig. 1 An example of sinusoidal, ‘uncertain’ and straight

trajectories of fish swimming in a vertical plane (raw 40 LogR

TVG uplooking echogram)

254 Hydrobiologia (2010) 654:253–265

123

Methods

Study area

The study was carried out in the meso- to eutrophic

Rımov Reservoir, Czech Republic (48�500N, 14�290E;

170 km south of Prague), which has an area of 210 ha

and maximum surface elevation of 471 m a.s.l. The

maximum depth of the reservoir is 45 m and the

volume 33.6 millions m3 (Seda et al., 2000). This

dimictic reservoir is inhabited by 30 species of fish,

five of which, common bream, roach, bleak Alburnus

alburnus, perch Perca fluviatilis and ruff Gymno-

cephalus cernuus, have the highest ecologically

relevant abundances. The reservoir has had a stable

fish composition, with a relatively low proportion of

predatory fish, since 1988–1989 when a percid phase

was replaced by a cyprinid phase (Seda & Kubecka,

1997; Rıha et al., 2009).

Sampling

A scientific echosounder, Simrad EY 500, along with

an ES 120-7G circular split-beam transducer (nom-

inal angle 7.1�) was used for obtaining the data.

Signal frequency was 120 kHz, pulse length 0.1 ms,

pulse interval 0.2 s and output power 63 W. As in the

previous work of Cech & Kubecka (2002), the

uplooking transducer was placed at a depth of 38 m

ca. 1 m above the bottom, in the lacustrine part of the

reservoir near the course of the old Malse River

(for location see Vasek et al., 2008). The transceiver

and PC were placed 80 m away from the study area

in the enclosed floating boat shed.

The echosounder was continually in operation

from April to October 2005. Four limnologically

distinctive periods were chosen for further analysis:

(1) the beginning of spring stratification (14–16

April); (2) the clear-water phase (30 May–5 June); (3)

the summer maximum of phytoplankton and zoo-

plankton (8–12 August) from which there is a

comparable data set from a previous study of Cech

& Kubecka (2002); (4) the beginning of autumn

mixing (12–14 October). During all these periods,

there was stable, sunny weather, except in June and

August when sunny days alternated with cloudy days.

Simultaneously with the acoustic observations,

fish were sampled using a series of Nordic multi-

mesh gillnets (epi-, meso- and bathypelagic nets

4.5 m high, 16 panels 2.5 m wide each, mesh size

from knot to knot 5.5, 6, 8, 10, 12.5, 16, 19.5, 24, 29,

35, 43, 55, 70, 90, 110, 135 mm) and a purse seine

(120 m long, 12 m deep, front/mid/rear mesh size

6/8/10 mm; used in August only) (Vasek et al.,

2008). Gillnets were installed 2 h before the sunset

and lifted 2–3 h after the sunset on each sampling

occasion. The exposition time throughout sunset was

chosen on the basis of previous studies that found

increased gillnet catchability at twilight period

(Vasek et al., 2009) and gut fullness of planktivorous

fish markedly higher during the day and evening than

in the morning and at night (Vasek & Kubecka,

2004). Extensive purse seining was performed during

daylight hours (noon–evening). In total, 850 fish were

captured by gillnets and 182 fish were captured using

purse seine. Fish B200 mm LS were immediately

anesthetised in MS 222 (500 mg l-1; left in the

solution for 10 min), identified to the species, mea-

sured to the nearest 5 mm LS and transferred into

10% formaldehyde solution. Fish [200 mm LS were

killed by overdosing in MS 222 (1 g l-1; left in the

solution for 10 min), dissected immediately and only

their guts were preserved in 5% formaldehyde for

later processing (Vasek et al., 2008). The death of all

fish prior to immersion into the formaldehyde or prior

to dissection was assured by decapitation.

Zooplankton was collected by duplicate vertical net

hauls (net diameter 20 cm, mesh size 200 lm) drawn

through 5–0 m depth, which approximately corre-

sponded to the extent of the epilimnion during the

periods studied (Vasek et al., 2008). These gillnet,

purse seine and diet data, along with information on the

density and composition of the reservoir zooplankton,

have been described elsewhere (Vasek et al., 2008).

The acoustic records were processed by Sonar5 Pro

post-processing software. A total of 5,898 fish were

manually tracked, i.e. echoes were combined into

tracks and counted, since automatic tracking was not

able to distinguish individual fish when their trajec-

tories overlapped (Balk & Lindem, 2000). The TS

threshold was set to -56 dB. Time of day, depth in the

water column, TS, trajectory shape and change of

vertical range were recorded for individual fish. Three

basic types of fish swimming trajectories in the

vertical plane were distinguished—sinusoidal, straight

and ‘uncertain’. Sinusoidal swimming was defined as

swimming along the trajectory that resembled a

sinusoid curve with at least one full sinusoidal cycle

Hydrobiologia (2010) 654:253–265 255

123

on the echogram (Cech & Kubecka, 2002). A straight

trajectory (straight swimming) resembled a straight

line with no apparent change of vertical range.

Sometimes the fish trajectory was classified as

‘uncertain’ in cases when it was not swimming

sinusoidally (no regular up-and-down movement)

but did seem to change its vertical range (‘uncertain’

swimming; Fig. 1). Fish trajectories shorter than 15 s

(ca. the average duration of one single sinusoidal cycle

according to Cech & Kubecka, 2002; 586 trajectories

in total) were excluded from further analysis to

minimise misinterpretation of sinusoidal and ‘uncer-

tain’ trajectories. In total (day and night, April to

October records), 2,496 straight trajectories, 941

‘uncertain’ trajectories and 1,875 sinusoidal trajecto-

ries were included into individual analyses.

Vertical profiles of temperature (�C) and dissolved

oxygen (mg l-1) were measured with a calibrated

OXI 196 probe (WTW, Germany). Meteorological

data were recorded by an HBI meteorological probe

MS16 (Fiedler-Magr, Czech Republic), which com-

prises a global radiation probe GR01, temperature

probe HST005 and rain gauge SR02. Global radiation

(GR, in W m-2), comprised of direct radiation,

reflected radiation from the ground and diffused

radiation of the visible spectrum (Gjelland et al.,

2004), was used for defining day and night. The time

interval when GR = 0 W m-2 was defined as sunrise/

sunset, GR [ 0 W m-2 corresponds to day (daylight

hours).

Calculations and statistics

The LS of fish was calculated from the acoustic TS by

using the formula:

LS ¼ 10 TS�mð Þ=n; ð1Þ

where n = 19.65, m = -92.8; ventral aspect (Fro-

uzova et al., 2005). For sinusoidally swimming fish

only TS readings in the uppermost and lowermost

positions of the sinusoidal cycle, when the fish are not

tilted, were considered (Cech & Kubecka, 2002).

Weights of fish were calculated using the length–

weight relationship:

W ¼ a� LbS; ð2Þ

where a = 1.0943 9 10-5, b = 3.1387. Parameters

a and b were obtained from length–weight

relationships for common bream and roach caught

in Rımov Reservoir in 2005. Those two cyprinids

made up the majority of the pelagic fish stock that

year (Peterka et al., 2007). Biomass of fish was

calculated for each hour using the formula:

B ¼ W � N � t

S� 10; ð3Þ

where B is total biomass (kg ha-1), W is hourly

averaged fish weight (g), N is the number of

observations per h, S is water surface area covered

by the acoustic beam (m2), t is the average time spent

in the beam (s) and 10 is the coefficient for correction

to kg ha-1.

Abundance of fish (ind. ha-1) was computed as:

A ¼ B

W: ð4Þ

The role of season and density of large zooplankton

The data were tested using regression analysis

[course of (1) fish abundance, (2) fish biomass and

(3) proportion of sinusoidal fish in the open water of

Rımov Reservoir from April to August; course of the

mass of zooplankton in the contents of fish guts from

April to August] and t test (depth distribution of

sinusoidally compared to straight swimming fish).

The role of time of the day and fish length

The data were tested using ANOVA (proportion of

juvenile fish in the open water fish stock during day

compared to night; comparison of LS of sinusoidal

fish during individual periods of the year) and

Wilcoxon rank sum test (size composition of the

stock of sinusoidally compared to straight swimming

fish).

The role of time of the day, weather condition

and solar radiation

The difference in percentage of sinusoidal fish during

sunny compared to cloudy days in June and August

was tested using t test.

Time series analyses were performed in R (a free

software environment for statistical computing; pack-

age stats version 2.7.0). A seasonal-trend decompo-

sition procedure, based on the loess model (STL,

nonparametric analogy of Fourier analysis; see

256 Hydrobiologia (2010) 654:253–265

123

Becker et al., 1988; Cleveland et al., 1990), was

applied. The seasonal component (originally Sv;

Cleveland et al., 1990) was substituted by the daily

trend component (Dv). The trend component (origi-

nally Tv; Cleveland et al., 1990) was substituted by

the general trend component (GTv). Remainders

(originally Rv; Cleveland et al., 1990) were substi-

tuted by residuals (Rv). Raw data (Yv) are equal to the

sum of all components (the original equation in

Cleveland et al., 1990, was adapted with regard to

changes in names of individual components):

Yv ¼ GTv þ Dv þ Rv: ð5ÞThe model worked on the basis of iterations (two

inner iterations were used), when the significances of

all fitted elements were compared with each other and

with the variability of the residuals. R functions, ts

(with argument freq equal to 24) and stl (with

argument s.window equal to ‘periodic’ and argument

t.window equal to 6), were used to create and plot

time series.

All regression analyses, ANOVAs, Wilcoxon rank

sum tests and t tests were performed in Statistica

(StatSoft).

Results

The role of season and density of large

zooplankton

The acoustic abundance of fish as well as their

acoustic biomass in the open water of Rımov Reser-

voir steadily increased from April to August (regres-

sion analysisday abundance; F1,11 = 5.12, r2 = 0.32,

P \ 0.05; regression analysisday biomass; F1,11 =

26.68, r2 = 0.71, P \ 0.001; Fig. 2a, b). The only

exceptions were the nights in June when very low

abundance, and biomass of fish was recorded (coin-

cidence with intensive bream spawning events in the

littoral zone). There were also very few fish in

October in the open water, especially the daytime

abundance and biomass was very low (over 8 times

lower abundance and 16 times lower biomass com-

pared to August).

The temperature profiles showed sharp stratifica-

tion in April, June and August with an apparent

thermocline shallower than 10 m and relatively

weak temperature stratification in October with a

thermocline below 25 m. A lack of dissolved oxygen

in the deepest strata was recorded in all months studied

except April, but in all periods the oxygen values were

lower in the thermocline compared to the epilimnetic

layers. The water transparency was highest during the

clear-water phase at the beginning of June (Fig. 3).

Daphnia dominated the reservoir zooplankton in

late May/early June (28.8 ind. l-1) and in August

(31.7 ind. l-1). The density of Leptodora peaked in

August (0.6 ind. l-1). In April, large cladoceran

species were almost absent from the reservoir

(0.4 ind. l-1), but they were partly replaced by larger,

early-spring cyclopoid copepods (14.6 ind. l-1)

mostly Cyclops vicinus. In October, both large

cladocerans and cyclopoid copepods had almost

vanished from the open water of the reservoir

(2.5 ind. l-1) (for more details see Vasek et al., 2008).

0

50

100

150

200

250

300

350

400

Abu

nd

ance

[in

d. h

a-1

h-1

]

0

50

100

150

200

250

April June August October

Bio

mas

s [k

g h

a-1

h-1

]

(a)

(b)

Fig. 2 a Abundance (mean ? SD) and b biomass

(mean ? SD) of all fish during (square) day and (filled square)

night in April, June, August and October 2005 in the open

water of Rımov Reservoir

Hydrobiologia (2010) 654:253–265 257

123

The proportion of sinusoidal fish clearly increased

from April to August (regression analysisall fish;

F1,11 = 10.82, r2 = 0.50, P \ 0.01). On the other

hand, this swimming behaviour was almost absent in

October (Fig. 4a–f). The increase in the occurrence of

sinusoidal swimming between April, June and,

finally, August resulted in an increase of zooplankton

in the contents of fish guts (regression analyses;

F1,90 = 41.10, P \ 0.001, combining simultaneous

data from this study and Vasek et al., 2008).

The mean depths of the fish trajectories during the

day differed between the sinusoidally and straight

swimming fish in April (t test; df = 555, P \ 0.001)

and in August (t test; df = 1,246, P \ 0.001), and at

both these times, the sinusoidally swimming fish

were higher in the water column (i.e. closer to the

water surface), whereas the depths were not signif-

icantly different in June (t test; df = 1,780,

P = 0.95) (Fig. 5). All the time, the vast majority

of sinusoidally swimming fish (�90%) occurred in

the epilimnion.

The role of time of the day and fish length

The sinusoidal swimming pattern was observed

throughout the day and in both larger (LS [ 100 mm;

Fig. 4a, b) and small fish (LS B 100 mm; Fig. 4c, d).

However, the vast majority of sinusoidal trajectories

were sourced from larger fish and day records (95.8%

of all observations; Fig. 6a–h). A few fish (51

individuals; sum of all April, June, August and

October records) continued swimming sinusoidally

after sunset. This behaviour was observed in 21

individuals up to 20 min after sunset and in 22

individuals up to 1 h after sunset. Sinusoidal swim-

ming was observed not any longer than 80 min after

sunset. No such swimming trajectory was observed

before sunrise.

The length frequency distributions in April showed

a distinct peak of 1? fish [2004 year class (YC);

Fig. 6a, b]. The situation was very similar in the June

records (Fig. 6c, d), while August records showed the

appearance of 0? fish in the open water (2005 YC;

Fig. 6e, f). Those 0? fish clearly dominated both day

and night records in October (Fig. 6g, h). The

proportion of small fish (LS B 100 mm) in the open

water fish stock was higher at night than during the

day (ANOVA; F1,11 = 3.52, P \ 0.01). No peak of

0? fish was observed in April because there were no

0? fish, while in June the reason for the lack of a 0?

peak is the fact that 0? fish were probably too small

and they fell below the acoustic threshold (cf. Cech

et al., 2005; Cech & Kubecka, 2006).

Fish with sinusoidal swimming patterns comprised

the majority of observations of individuals LS C

200 mm during the day (53.8% in April, 75.4% in

June and 75.9% in August). The proportion of fish of

LS \ 200 mm with a sinusoidal swimming mode

noticeably increased in August. The LS of sinusoidal

fish differed significantly between individual months

(ANOVA; F3,1872 = 12.82, P \ 0.001), being on

average 243 mm (min–max 49–718 mm) in April,

295 mm (min–max 55–783 mm) in June and 237 mm

(min–max 90–512 mm) in August. The size compo-

sitions of the samples of sinusoidally swimming fish

and straight swimming fish differed significantly in

0

10

20

30

0 5 10 15 20 5 10 15 205 10 15 205 10 15 200 0 0

(a) (b) (c) (d)

S. depth1.5 m

S. depth5.0 m

S. depth1.5 m

S. depth3.1 m

Dep

th [m

]

T [oC] O2 [mg l-1]

Fig. 3 Comparison of the

vertical distribution of

temperature (solid line) and

dissolved oxygen (dashedline) measured during the

noon period in Rımov

Reservoir in a April, b June,

c August and d October

2005. The transparency, as

Secchi disc depth, is given

for each sampling period

258 Hydrobiologia (2010) 654:253–265

123

April (Wilcoxon rank sum test; W = 37,426, P \0.001), in June (Wilcoxon rank sum test; W =

498,406, P \ 0.001) and in August (Wilcoxon rank

sum test; W = 206,870, P \ 0.001). Sinusoidally

swimming fish were always larger than straight

swimming fish (60–130 mm LS difference, indicating,

however, more likely difference in foraging cohort

membership than difference in species identity). The

October data were not tested because of the lack of

sinusoidal trajectories (only two records of sinusoi-

dally swimming fish).

The role of time of the day, weather condition

and solar radiation

The proportion of sinusoidal trajectories in a

sum of trajectories of ‘potentially sinusoidal’ fish

0%

20%

40%

60%

80%

100%

0%

20%

40%

60%

80%

100%

0%

20%

40%

60%

80%

100%

April June August October April June August October

n=465 n=1363 n=1038 n=46

n=209 n=756 n=430 n=83

n=674 n=2119 n=1468 n=129

n=38 n=29 n=149 n=124

n=104 n=90 n=349 n=39

n=142 n=119 n=498 n=163

(a)

(f)(e)

(d)(c)

(b)Fig. 4 Composition of

individual types of fish

swimming trajectories

(white square straight,

gray-shaded square‘uncertain’ and blacksquare sinusoidal) in the

open water of Rımov

Reservoir during individual

months. a Day: larger fish

(LS [ 100 mm), b night:

larger fish (LS [ 100 mm),

c day: small fish

(LS B 100 mm), d night:

small fish (LS B 100 mm),

e day: all fish, f night: all

fish

Apr

il -

sin

Apr

il -

str

June

- s

in

June

- s

tr

Aug

ust -

sin

Aug

ust -

str

0

5

10

15

20

25

Dep

th [m

]

Fig. 5 Mean depths of sinusoidal (sin) and straight swimming

fish (str) during the day in April, June and August in the open

water of Rımov Reservoir. Plots = the median, 10th, 25th,

75th and 90th percentiles as vertical boxes with error bars and

(circle) outliers

Hydrobiologia (2010) 654:253–265 259

123

[LS [ 100 mm, depth B4 m according to Cech &

Kubecka (2002)] showed an apparent pattern

throughout the day in April, June and August; it

increased sharply around sunrise, was highest

within 5–6 h around solar noon, and sharply

decreased around sunset (Fig. 7). This pattern

was described by a sigmoidal model (Eq. 6), which

in all cases explained more than 59% of the

variability:

s ¼ a

1� e�x�x0

bð Þ; ð6Þ

where x represents time shift from solar noon, x0

corresponds to time shift from solar noon when GR =

0 W m-2 (i.e. sunrise/sunset), b determines slope of

the curve, a indicates maximum % of sinusoidally

swimming fish during solar noon and s shows calcu-

lated % of sinusoidally swimming fish at a given time.

N

0

5

10

15

20

25

30

35

N

00 100 200 300 400 500 600 0 100 200 300 400 500 600

5

10

15

20

25

30

35

N

0

10

20

30

40

50

60

N

0

10

20

30

40

50

60

70

1+ 1+

N

0

5

10

15

20N

0

20

40

60

80

100

120

140

1+ 1+

N

0

20

40

60

80

100

120

N

020406080

100120140160180200

+0+0

(b)(a)

(d)(c)

(f)(e)

(g) (h)

0+0+

L s (mm)

Fig. 6 Frequency

distributions of standard

length (LS) of fish with

(filled square) sinusoidal

and (square) straight

swimming patterns during

day and night in April

(a, b), June (c, d), August

(e, f) and October (g, h).

Note that sinusoidal

swimming was negligible in

October—only 0.63% of all

fish performed this

behaviour. Day records

(a, c, e, g), night records

(b, d, f, h). Very small 0?

fish in October (smaller

than in August) came from

a portional spawning of

common bream and bleak in

July (M. Vasek, pers.

observation) or these are

smallest individuals of

pelagic 0? perch, still prior

to their final migration

into the littoral zone (Cech

& Kubecka, 2006; Vasek

et al., 2006)

260 Hydrobiologia (2010) 654:253–265

123

The percentage of sinusoidal fish during the day

was dependent on weather conditions (amount of

clouds, surface global radiation). Fish performed

sinusoidal swimming preferentially during sunny

days in June and August (62 and 78% of fish swam

sinusoidally) rather than during cloudy days in June

and August (30 and 42% of fish swam sinusoidally)

(t testJune; df = 66, P \ 0.01; t testAugust; df = 44,

P \ 0.01; Fig. 8). The similarity between daily

trends and the general trend of GR and % of

sinusoidal fish was highly apparent. There was a

decline of both GR and % of sinusoidal fish during

the cloudy days.

Discussion

The presence of larger fish (LS [ 100 mm) in the

open water of Rımov Reservoir, as well as the

occurrence of sinusoidal swimming, apparently cor-

relates with the presence of large zooplankton

(Cladocera) in the epilimnion. Daphnia dominated

the reservoir zooplankton in late May/early June and

in August, while the density of Leptodora, the largest

zooplankter in the reservoir open water, peaked in

August (Vasek et al., 2008). At the same time, 60.3

and 64.9% of fish were observed to be swimming

sinusoidally. In April, large cladoceran species were

almost absent from the reservoir, but they were partly

replaced by larger, early-spring cyclopoid copepods

mostly Cyclops vicinus (Vasek et al., 2008). Still,

42.4% of the fish were performing sinusoidal swim-

ming. In contrast, in October both large cladocerans

and cyclopoid copepods had almost vanished from

the open water of the reservoir (Vasek et al., 2008).

At the same time, there were very few larger fish in

the open water and sinusoidal swimming had almost

ceased (\5% of trajectories). The extremely low

occurrence of sinusoidal swimming in October cor-

responds with previous findings of Cech & Kubecka

(2002) from November, when no fish were observed

using this movement pattern.

In their previous work, Cech & Kubecka (2002)

argued that sinusoidal swimming in fishes relates to

planktivory since the period of sinusoidal swimming

coincided with the period of intensive feeding

on zooplankton, mainly Daphnia and Leptodora.

Thetmayer & Kils (1995) have shown that the

0

20

40

60

80

100%

of s

inus

oida

l fis

h

0

20

40

60

80

100

0 2 4 6 8 10 120

20

40

60

80

100

x0=6.80

x0=7.70

x0 =6.90

(a)

(b)

(c)

Time shift from noon [h]

Fig. 7 The relationship between the percentage of sinusoidal

fish in the open water of Rımov Reservoir and the time shift

from solar noon (noon is shown as 0 h, i.e. 13 h of Central

European summer time) in a April, b June and c August. Each

point represents the hourly average of each processed day. The

vertical dotted line demonstrates both sunrise and sunset,

which well corresponds to the inflection point of the curve.

Note that, e.g. 6 h means both 7 and 19 h of Central European

summer time and that the biggest change in sinusoidal

swimming occurred directly around sunrise and sunset. The

sigmoidal relationship was highly significant for all tested

months (regression analysisApril; F1,56 = 40.68, r2 = 0.59, P \0.001; regression analysisJune; F1,125 = 110.55, r2 = 0.64,

P \ 0.001; regression analysisAugust; F1,76 = 181.90, r2 =

0.83, P \ 0.001)

Hydrobiologia (2010) 654:253–265 261

123

transparency of zooplankton found in open water is

an excellent strategy for staying invisible to horizon-

tally scanning eyes. However, planktonic prey might

appear more visible against the bright light of the sky

(darker due to absorbance at the opaque parts of the

body) or dark depths (brighter due to light scattered

in their tissue) compared with the less contrasting

background directly in front of fishes swimming in a

horizontal plane (Janssen, 1981, 1982; Lazzaro,

1987; Thetmayer & Kils, 1995). During the ascend-

ing and descending phase of the sinusoidal cycle

(ca. 75% of total swimming time), the fish tilts its

body to 30� compared to the horizontal plane, i.e.

facing towards the surface when ascending and

towards the bottom when descending in the water

column (Cech & Kubecka, 2002). This tilt of the fish

body is not far from an optimal attack angle when

even transparent prey is clearly visible to the predator

(Thetmayer & Kils, 1995).

Using an underwater camera in the open water of

Rımov Reservoir, Peterka et al. (2007) revealed that

in August 2005 sinusoidal swimming was typical for

Fig. 8 Time series analysis

of a, c surface global

radiation (GR) and

b, d percentage of

sinusoidal fish trajectories

in five sequential days in

June (a, b) and in four

sequential days in August

(c, d) showing an apparent

relationship between

sinusoidal swimming and

global radiation. Percentage

of sinusoidal fish was

relatively low in the first

(June, August) and fourth

days (June); those days

were very cloudy (see

arrows in trend component

of both time series). The

increase of sinusoidal

swimming was clear during

other days, which were

relatively cloudless. Note

that the curves describing

the general trends of GR

and percentages of

sinusoidal fish are very

similar. The thick grey lineson the right side of each plot

gives the same interval of

GR (300 W m-2) or

percentage of sinusoidal

fish (22%) for all panels.

Weather icons are used to

simplify the day values of

global radiation

262 Hydrobiologia (2010) 654:253–265

123

larger individuals of common bream and roach,

which dominated the record. Both these species were

frequently seen swimming sinusoidally, both as

single fish or in groups of several individuals. The

fish were clearly searching for zooplankton and they

were sucking them in.

Simultaneous purse seine catches confirmed that

common bream and roach of age [1? comprised

84% of the pelagic fish stock in Rımov Reservoir

(Peterka et al., 2007) and gut content analyses

confirmed that zooplankton was the only prey of

those two species (Vasek et al., 2008). Common

bream and roach were also abundant in pelagic gillnet

catches from April to August while, in contrast, in

October these species completely vanished from

pelagic gillnet catches (Vasek et al., 2008). It seems

that the absence of both large cladocerans and

cyclopoid copepods, along with the disappearance

of thermal and oxygen stratification in the upper

25 m of the water column, forced the common bream

and roach to leave the open water of the reservoir.

This fact resulted in (1) very low abundance and

biomass of larger fish in the open water and (2) very

low occurrence of sinusoidal swimming. Under such

unfavourable feeding conditions, it might be disad-

vantageous for these two species to stay longer in the

open water and perform even sinusoidal swimming,

an efficient foraging behaviour. On the other hand,

the increase in the occurrence of sinusoidal swim-

ming between April, June and, finally, August clearly

resulted in an increase of zooplankton in the contents

of fish guts.

It is evident that sinusoidal swimming is a visual

feeding behaviour of zooplanktivorous fishes such as

common bream and roach. This assumption is well

supported by Vasek & Kubecka (2004) who have

described circadian feeding activity of adult common

bream and roach with clear afternoon peaks and

distinct nighttime declines in gut fullness. Sinusoidal

swimming is restricted to daylight hours and to the

epilimnion of the reservoir (Cech & Kubecka, 2002,

this study) where there are appropriate light (Cech

et al., 2005), temperature, oxygen (Bohl, 1980;

Prchalova et al., 2008; Vasek et al., 2008) and feeding

conditions (Hudcovicova & Vranovsky, 2006; Seda

et al., 2007). Only 51 individuals out of 1,875

sinusoidally swimming fish performed this behaviour

outside daylight hours. Since 43 of those fish were

observed within 1 h after sunset, those records could

be classified as an example of residual behaviour.

However, it could also be hypothesized that moonlight

and, e.g. light from the dam construction or a nearby

village, in combination with eyes equipped with a

tapetum lucidum (common bream; Kuhne & Sewall,

1880) may improve optical conditions to an extent that

during the first part of the night this foraging behaviour

is still reasonable for some unsatiated individuals.

Generally, the disappearance of sinusoidal swimming

after sunset coincided with lower abundance and

biomass of larger fish in the open water, a result of

nocturnal inshore migrations (Schulz & Berg, 1987;

Kubecka, 1993; Kubecka & Duncan, 1994; Rıha et al.,

2008).

Preliminary results have shown that sinusoidal

swimming is also dependent on the weather condi-

tions. It was observed during sunny days and stable

weather, but heavy rain, storms and strong winds

stopped this behaviour and, moreover, led to a rapid

emigration of larger fish from the open water

(M. Cech, pers. observation). In the present study,

those findings are confirmed by the fact that a

significantly lower occurrence of sinusoidally swim-

ming fish was recorded during cloudy days compared

to sunny days. It appears that high values of GR and

stable calm weather during high pressure periods,

result in optimal optical conditions for sinusoidal

swimming, making this foraging behaviour more

efficient. Using sinusoidal swimming in the pelagic

habitat, both common bream and roach are able to

ingest specifically large zooplankton and avoid

cyanobacterial clumps, which are of the same size

and highly abundant in the reservoir during the

summer months (Znachor et al., 2006). In the evening,

the gut contents of common bream and roach were

found to consist of mainly cladocerans (consumption

up to 100,000 Daphnia individuals fish-1 day-1;

M. Vasek, unpubl. data) but cyanobacterial clusters

were completely absent (Vasek et al., 2008).

In contrast to the original study of Cech & Kubecka

(2002, nanalysed sinusoidal fish = 201), when using more

powerful data (this study; nanalysed sinusoidal fish

= 1,875) sinusoidal swimming was observed in a

much wider size range of fish (min–max LS

49–783 mm). The vast majority of records ([90% in

April–August), however, came from fish of LS

100–400 mm (average ca. 250 mm), which corre-

sponds well to the previous results of Cech &

Kubecka (2002) and represents the typical size range

Hydrobiologia (2010) 654:253–265 263

123

of common bream and roach of age [1? in Rımov

Reservoir (Vasek et al., 2008, 2009). The largest

sinusoidally swimming fish were most probably

common carps Cyprinus carpio L. since those fish

were sporadically seen on the UW camera records

(Peterka et al., 2007). This illustrates the fact that

sinusoidal swimming is widely used by all size

categories of fishes (except small 0? fish) exploiting

the zooplankton production in the reservoir.

Recently, a visual observation confirmed intensive

sinusoidal swimming for larger common bream

during June peak of Daphnia and Leptodora also in

case of Vır Reservoir, Czech Republic (Black Sea-

drainage area; sunny days) (M. Cech, unpubl. data).

Conclusion

The present study has shown that sinusoidal swim-

ming, the foraging behaviour, is widely used in fishes

during spring and summer, which periods correlate

with presence of large zooplankton (Daphnia, Lept-

odora, Cyclops vicinus) in the epilimnion of Rımov

Reservoir. The behaviour was observed for fish of LS

49–783 mm, however, the vast majority of records

came from fish of LS 100–400 mm, which corre-

sponds to older (age[1?) common bream and roach,

the dominant zooplanktivorous fish in the reservoir.

The sinusoidal swimming is influenced by weather

conditions and intensity of solar radiation. Fish

performed sinusoidal swimming preferentially during

sunny days rather than during cloudy days. The

proportion of fish performing sinusoidal swimming

increased sharply around sunrise was the highest

within 5–6 h around solar noon and sharply

decreased around sunset. Both relationships between

sinusoidal swimming and (1) weather conditions and

(2) intensity of solar radiation indicate the need of

very good optical conditions for this foraging

behaviour.

In order to conclude these findings to a conceptual

model, the sinusoidal swimming in fishes could be

observed (probability *100%) when following

requirements are fulfilled: (1) Both large zooplank-

ton, particularly Daphnia, Leptodora, Cyclops vicinus

and larger (LS [ 100 mm) zooplanktivorous fishes,

particularly common bream and roach, are present in

the open water of lakes and reservoirs during spring

and summer, (2) calm, sunny days occur.

Acknowledgments The authors thank J. Turek for providing

the environmental parameters data, M. Morris for careful

reading and correcting the English and two anonymous referees

for valuable comments on an earlier version of the manuscript.

The study was supported by the Grant Agency of the Czech

Republic (project No. 206/07/1392 and 206/09/P266) and the

Academy of Sciences of the Czech Republic (project No.

1QS600170504 and AVOZ60170517).

References

Balk, H. & T. Lindem, 2000. Improved fish detection in data

from split-beam sonar. Aquatic Living Resources 13:

297–303.

Becker, R. A., J. M. Chambers & A. R. Wilks, 1988. The New

S Language: A Programming Environment for Data

Analysis and Graphics. Wadsworth & Brooks/Cole

Advanced Books & Software, Monterey.

Bohl, E., 1980. Diel pattern of pelagic distribution and feeding

in planktivorous fish. Oecologia 44: 368–375.

Cech, M. & J. Kubecka, 2002. Sinusoidal cycling swimming

pattern of reservoir fishes. Journal of Fish Biology 61:

456–471.

Cech, M. & J. Kubecka, 2006. Ontogenetic changes in the

bathypelagic distribution of European perch fry Percafluviatilis monitored by hydroacoustic methods. Biologia,

Bratislava 61(2): 211–219.

Cech, M., M. Kratochvıl, J. Kubecka, V. Drastık & J. Matena,

2005. Diel vertical migrations of bathypelagic perch fry.

Journal of Fish Biology 66: 685–702.

Cleveland, R. B., W. S. Cleveland, J. E. McRae & I. Terpen-

ning, 1990. STL: a seasonal-trend decomposition proce-

dure based on loess. Journal of Official Statistics 6: 3–73.

Frouzova, J., J. Kubecka, H. Balk & J. Frouz, 2005. Target

strength of some European fish species and its dependence

on fish body parameters. Fisheries Research 7: 86–96.

Gjelland, K. Ø., T. Bøhn, F. R. Knudsen & P. Amundsen, 2004.

Influence of light on swimming speed of coregonids in

subarctic lakes. Annales Zoologici Fennici 41: 137–146.

Haberlehner, E., 1988. Comparative analysis of feeding and

schooling behaviour of the Cyprinidae Alburnus alburnus,

Rutilus rutilus and Scardinius erythrophthalmus in a

backwater of the Danube near Vienna. Internationale

Revue des Gesamtes Hydrobiologie 73: 537–546.

Hudcovicova, M. & M. Vranovsky, 2006. Vertical distribution

of pelagial zooplankton in a middle-sized dimictic valley

reservoir. Biologia 61: 171–177.

Janssen, J., 1981. Searching for zooplankton just outside Snell’s

window. Limnology and Oceanography 26: 1168–1171.

Janssen, J., 1982. Comparison of searching behaviour for

zooplankton in obligate planktivore, blueback herring

(Alosa aestivalis) and facultative planktivore, bluegill

(Lepomis macrochirus). Canadian Journal of Fisheries and

Aquatic Sciences 39: 1649–1654.

Kubecka, J., 1993. Night inshore migration and capture of adult

fish by shore seining. Aquaculture and Fisheries Man-

agement 24: 685–689.

Kubecka, J. & A. Duncan, 1994. Low fish predation pressure in

the London reservoirs: I. species composition, density and

264 Hydrobiologia (2010) 654:253–265

123

biomass. Internationale Revue gesampten Hydrobiologie

79: 143–155.

Kuhne, W. & H. Sewall, 1880. On the physiology of the retinal

epithelium. Journal of Physiology 3: 88–92.

Lazzaro, X., 1987. A review of planktivorous fishes: their

evolution, feeding behaviour, selectivities, and impacts.

Hydrobiologia 146: 97–167.

Lindsey, C. C., 1978. Form, function and locomotory habits in

fish. In Hoar, W. S. & D. J. Randall (eds), Fish Physiology

– Locomotion, Vol. 7. Academic Press, London: 1–100.

Peterka, J., M. Cech, M. Vasek, T. Juza, V. Drastık, M.

Prchalova, J. Kubecka & J. Matena, 2007. Fish occurrence

in the open water habitat of the eutrophic canyon shaped

Rımov reservoir (Southern Bohemia): comparing indirect

and direct methods of investigation. In Kubecka, J. (ed),

Fish Stock Assessment Methods for Lakes and Reservoirs.

HBU BC AVCR, Ceske Budejovice: 42 pp.

Prchalova, M., J. Kubecka, M. Vasek, J. Peterka, J. Seda,

T. Juza, M. Rıha, O. Jarolım, M. Tuser, M. Kratochvıl,

M. Cech, V. Drastık, J. Frouzova & E. Hohausova, 2008.

Distribution patterns of fishes in a canyon-shaped reser-

voir. Journal of Fish Biology 73: 54–78.

Rıha, M., J. Kubecka, T. Mrkvicka, M. Prchalova, M. Cech, V.

Drastık, J. Frouzova, M. Hladık, E. Hohausova, O. Jar-

olım, T. Juza, M. Kratochvıl, J. Peterka, M. Tuser & M.

Vasek, 2008. Dependence of beach seine net efficiency on

net length and diel period. Aquatic Living Resources 21:

411–418.

Rıha, M., J. Kubecka, M. Vasek, J. Seda, T. Mrkvicka, M.

Prchalova, J. Matena, M. Hladık, M. Cech, V. Drastık, J.

Frouzova, E. Hohausova, O. Jarolım, T. Juza, M. Kra-

tochvıl, J. Peterka & M. Tuser, 2009. Long-term devel-

opment of fish populations in the Rımov Reservoir.

Fisheries Management and Ecology 16: 121–129.

Schulz, U. & R. Berg, 1987. The migration of ultrasonic-tag-

ged bream, Abramis brama (L.), in Lake Constance

(Bodensee-Untersee). Journal of Fish Biology 31:

409–414.

Seda, J. & J. Kubecka, 1997. Long-term biomanipulation of

Rımov reservoir (Czech Republic). Hydrobiologia 345:

95–108.

Seda, J., J. Hejzlar & J. Kubecka, 2000. Trophic structure of

nine fish reservoirs regularly stocked with piscivorous

fish. Hydrobiologia 429: 141–149.

Seda, J., K. Kolarova, A. Petrusek & J. Machacek, 2007.

Daphnia galeata in the deep hypolimnion: spatial

differentiation of a ‘‘typical epilimnetic’’ species. Hyd-

robiologia 594: 47–57.

Simmonds, E. J. & D. N. MacLennan, 2005. Fisheries

Acoustics: Theory and Practice. Blackwell Science Ltd,

Oxford.

Thetmayer, H. & U. Kils, 1995. To see and not to be seen.

Marine Ecology Progress Series 126: 1–8.

Vasek, M. & J. Kubecka, 2004. In situ diel patterns of zoo-

plankton consumption by subadult/adult roach Rutilusrutilus, bream Abramis brama, and bleak Alburnus al-burnus. Folia Zoologica 53: 203–214.

Vasek, M., J. Kubecka & J. Seda, 2003. Cyprinid predation on

zooplankton along the longitudinal profile of a canyon-

shaped reservoir. Archiv fur Hydrobiologie 156: 535–550.

Vasek, M., J. Kubecka, J. Matena & J. Seda, 2006. Distribution

and diet of 0? fish within a canyon-shaped European

reservoir in late summer. International Review of

Hydrobiology 91(2): 178–194.

Vasek, M., O. Jarolım, M. Cech, J. Kubecka, J. Peterka &

M. Prchalova, 2008. The use of pelagic habitat by cypri-

nids in a deep riverine impoundment: Rımov reservoir,

Czech Republic. Folia Zoologica 57: 324–336.

Vasek, M., J. Kubecka, M. Cech, V. Drastık, J. Matena, T.

Mrkvicka, J. Peterka & M. Prchalova, 2009. Diel variation

in gillnet catches and vertical distribution of pelagic fishes

in a stratified European reservoir. Fisheries Research 96:

64–69.

Videler, J. J., 1993. Fish Swimming. Chapman & Hall,

London.

Webb, P. W., 1984a. Bodyform, locomotion and foraging in

aquatic vertebrates. American Zoologist 24: 107–120.

Webb, P. W., 1984b. Form and function in fish swimming.

Scientific American 251: 58–68.

Weihs, D., 1973. Mechanically efficient swimming techniques

for fish with negative buoyancy. Journal of Marine

Research 31: 194–209.

Weihs, D., 1974. Energetic advantages of burst swimming of

fish. Journal of Theoretical Biology 48: 215–229.

Wootton, R. J., 1998. The Ecology of Teleost Fishes. Kluwer

Academic Publishers, Dordrecht.

Znachor, P., T. Jurczak, J. Komarkova, J. Jezberova, J.

Mankiewicz, K. Kastovska & E. Zapomelova, 2006.

Summer changes in cyanobacterial bloom composition

and microcystin concentration in eutrophic Czech reser-

voirs. Environmental Toxicology 21: 236–243.

Hydrobiologia (2010) 654:253–265 265

123