This is an elec onic ep in o he o iginal a icle.
This ep in may di e om he o iginal in pagina ion and ypog aphic de ail.
Au ho (s):
Ti le:
Yea :
Ve sion:
Please ci e he o iginal e sion:
All ma e ial supplied ia JYX is p o ec ed by copy igh and o he in ellec ual p ope y igh s, and
duplica ion o sale o all o pa o any o he eposi o y collec ions is no pe mi ed, excep ha
ma e ial may be duplica ed by you o you esea ch use o educa ional pu poses in elec onic o
p in o m. You mus ob ain pe mission o any o he use. Elec onic o p in copies may no be
o e ed, whe he o sale o o he wise o anyone who is no an au ho ised use .
Dis ibu ion o He bi o ous Fish Is F ozen by Low Tempe a u e
Vejříko á, I ana; Vejřík, Lukáš; Sy ä an a, Ja i; Kiljunen, Mikko; Čech, Ma in; Blabolil,
Pe ; Vašek, Mojmí ; Sajdlo á, Zuzana; Chung, Son Hoang The; Šmejkal, Ma ek;
F ouzo á, Ja osla a; Pe e ka, Jiří
Vejříko á, I., Vejřík, L., Sy ä an a, J., Kiljunen, M., Čech, M., Blabolil, P., Vašek, M.,
Sajdlo á, Z., Chung, S. H. T., Šmejkal, M., F ouzo á, J., & Pe e ka, J. (2016).
Dis ibu ion o He bi o ous Fish Is F ozen by Low Tempe a u e. Scien i ic Repo s, 6,
A icle 39600. h ps://doi.o g/10.1038/s ep39600
2016
1
Scien i ic RepoR s | 6:39600 | DOI: 10.1038/s ep39600
www.na u e.com/scien i ic epo s
Dis ibu ion o He bi o ous Fish Is
F ozen by Low Tempe a u e
I ana Vejříko á1,2, Lukáš Vejřík1, Ja i Sy ä an a3,4, Mikko Kiljunen5, Ma in Čech1,
Pe Blabolil1, Mojmí Vašek1, Zuzana Sajdlo á1, Son Hoang The Chung1, Ma ek Šmejkal1,
Ja osla a F ouzo á1 & Jiří Pe e ka1
The numbe o he bi o es in popula ions o ec o he mic e eb a es dec eases wi h inc easing
la i ude. A highe la i udes, ish consuming plan ma e a e exclusi ely omni o ous. We assess
whe he omni o ous ish eadily shi o he bi o y o whe he animal p ey is ypically p e e ed.
We add ess empe a u e as he key ac o causing hei absence a highe la i udes and discuss he
po en ial polewa d dispe sion caused by clima e changes. A con olled expe imen illus a es ha
udd (Sca dinius e y h oph halmus) eadily u ilize plan ma e a wa e empe a u es abo e 20 °C
and a oid i s consump ion below 20 °C. Field da a suppo hese esul s, showing ha plan ma e
domina es udd die s du ing he summe and is absen du ing he sp ing. U ilizing cellulose equi es he
enzyme cellulase, which is p oduced by mic oo ganisms g owing a empe a u es o 15–42 °C. Wa e
empe a u es a highe la i udes do no each 15 °C yea - ound; a ou la i ude o 50°N~150 days/yea .
Hence, he species ichness o omni o ous ish dec eases d ama ically abo e 55° la i ude. Ou esul s
p o ide suppo o he hypo hesis ha s ic he bi o ous specialis s ha e de eloped only in he opics.
Tempe a u es below 15 °C, e en o a sho ime pe iod, inac i a e cellulase and cause die limi a ions
o omni o ous ish. Howe e , we may expec inc eases in he bi o y a highe la i udes caused by
clima e change.
A specializa ion o he bi o y in aqua ic ecosys ems was p e iously conside ed a spo adic phenomenon wi h i el-
e an impac s on he wa e communi ies1,2. Howe e , ecen s udies indica e ha he bi o y in aqua ic ecosys ems
has been o e looked by scien is s in compa ison o e es ial ecosys ems3–6, and he impac o he bi o y by bo h
e eb a es and in e eb a es on aqua ic ecosys ems is subs an ial7–9. Fishes ha e he ollowing wo me hods o
diges ing ba ely deg adable high molecula weigh polysaccha ide ma e ial: i) p edominan ly chemical p ocess-
ing by acid hyd olysis in he s omach (e.g., Acan hu idae, Kyphosidae, Pomacen idae) o ii) mechanical p ocess-
ing by b eakdown o plan ma e using a muscula s omach (e.g., Acan hu idae, Spa idae) o pha yngeal ee h
(e.g., Cichlidae, Cyp inidae, Sca idae)10–12.
A specializa ion in he bi o y is o en de eloped by ish and o he ec o he mic aqua ic and e es ial e e-
b a es a lowe la i udes13,14. In con as , ec o he mic e eb a es (mainly ish) a highe la i udes ha use plan
ma e a e omni o es, consuming bo h plan and animal sou ces in hei die s. A hese la i udes, ei he a he bi -
o ous specializa ion has no de eloped o e idence o his ype o specializa ion is lacking in he ossil eco d14–16.
The quan i y o plan ma e in he die o omni o es and he numbe o ec o he mic species capable o consum-
ing plan ma e apidly dec ease polewa d16–19, which can po en ially be explained by a numbe o heo ies as
ollows: (i) sho - e m e olu ion and he inabili y o ish o mig a e along la i udinal g adien s20, (ii) a ailabili y o
a eadily pala able plan die a lowe la i udes21, (iii) absence o plan die a highe la i udes du ing he win e 22,
and (i ) cons ain s in he diges ion o plan ma e a low empe a u es14,18,23. The las heo y is suppo ed by
obse a ions ha ma ine he bi o ous ish ha e sp ead o highe la i udes due o global wa ming24,25.
O he Eu opean na i e ish species ound in len ic wa e , he bi o y is mos de eloped in udd (Sca dinius
e y h oph halmus). The p opo ion o plan -based die consumed (mainly mac ophy es) is ma kedly highe o
udd han o o he species, e.g., oach (Ru ilus u ilus), e en when a nu ien - ich die is a ailable26. Recen
1Biology Cen e o he Czech Academy o Sciences, Ins i u e o Hyd obiology, Na Sádkách 7, 37005 České Budějo ice,
Czech Republic. 2Facul y o Science, Uni e si y o Sou h Bohemia in České Budějo ice, B anišo ská 31, 37005
České Budějo ice, Czech Republic. 3Depa men o Bioscience, Aa hus Uni e si y, Vejlsø ej 25, 8600 Silkebo g,
Denma k. 4Depa men o En i onmen al and Biological Sciences, Uni e si y o Eas e n Finland, P.O. Box 111, FI-
80101 Joensuu, Finland. 5Depa men o Biological and En i onmen al Science, Uni e si y o Jy äskylä, P.O. Box 35,
FI-40014 Uni e si y o Jy äskylä, Finland. Co espondence and eques s o ma e ials should be add essed o I.V.
(email: [email p o ec ed])
Recei ed: 18 May 2016
accep ed: 24 No embe 2016
Published: 22 Decembe 2016
OPEN
www.na u e.com/scien i ic epo s/
2
Scien i ic RepoR s | 6:39600 | DOI: 10.1038/s ep39600
s udies examining mac ophy e cha ac e is ics and he causes o selec i e he bi o y in udd asse ha he g ea es
ac o s a e C:N a io and phenolic concen a ion, bo h o which a e nega i ely co ela ed wi h he bi o y a es.
Lesse impo ance is assigned o d y ma e con en and he concen a ion o o al soluble p o eins15,27,28. The
sha ed e olu ion o udd and mac ophy es also seems o be a ele an ac o 28.
Al hough udd a e shown o p e e an animal die in expe imen al condi ions (aqua ium, mesocosm)7,15, na -
u al obse a ions indica e a p e e ence o a plan -domina ed die in he wild29,30. The pe cen age o plan ma e
in udd die usually inc eases wi h size and age31. Howe e , he key en i onmen al ac o d i ing hese p e e -
ences is s ill unknown. Acco ding o p e ious indings, empe a u e seems o play an impo an ole, as i is pos-
i i ely co ela ed wi h he in ake o plan ma e by omni o ous ish14,26,30,32. This end has also been obse ed in
o he ish species33. I is o en a consequence o changes in o he en i onmen al ac o s, such as plan a ailabili y22
and he inc eased celluloly ic ac i i y media ed by symbio ic mic oo ganisms10,30.
This s udy ocuses on he ood p e e ences o omni o ous ish using udd as a model o ganism in wo olig-
o ophic lakes exhibi ing di e ences in bo h mac ophy e occu ence and animal ood a ailabili y. Ou p incipal
conclusions we e d awn om mesocosm expe imen s es ing udd p e e ences o animal p ey e sus plan ma -
e in he p esence o di e en concen a ions o ood and unde a ange o empe a u e condi ions (16, 20, 24 °C).
The p ima y aim o he s udy was o de e mine i omni o ous ish eadily shi o he bi o y, o whe he animal
p ey is always p e e ed. These esul s will enhance ou unde s anding o he key ac o s causing he absence o
ec o he mic he bi o es a highe la i udes. Addi ionally, ecen clima e change models p edic apid wa ming,
pa icula ly a highe la i udes. The esul s o his s udy can p o ide impo an insigh s in o how many he bi -
o ous ish species may dispe se o new a eas as wa e empe a u es inc ease and ish species gain ad an ages in
ophic compe i ion by u ilizing esou ces ha a e no a ailable o o he aqua ic animals, including na i e species.
Resul s
An assessmen o po en ial die o udd in s udied lakes. In bo h lakes, mac ophy es and mac oalgae
(Cha a and Vauche ia, hence o h e e ed o as mac ophy e co e age o simplici y) occu ed down o a dep h
o 12 m, bu he abundance in he wo loca ions di e ed no iceably. In Milada Lake in Sep embe , dense co e age
by subme ged mac ophy es eached 91% a he p ima y udd habi a dep h o 0–3 m, wi h a mean we mass o
2,720 g m−2 (Table1). In May, he co e age was lowe (60.1%), bu mac ophy es we e s ill p esen wi h a mean
mass o 1,275 g m−2. The species occu ing a his dep h we e as ollows: Po amoge on pec ina us (Sep embe :
40%; May: 30%), My iophyllum spica um (20%; 12%), Cha a sp. (19%; 11%), Vauche ia (11.5%; 7%), Po amoge on
ichoides (0.2%; 0%), Spha ganium eme sum (0.15%; 0%), My iophyllum e icilla um (0.05%; 0%), Po amoge on
c ispus (0.05%; 0.1%) and Elodea canadensis (0.04%; 0%). In Mos Lake, he e was only a spa se mac ophy e
co e age o 1.6% and 0.1% a a dep h o 0–3 m, and a mean mass o only 51 g m−2 and 11 g m−2 in Sep embe
and May, espec i ely. The species occu ing a his dep h we e My iophylum spica um (Sep embe : 0.8%; May:
0.03%), Po amoge on pec ina us (0.3%; 0.04%), Cha a sp. (0.2%; 0.03%), Spha ganium eme sum (0.2%; 0%)
and Po amoge on c ispus (0.1%; 0%). The sho elines o bo h lakes comp ise s ones co e ed by pe iphy on laye s o
simila densi ies, bu an accu a e biomass measu emen was no aken.
In e eb a es li ing in he ine sedimen , on he su ace o he s ones and mac ophy es had a mean biomass
5.4 g m−2 and 4.2 g m−2 in Milada and Mos Lakes, espec i ely. In bo h lakes, he ollowing gene a occu ed
in o de o descending biomass: wa e louse (Asellus aqua icus), d agon ly la ae (Odona a), chi onomid la -
ae (Chi onomus spp.) and caddis ly la ae (T ichop e a). Addi ionally, zeb a mussels (D eissena polymo pha)
occu ed a high densi ies in bo h s udied lakes and we e he only po en ial udd ood sou ce ha was mo e
abundan in Mos Lake han in Milada Lake (Table1). Howe e , his ood sou ce was no ound in udd s omachs
in ei he o he s udied lakes (Table2).
The mean densi y o zooplank on a 0–20 m dep h was 36.3 ind. L−1 and 27.4 ind. L−1 in Milada and Mos
Lakes, espec i ely. Copepods (Copepoda) and la ge Daphnia we e mo e abundan in Milada Lake, while small
Daphnia we e mo e abundan in Mos Lake (Table1).
Die esul s e ealed by Gu Con en Analysis (GCA) and S able Iso ope Analysis (SIA). GCA
e ealed ha plan ma e domina ed he die o udd in Sep embe 2013 and 2014 (su ace wa e empe a u e:
19.1–21 °C) in bo h lakes. Fo udd olde han one yea in Milada Lake, 92.5% o he ood consumed was plan
ma e in he o m o mac ophy es, and he es was animal p ey (Table2). In Mos Lake, plan ma e also
Dep h 0–3 m Dep h 0–20 m
Mac ophy e co e .(%) Dep h 0–3 m Mean biomass o po en ial sou ces (g m−2)Dep h 0–20 m Mean densi y (ind. L−1)
Pala able
mac oph.
Unpala able
mac oph.
Pala able
mac oph.
Unpala able
mac oph.
Ben hic
in e . D eissena Copepods
Cladoce ans
(Daphnia)
small la ge
Milada 71 20 2122 598 5.4 167 23.5 10.1 2.7
Mos 0.9 0.8 27 24 4.2 242 0.3 26.8 0.3
Table 1. Summa y o he assessmen o po en ial die o udd in Milada and Mos Lakes. We assessed
mean co e age and mean biomass o pala able and unpala able mac ophy es on he bo om in he dep h
o 0–3 m, and biomass o ben hic in e eb a es in he same dep h. Fu he , we assessed mean densi y o
zooplank on in he wa e column in he dep h o 0–20 m.
www.na u e.com/scien i ic epo s/
3
Scien i ic RepoR s | 6:39600 | DOI: 10.1038/s ep39600
domina ed bu consis ed o pe iphy on (68%) and de i us (25%). Animal p ey accoun ed o 7% o he gu con-
en s. In Sep embe in Milada Lake, ju enile udd we e s ic he bi o es consuming mac ophy es, whe eas in
Mos Lake hey we e s ic zooplank i o es. In con as o he wa me mon h o Sep embe , in May 2015 (su ace
wa e empe a u e: 13.1–14.2 °C), only animal p ey was ound in udd diges i e ac s. In Milada Lake, die con-
sis ed solely o he ae ial s age o aqua ic insec s. In Mos Lake, i consis ed o he ae ial s age o aqua ic insec s,
ben hic in e eb a es and zooplank on (Table2).
The esul s o SIA indica ed a lowe p opo ion o plan ma e in udd die in Sep embe han was sug-
ges ed by GCA (Table2 and Fig.1). The SIAR s able iso ope mixing model showed 51% and 62% plan ma -
e consumed by udd in Milada and Mos Lakes, espec i ely. In con as , SIA showed a highe pe cen age o
animal p ey consump ion (Milada: 49%, Mos : 38%) han he GCA (7.5%, 7%) (Table2 and Fig.1). No clea
end was obse ed be ween ish size and δ
15N o ish size and he p opo ion o plan ma e in gu con en s
(Supplemen a yFigsS1andS2). Addi ionally, he e we e no s a is ically signi ican di e ences in δ
13C and δ
15N
be ween sampling yea s 2013 and 2014 (see Supplemen a yTableS1, and Fig.S3 o SIA biplo s).
Resul s o eeding expe imen s. Unde expe imen al condi ions e lec ing he na u al condi ions in
Milada Lake (an animal p ey o plan ma e a io o 1:400), plan ma e domina ed udd die s a 24 °C (99.8%)
and 20 °C (88.8%). In con as , plan ma e cons i u ed an a e age o only 1.6% o gu con en s a 16 °C (Fig.2 and
Table3). The pe cen age o plan ma e a 24 °C and 20 °C dec eased wi h changing die a ios. In he p esence
o a a io o animal p ey o plan ma e o 1:10, plan ma e comp ised 52.5% and 44.5% o he gu con en s,
espec i ely. A 16 °C and a a io o 1:10, he p opo ion o plan ma e emained e y low (1.4%). The inal
expe imen used a a io o animal p ey o plan ma e o 1:1, which di e ed he mos om na u al condi ions o
all he ea men s. Unde hese condi ions, he pe cen ages o plan ma e a 24 °C and 20 °C dec eased o 31.9%
and 15.1%, espec i ely. A 16 °C and a 1:1 a io, he p opo ion o plan ma e was only 1.1% (Table3 and Fig.2).
Bo h empe a u e and he a io o a ailable ood sou ces signi ican ly a ec ed he die p e e ences o udd. The
impac o empe a u e was g ea e (F2,15 = 10.35, p < 0.01) han he impac o die a io (F2,15 = 2.67, p < 0.10). The
Da e Lake N
P opo ion o gi en die ca ego y in gu con en , mean ± SD (%)
Mac oph. Pe iphy on De i us Ben hos Zoopl. D eissena A. insec
Sep embe Milada 80 92.5 ± 22 0 0 7.5 ± 22 0 0 0
2013 and 14 Mos 80 0 68 ± 34 25 ± 31 7 ± 13 0 0 0
May 2015 Milada 25 0 0 0 0 0 0 100
Mos 25 0 0 0 42 ± 41 13 ± 20 0 45 ± 40
Table 2. Mean p opo ion o die ca ego ies (% ± s anda d de ia ion) in Milada and Mos Lakes du ing
Sep embe and May acco ding o gu con en analysis. The size ange o udd was 98–430 mm (TL; > 1 yea
old). N = numbe o dissec ed indi iduals, Zoopl. = zooplank on and A. insec = ae ial s age o aqua ic insec .
Figu e 1. P obabili y p opo ion o plan ma e and animal p ey in assimila ed die o udd in Milada
and Mos Lakes acco ding o s able iso ope analysis. Plan ma e is p esen ed by he ca ego ies mac ophy es
(Milada Lake), pe iphy on and de i us (bo h Mos Lake). Animal p ey is p esen ed by he ca ego ies
zooplank on, ben hos, and zeb a mussel (D eissena polymo pha). The c edibili y in e als a e 95, 75 and 25%.
(see Supplemen a yFig.S3 o SIA Biplo s).
www.na u e.com/scien i ic epo s/
4
Scien i ic RepoR s | 6:39600 | DOI: 10.1038/s ep39600
p opo ion o plan ma e in udd die dec eased wi h dec easing empe a u e. The in e ac ion o empe a u e
and die a io also had a s a is ically signi ican impac (F8,8 = 31.11, p < 0.001).
A 16 °C, all s udied udd exhibi ed e y simila eeding beha iou ega dless o die a io, and he pe cen age
o plan ma e consumed was consis en ly e y low (an amoun g ea e han 4% was ound in only one indi-
idual). A 20 °C, ood consump ion was mo e a iable, and we ound indi iduals wi h diges i e ac con en s
anging om 100% plan ma e o 100% animal p ey. A 24 °C, he a iabili y was lowe han a 20 °C, wi h he
highes a iabili y ound when using a 1:10 die a io (Table3 and Fig.2).
In he addi ional expe imen wi h only animal p ey a ailable as a ood sou ce a 24 °C, all eigh indi iduals
had diges i e ac s ull o animal p ey (100%). In con as , in he ial wi h only plan ma e a ailable as a ood
sou ce and a empe a u e o 16 °C, he diges i e ac s o all en indi iduals we e en i ely emp y (e en when using
he spa e indi iduals, see Me hods). In he inal addi ional expe imen wi h only plan ma e as a ood sou ce a
13 °C o 168 h (i.e., assuming majo s a a ion), all eigh indi iduals u ilized plan ma e (100%; Table3) and he
specimens’ en i e diges i e ac s we e e enly illed wi h he plan ma e ial.
A sligh di e ence in mass was obse ed be o e and a e expe imen s o all udd indi iduals. In mos cases
(excep o he wo ials discussed below), he sligh mass inc eases we e caused by he illing o he diges i e
ac . The e was i ually no di e ence obse ed a e sub ac ing he mass o gu con en s. In he expe imen s
wi h bo h animal p ey and plan ma e , he mean di e ence ± SD was 0.35 g ± 0.24 and was s a is ically insig-
ni ican o all expe imen s (p > 0.1). In he case o expe imen s wi h only plan ma e a ailable a 16 °C, a sligh
dec ease in mass was obse ed (mean ± SD: 0.06 g ± 0.05) bu i was also s a is ically insigni ican (F1,10 = 1.13,
p > 0.1). A g ea e dec ease in mass was obse ed in he expe imen wi h only plan ma e a 13 °C. In his ial,
Figu e 2. P opo ion o plan ma e in gu con en o expe imen al udd in gi en empe a u es and
h ee di e en die a ios. The die a ios o animal p ey s. plan ma e we e 1:1 (g ey), 1:10 (blue) and 1:400
(g een). Box and whiske s plo s: uppe and lowe qua iles (boxes), median alues (line inside he boxes),
maximum and minimum alues (whiske s), and ou lie s (ci cles) a e shown.
T, Time Repe i ion no.
P opo ion o plan ma e in gi en die a io (%)
1:400 1:10 1:1 0:1 1:0
13 °C, 168 h 1 — — — 100 ± 0 —
13 °C, 168 h 2 — — — 100 ± 0 —
16 °C, 24 h 1 2 ± 4 1.3 ± 1.5 1.3 ± 1.5 Emp y —
16 °C, 24 h 2 1.1 ± 1.3 1.5 ± 1.9 1 ± 1.2 Emp y —
20 °C, 24 h 1 87.5 ± 2 42.5 ± 41.9 14 ± 6.9 — —
20 °C, 24 h 2 90 ± 14.1 46.5 ± 36.5 16.3 ± 22.8 — —
24 °C, 24 h 1 100 ± 0 36.3 ± 21.4 23.8 ± 7.5 — 0 ± 0
24 °C, 24 h 2 99.5 ± 1 68.8 ± 27.8 40 ± 40.8 — 0 ± 0
Table 3. Mean p opo ion o plan ma e in gu con en o expe imen al udd (% ± s anda d de ia ion)
o each expe imen in gi en die a ios (animal p ey s. plan ma e ) and empe a u es. “Emp y” deno es
si ua ions when ish we e ound wi h no gu con en s, whe eas 0% o plan ma e indica es p esence o animal p ey.

www.na u e.com/scien i ic epo s/
5
Scien i ic RepoR s | 6:39600 | DOI: 10.1038/s ep39600
he mean dec ease was 0.75 g ± 0.32, which co esponds o a mean mass dec ease o 2.7% o o al mass and was
s a is ically signi ican (F1,8 = 5.4, p = 0.02).
Discussion
Based on he gu con en s o udd in bo h lakes in Sep embe (su ace wa e empe a u e: 19.1–21 °C), udd show
a endency owa ds he bi o y du ing summe when a su icien amoun o mac ophy e plan ma e ial is a ailable.
The esul s o Milada Lake demons a e ha his is also ue o ju enile udd, which con adic s he wo k by
Nu minen e al.34, who obse ed ha udd did no u ilize a plan die du ing hei i s yea . Howe e , i he only
plan ma e a ailable is pe iphy on and/o de i us ju enile udd igno e i and u ilize zooplank on, as shown by
he esul s o Mos Lake. In con as , olde udd consume pe iphy on and de i us ex ensi ely in he absence o
mac ophy es. When a su icien numbe o mac ophy es a e p esen , pe iphy on and de i us a e igno ed; his was
demons a ed by he esul s o Milada Lake.
Gu con en analyses conduc ed in May (su ace wa e empe a u e: 13.1–14.2 °C) show ha he udd com-
ple ely dis ega ded plan ma e e en hough mac ophy es and pe iphy on we e p esen . The obse ed endency
owa ds swi ching be ween ood sou ces is consis en wi h o he s udies on udd. Based on ou esul s, empe -
a u es app oxima ely 20 °C seem o be a c ucial h eshold o ansi ioning o he bi o y26,30,31. This end is no
only alid in ou chosen model o ganism bu can also be obse ed in o he species. Fo example, a s udy ocused
on a udde ish (Gi ella nig icans) showed ha hey exhibi ed he highes RNA:DNA a io ( he mos in ensi e
g ow h in a sho - e m pe iod) when hey u ilized plan ma e a empe a u es abo e 22 °C. Whe eas a 17 °C, a
low RNA:DNA a io indica ed ha ish we e expe iencing s ess while u ilizing plan ma e 14. Acco ding o he
SIA conduc ed o udd in Sep embe , he pe cen age o plan ma e consump ion was lowe han ha shown
by GCA. This di e ence is likely caused by iso ope u no e in ish issue35. The issue o udd s ill con ained an
iso opic signal om ood assimila ed du ing colde pe iods, when animal p ey was p e e ed. Ne e heless, we
canno exclude he possibili y ha he iso opic signal was en iched by he assimila ion o bac e ia esponsible o
cellulose diges ion, a he han by he plan s hemsel es36.
The expe imen al po ion o ou s udy p o ided suppo o he hypo heses ha bo h empe a u e and a ail-
abili y o ood play impo an oles in he ood p e e ences o udd. Rudd we e almos en i ely he bi o ous in
condi ions simila o hose o Milada Lake, ep esen ed by ials wi h he highes a io o plan ma e o animal
p ey and empe a u es o 20 °C and 24 °C. Dec easing plan ma e and inc easing animal p ey led o a dec ease
in obse ed he bi o y. This end was mo e appa en a 20 °C han a 24 °C. Thus, plan ma e seems o be mo e
eadily u ilized in wa m wa e when i is eadily a ailable.
The c i ical poin o ansi ion o plan u iliza ion was obse ed a 16 °C. Rudd essen ially consumed only
animal p ey a his empe a u e in all h ee expe imen al concen a ions o animal p ey and plan ma e .
In e es ingly, in he expe imen a 16 °C which p o ided only plan ma e , he udd p e e ed o consume no ood
a he han o consume only plan ma e . I should be emphasized ha his occu ed a e a pe iod o s a a ion
las ing ou days. Such a s a egy has been obse ed among ca ni o ous ish bu is no common o omni o ous
o he bi o ous species22.
These indings indica e ha low empe a u e is he key ac o d i ing he elimina ion o plan ma e om he
die o udd. This e ec does no only apply o ish bu is also obse ed in in e eb a e omni o es such as cope-
pods37 and snails (Lymnaea s agnalis)38. P e e ence o plan ma e dec eases apidly when he empe a u e alls
below 20 °C. Plan ma e is comple ely igno ed when he empe a u e alls o 16 °C, which likely explains why
Do enbosch and Bakke 9 no ed low p e e ence o plan ma e among udd and g ass ca p (C enopha yngodon
idella). The s ong p e e ence o animal p ey obse ed by Do enbosch and Bakke 15 was likely in luenced by
he low empe a u e in hei expe imen (18 °C), and by he a io o a ailable animal p ey o plan ma e (2.9:1).
These esul s a e consis en wi h ou indings ha udd in bo h lakes u ilized only animal p ey in he sp ing,
when he su ace wa e empe a u e anged om 13.1 o 14.2 °C. Pisci o y was e en obse ed in udd du ing he
sp ing, in low empe a u es in he Niaga a Ri e 26.
In ex eme cases, ish a e likely capable o gaining a minimal amoun o ene gy om plan sou ces e en in
cold wa e , as shown in when udd s a ed o u ilize plan ma e a 13 °C a e 168 h o s a a ion in one o ou
expe imen s. This sugges s ha ish a e able o inges plan ma e a low empe a u es, bu his beha iou is no
sus ainable in he long- e m. The mean dec ease in mass o udd indi iduals in ou expe imen was 2.7%, sup-
po ing he heo y o Beh ens and La e y14 ha ish begin o expe ience s ess a empe a u es below 17 °C when
consuming plan ma e . Howe e , excep ions we e disco e ed in An a c ica, whe e ex eme he bi o y is ound in
ou species o no o henioid ishes (No o henioidei), which u ilize algae19, and e e ences he ein.
To be e unde s and he obse ed ends ela ed o empe a u e, i is necessa y o unde s and how ish diges
plan ma e . Fish lack speci ic o gans o plan ma e diges ion, such as he specialized s omachs ound in he -
bi o ous mammals and some bi ds10,39. The e o e, o use cellulose as a nu ien sou ce hey equi e he enzyme
cellulase. B-1,4 glycosidic bonds mus be clea ed o elease glucose uni s40. Fo his p ocess, ish depend on he
p oduc ion o enzymes by symbio ic mic oo ganisms li ing in hei diges i e ac s10,11. The mic oo ganisms
en e he diges i e ac ia consumed ood, p ima ily h ough de i us10,31,41. The cellulase le els in he diges i e
ac s o udd a e ypically highes in la e summe and lowes in ea ly sp ing10. This is known om he wo k by
Saha e al.42, who isola ed a se ies o Bacillus ci culans and B. mega e ium om g ass ca p and ilapia (O eoch omis
mossambica). A cha ac e iza ion o he isola ed mic oo ganisms e ealed ha hey could g ow in a wide ange
o pH le els (5–11) and empe a u es (15–42 °C). Simila ly, a wide empe a u e ange was ound o be sui able
o mic oo ganisms wi h cellulase isola ed om he cyp inid Labeo ohi a43. The e o e, he cellulase p oduced
by mic oo ganisms capable o li ing in he diges i e ac s o ec o he ms can unc ion a lowe empe a u es in
compa ison o o he wa e o ganisms, whe e op imum empe a u es o clea ing complex ca bohyd a es ange
om 30–60 °C44–46. Al hough he lowe limi o cellulose unc ioning in mic oo ganisms is 15 °C, he bi o y by
ish s ill seems o be unbene icial om 16–19 °C, (based on ou obse a ion26,30,32), p obably due o low cellulase
www.na u e.com/scien i ic epo s/
6
Scien i ic RepoR s | 6:39600 | DOI: 10.1038/s ep39600
ac i i y. The e o e, he p ecise empe a u e a which plan die s a s o be igno ed by ish will a y by species and
indi idual.
Howe e , a empe a u e o app oxima ely 15 °C is s ill ela i ely high o an aqua ic ecosys em. A ou s udy
si e (Milada Lake: 50.4°N), he su ace wa e empe a u e is wa me han 15 °C o only 149 days (41%) each yea
on a e age (Pali o ý kombiná Ús í s.p., unpubl. da a). The pe iod o e ec i e cellulose diges ion is d ama ically
sho ened wi h inc easing la i ude. This is likely he eason ha species ichness in omni o ous ish dec eases
d ama ically abo e 55°, and why hey a e absen abo e 60°16. This end is s onge in eshwa e han in ma ine
ecosys ems, p obably due o ocean cu en s ha in luence wa e empe a u e16. Wa m cu en s can inc ease he
empe a u e a highe la i udes, and he opposi e can occu a lowe la i udes. In con as o ma ine condi ions,
in eshwa e he empe a u e is mo e s ic ly co ela ed wi h la i ude. The end o dec eases in he numbe o
species u ilizing plan ma e a highe la i udes is also s onge o ish han o e es ial ec o he ms, such as
ep iles16,47. This is likely due o di e ences in hei habi a s. In cold clima es, ep iles ha e sol ed he p oblem o
plan diges ion by educing hei size. A small body can be mo e quickly wa med by ai , especially wi h a wa med
unde laye ha enables cellulase ac i i y. This s a egy is no iable in wa e ecosys ems, which ha e a high spe-
ci ic he mal capaci y.
Al hough hese ends in ish he bi o y a e sligh ly less appa en in ma ine han in eshwa e ecosys ems,
he polewa d mo emen o he bi o ous species in ma ine ecosys ems due o clima e change has al eady had an
appa en nega i e impac . Fo example, losses o kelp o es s caused by inc eases in he bi o ous consume s ha e
been obse ed24,25,48,49. The sp ead o he bi o ous species in eshwa e ecosys ems has likely no been ex ensi ely
s udied o da e. Ne e heless, we can assume ha hese changes a e occu ing. The impac o non-na i e ish
species on mac ophy es is e iden as obse ed in s udies ega ding he in oduc ion o udd in New Zealand and
No h Ame ica27,28. We can also p esume ha ish popula o use in aquacul u e, such as Tilapia spp. o g ass
ca p9,50, a e likely o sp ead by human ac i i y and will ha e nega i e impac s on mac ophy es.
In conclusion, empe a u e is he cause o ish de eloping s ic he bi o ous specializa ions only in opical
egions, excep o a ew ex eme cases as men ioned abo e. A highe la i udes, omni o ous ish only u ilize plan
ma e 13,14, which is es ic ed by empe a u es lowe han 15 °C. He bi o ous ish need a egula in ake o ood;
he e o e an annual dec ease in wa e empe a u e below 15 °C, e en o a sho ime pe iod, p e en s ish speci-
a ion and he bi o ous specializa ion a highe la i udes. Clima e change is expec ed o ha e s onge impac s on
wa e empe a u es a highe la i udes, allowing he dispe sion o he bi o ous eeding beha iou s by ish. This can
bene i many omni o ous ish species and g ea ly shape he u u e o ish communi ies in lakes o highe la i ude
egions.
Me hods
S udy si e. The s udy was conduc ed in wo newly c ea ed opencas mine lakes, Milada Lake (50°39′N,
13°58′E) and Mos Lake (50°32′N, 13°32′E), in he Czech Republic (Fig.3). The oligo ophic o meso ophic
Milada Lake has a mean summe o al phospho us (TP) in he su ace laye o < 10 μ g L−1, and is loca ed 80 km
no hwes o P ague (Fig.3b). I has an a ea o 250 ha, a olume o 0.036 km3 and a maximum dep h o 25 m.
Aqua ic es o a ion s a ed in 2001 and was inished in 2011. Se e al species o mac ophy es and algae a e p esen
a high le els o biomass o a dep h o 12 m51. The biomass o udd was 3.58 kg ha−1 in 2013, and 2.50 kg ha−1 in
2014. O he ishes p esen a e pe ch (Pe ca lu ia ilis), oach, u e (Gymnocephalus ce nua), pike (Esox lucius),
Eu opean ca ish (Silu us glanis), ench (Tinca inca) and pikepe ch (Sande luciope ca). The oligo ophic Mos
Lake (TP < 5 μ g L−1) is 75 km no hwes o P ague. I has an a ea o 310 ha, a olume o 0.07 km3 and a maximum
dep h o 75 m. Aqua ic es o a ion s a ed in 2008 and was inished in 2014. Mac ophy es we e no abundan in
2014. The biomass o udd was 1.9 kg ha−1 in 2013 and 0.6 kg ha−1 in 2014. O he ishes p esen a e pe ch, oach,
u e, pike, Eu opean ca ish, ench and ma aena whi e ish (Co egonus ma aena).
Wa e empe a u e and dissol ed oxygen we e measu ed using a calib a ed YSI 556 MPS p obe (YSI
Inco po a ed - Yellow Sp ings, Ohio, USA) in bo h lakes, h ee imes each du ing sampling in Sep embe 2014
and May 2015.
Collec ion o plan and in e eb a e samples. To ob ain a quali a i e assessmen o mac ophy es, wo
SCUBA di e s isually assessed hei occu ence a en ansec s. T ansec s we e ma ked om he sho e o a
dep h o 12 m in bo h lakes in Sep embe 2013, Sep embe 2014 and May 2015 (Fig.3b), using measu ing apes.
The co e age o each mac ophy e species, he unco e ed bo om a ea, he pe cen age composi ion o each spe-
cies, and he pe cen age o unco e ed bo om a ea we e measu ed a 1 m dep h in e als. The esul s o duplica e
measu emen s we e a e aged o a mo e accu a e assessmen . To ob ain a quan i a i e assessmen o mac o-
phy es, SCUBA di e s s aked a quad angle (50 × 50 cm) a six loca ions a a dep h o 0–3 m (Fig.3b) wi h 100%
mac ophy e co e age. All mac ophy es and mac oalgae we e emo ed om he a ea, pu in o a sack, aised o he
boa and weighed a e being allowed o d ain o 10 minu es. Po amoge on spp., Cha a sp., Vauche ia and Elodea
canadensis we e de e mined o be po en ially pala able o udd, acco ding o pas s udies9,15,30 and ou obse a-
ions. My iophyllum sp. was classi ied as an unpala able species and was no included as a po en ial ood sou ce
o udd. The mean co e age o pala able species pe 1 m2 a a dep h o 0–3 m (whe e mos udd we e ound by
SCUBA di e s and in gillne ca ches) was es ima ed using bo h quan i a i e and quali a i e in o ma ion. Samples
o pala able mac ophy es we e collec ed and ozen o s able iso ope analysis. To ob ain in o ma ion abou pe i-
phy on, he s ony sho eline was assessed isually om he boa using pola ized glasses.
Sampling was conduc ed o ben hic in e eb a es (animal p ey), a a dep h o 0–3 m in bo h lakes in
Sep embe 2014. SCUBA di e s sampled six loca ions e enly dis ibu ed along he lake; hese we e iden ical
o he loca ions o quan i a i e assessmen o mac ophy es (Fig.3b). A each loca ion a plas ic co e (leng h
50 cm, diame e 8 cm) was used o collec wo samples o in e eb a es li ing in he ine sedimen in h ee dep h
zones (0.5, 1.5 and 3 m). All six subsamples we e hen combined o gi e one composi e sample pe locali y. Fo
www.na u e.com/scien i ic epo s/
7
Scien i ic RepoR s | 6:39600 | DOI: 10.1038/s ep39600
in e eb a es li ing on he s ones, a quad angle (50× 50 cm) was s aked ou , and all mac oscopic in e eb a es
(mos o which we e wa e louse o zeb a mussel specimens) we e collec ed using o ceps. Nex , all uppe s ones
we e aised om he s aked a ea o he boa o collec any emaining indi iduals. Two subsamples we e ob ained
and combined o gi e one sample pe loca ion. The bo om co e o ine sedimen and s ones was measu ed a
each loca ion o de e mine he pe cen co e o each ype o bo om ma e ial. In Milada Lake, in e eb a es li ing
on mac ophy es we e also collec ed om he 50× 50 cm s aked a ea. All mac ophy es we e collec ed, pu in o
a sack, aised o he boa and insed, and all in e eb a es we e collec ed using o ceps. Two subsamples we e
ob ained and combined o gi e one sample pe loca ion. The p e iously men ioned wo (ben hic in e eb a es
om ine sedimen and s ones; Mos Lake) o h ee (ben hic in e eb a es om ine sedimen , s ones and mac-
ophy es; Milada Lake) samples we e combined om each loca ion. All in e eb a es we e coun ed, weighed and
p ese ed in 4% o maldehyde o subsequen iden i ica ion o he species le el in he labo a o y. To es ima e
mean biomass o in e eb a es pe 1 m2 o he lake bo om, he mean biomass o in e eb a es li ing in ine
sedimen and on s ones and mac ophy es was ecalcula ed p opo ionally. Subsamples o ben hic in e eb a es
occu ing in he en i onmen we e collec ed and ozen o s able iso ope analysis.
In bo h lakes, zooplank on was sampled in iplica e wi h a plank on ne (diame e : 24 cm, mesh size: 100 μ m)
a h ee loca ions by a e ical haul a 0–20 m in Sep embe 2014 (Fig.3). The samples we e ixed wi h 4% o -
maldehyde. In he labo a o y, he o ganisms we e iden i ied o o de o genus and coun ed unde he mic oscope
(Olympus CX40, 100 magni ica ions) using a Sedgewick-Ra e chambe . Daphnia, he axa mos ulne able o
ish p eda ion, we e di ided in o he ollowing wo g oups: small (< 700 μ m ca apace leng h) and la ge (≥ 700 μ m).
The epo ed densi y o zooplank on is he mean o all nine samples. Zooplank on samples we e collec ed and
ozen o s able iso ope analysis in Ma ch, May, July, Sep embe and No embe 2014.
Fish sampling. All animal handling (including ish sampling, GCA, SIA and he mesocosm expe imen )
was pe o med acco ding o he guidelines o and wi h pe mission om he Expe imen al Animal Wel a e
Commission a he Minis y o Ag icul u e o he Czech Republic (Re . No. CZ 01679). The Expe imen al Animal
Wel a e Commission app o ed all expe imen al p o ocols.
In bo h lakes, ish we e sampled o GCA and SIA using ben hic mul i-mesh gillne s ollowing he Eu opean
s anda d52. Gillne s we e se o e nigh ; hey we e placed in he wa e 2 h be o e sunse and li ed 2 h a e sun-
ise. Sampling was conduc ed h ee imes each a dep hs o 0–3, 3–6 and 6–9 m a h ee loca ions. Al oge he , 72
gillne s we e se in bo h lakes and bo h in 2013 and 2014. All cap u ed udd we e anaes he ised using a le hal
dose o icaine me hanesul ona e (MS–222, Sigma Ald ich Co.). Rudd o he expe imen we e also sampled by
elec o ishing (600 W-pulsed DC cu en ) in Milada Lake in Sep embe 2015.
Figu e 3. A map showing (a) he loca ion o he Czech Republic wi hin Eu ope (la i ude o 50° no h as bold
line) and (b) he loca ion o he wo s udy si es, Milada and Mos Lakes, in he Czech Republic. A de ailed
iew o he ba hyme ic maps shows con ou lines wi h ele an dep hs and sampling locali ies: B = ben hic
in e eb a es sampling; G = gillne s sampling; M = mac ophy es sampling; Z = zooplank on sampling. The
igu e was gene a ed by he so wa e A cMap, e sion 10.2.259.
www.na u e.com/scien i ic epo s/
8
Scien i ic RepoR s | 6:39600 | DOI: 10.1038/s ep39600
In Sep embe 2013 and 2014, 80 udd olde han one yea and 30 ju eniles we e used o GCA. In May 2015,
25 indi iduals olde han one yea we e used; in all cases, he samples we e om bo h lakes. The analysis was con-
duc ed on he same day as sampling. The mean sizes o he adul udd used o GCA we e 281.7 mm ± 92.4 (SD)
and 191.3 mm ± 38.9 o Milada and Mos Lakes, espec i ely. The mean mass measu emen s we e 100.8 g ± .56.5
and 515.9 g ± .477.4, espec i ely. The mean sizes o he ju enile udd we e 65 mm ± 2.6 and 63.8 mm ± 2.4, and
he mean mass measu emen s we e 1.8 g ± 0.4 and 1.2 g ± 0.5 o Milada and Mos Lakes, espec i ely. Fish scales
we e collec ed and used o age he ish. A e collec ion, 76 udd om Milada Lake and 56 udd om Mos Lake
in Sep embe 2013 and 2014 we e ozen o p ese e hem o SIA. The mean sizes o he udd used o SIA we e
271.4 mm ± 98.6 and 184.8 mm ± 47.2 o Milada and Mos , espec i ely, and he mean mass measu emen s we e
476.3 g ± 467.9 and 94.6 g ± 66 o Milada and Mos , espec i ely.
GCA. Fo udd ob ained om he expe imen and he lakes, gu con en s we e iden i ied unde a dissec ing
mic oscope and he pe cen composi ion o hei die by olume was isually es ima ed. The die o expe imen al
udd was ca ego ized in o wo g oups based on wha hey we e o e ed du ing he expe imen (plan ma e o
animal p ey). The die o udd ob ained om he lakes was ca ego ized in o se en unc ional g oups (mac o-
phy es, pe iphy on, de i us, ben hos, zooplank on, D eissena and he ae ial s age o aqua ic insec s). Fo he
expe imen al indi iduals, he gu con en s (we mass) we e also weighed (Ke n CKE 3600–2, accu acy = 0.01 g)
o de e mine we mass.
SIA o ish and po en ial ood sou ces. Do sal muscle issue was dissec ed om indi idual ish om
bo h lakes and cleaned o scales and skin. All samples ( ish and die samples) we e washed wi h dis illed wa e .
The samples we e d ied a 60 °C o 48 h and g ound in o a homogenous powde using a ball-mill Re sch MM 200
(Re sch GmbH, Haan, Ge many). Small subsamples (0.520–0.770 mg) we e weighed in in cups o he analysis
o δ
13C and δ
15N. The iso ope analyses we e pe o med using a The mo Finnigan DELTAplus Ad an age con in-
uous low s able iso ope- a io mass spec ome e , connec ed o a Ca lo E ba Flash EA1112 elemen al analyse
(The mo Elec on Co po a ion, Wal ham, MA, USA) a he Uni e si y o Jy äskylä, Finland. Measu emen s o
δ
13C and δ
15N we e epo ed in pe mil (‰) no a ion, using in e nal s anda ds wi h a known ela ionship o he
in e na ional s anda ds o Vienna Pee Dee belemni e ( o ca bon) and a mosphe ic N2 ( o ni ogen). Sample
analyses also yielded he elemen al composi ion o ca bon and ni ogen (by mass). Because he C:N a ios we e
consis en ly lowe han 3.5, iso ope alues we e no lipid co ec ed53.
Feeding expe imen s. F om Sep embe 14 o 24, 2015, 12 di e en expe imen s we e conduc ed and dupli-
ca ed on he sho e o Milada Lake, using di e en die a ios and wa e empe a u es. Each expe imen las ed
24 h, excep o one ha las ed 168 h (Table4). Fib eglass anks we e placed unde a en pa ilion o p e en
con ac wi h ain and sunligh , and we e illed wi h wa e om Milada Lake il e ed h ough a sie e (mesh size
170 μ m) ( ank diame e = 1 m, su ace o ank bo om = 0.785 m2, o al heigh o he ank = 0.8 m, heigh o wa e
column = 0.64 m and o al wa e olume in he ank = 502 L). Du ing he expe imen s, wa e empe a u es we e
measu ed a 4 h in e als, and he empe a u e was main ained wi h a maximum de ia ion o ± 0.6 °C. Oxygen
olume was also measu ed a he same ime in e als ( anging om 11.7 o 6.8 mg L−1 wi h he highes and lowes
alues a 13 °C and 24 °C, espec i ely). Wa e was hea ed o cooled o he equi ed empe a u e using aqua ium
hea e s (Eheim Jäge 300 W) and eeze packs con aining e ige an gel.
Die a ios o he expe imen s we e selec ed acco ding o he occu ence o pala able plan s and animal p ey
in he lake (see abo e and Table1) es ima ed o he ank su ace a ea. The es ima ed a ailable a io (by we
mass) o animal p ey o plan ma e in Milada Lake was 1:395.2, p esen ed as 1:400 o simpli ica ion. Then, die
masses co esponding wi h each o he equi ed a ios (1:10, 1:1, 1:0, 0:1) we e calcula ed o each expe imen
(Table4). Po amoge on pec ina us om Milada Lake and Chi onomus sp. om cul i a ion we e placed in he
equi ed amoun s on he ank bo om. Bo h ypes o die ha e been shown o be eadily u ilized by udd. P io
o placemen in he ank, mac ophy es we e insed p ope ly o emo e in e eb a es, and he equi ed mass was
weighed a e being allowed o d ain o 10 minu es. Li e Chi onomus sp. we e placed in shallow plas ic bucke s
(25 × 25x 10 cm) o p e en hem om hiding.
Th ee days p io o he expe imen , he udd we e caugh om Milada Lake by elec o ishing. The udd
we e s a ed o one day in cage ne s in he lake, and u he acclima ized o wo mo e days a he selec ed
Die a io Mass a io (g) Du a ion (h)
Da e o he expe imen in 2015
T = 13 °C T = 6 °C T = 20 °C T = 24 °C
1:400* 4.2:1,666 24 — Sep . 14 Sep . 14 Sep . 14
1:10 152:1,518 24 — Sep . 15 Sep . 15 Sep . 15
1:1 835:835 24 — Sep . 16 Sep . 16 Sep . 16
1:0 1,670:0 24 — — — Sep . 17
0:1 0:1,670 24 — Sep . 17 — —
0:1 0:1,670 168 Sep . 18–24 — — —
Table 4. Summa y o conduc ed expe imen s, he die a ios used (animal p ey s. plan ma e ) and mass
a ios coun ed, empe a u es, du a ion and da es o he expe imen s. The expe imen s we e conduc ed
always in wo epe i ions. No e ha he o al usable die mass was ca. 1,670 g pe ank in all expe imen s. * eal
lake die a io, Sep . means Sep embe .