Comparison of epifluorescence microscopy and flow cytometry in counting freshwater picophytoplankton

This is a sel -a chi ed e sion o an o iginal a icle. This e sion may di e om he o iginal in pagina ion and ypog aphic de ails. Au ho (s): Ti le: Yea : Ve sion: Copy igh : Righ s: Righ s u l: Please ci e he o iginal e sion: CC BY 4.0 h ps://c ea i ecommons.o g/licenses/by/4.0/ Compa ison o epi luo escence mic oscopy and low cy ome y in coun ing eshwa e picophy oplank on © Au ho s, 2021 Published e sion Salmi, Pauliina; Mäki, Ani a; Mikkonen, Anu; Pupponen, Veli-Mikko; Vuo io, K is iina; Tii ola, Ma ja Salmi, P., Mäki, A., Mikkonen, A., Pupponen, V.-M., Vuo io, K., & Tii ola, M. (2021). Compa ison o epi luo escence mic oscopy and low cy ome y in coun ing eshwa e picophy oplank on. Bo eal En i onmen Resea ch, 26, 17-27. h p://www.bo en .ne /BER/a chi e/pd s/be 26/be 26-017-027.pd 2021 BOREAL ENVIRONMENT RESEARCH 26: 17–27 © 2021 ISSN 1797-2469 (online) Helsinki 11 Feb ua y 2021 Edi o in cha ge o his a icle: Johanna Ma ila Compa ison o epi luo escence mic oscopy and low cy ome y in coun ing eshwa e picophy oplank on Pauliina Salmi1), Ani a Mäki1), Anu Mikkonen1), Veli-Mikko Puupponen1), K is iina Vuo io2)* and Ma ja Tii ola1) 1) Nanoscience Cen e , Depa men o Biological and En i onmen al Sciences, Uni e si y o Jy äskylä, Su on ie 9C, P.O. Box 35, FI-40014 Jy äskylä, Finland 2) Finnish En i onmen Ins i u e (SYKE), P.O. Box 140, FI-00251 Helsinki, Finland (*co esponding au ho ’s e-mail: [email p o ec ed]) Recei ed 20 Ap . 2020, inal e sion ecei ed 10 Dec. 2020, accep ed 29 No . 2020 Salmi P., Mäki A., Mikkonen A., Pupponen V-M., Vuo io K. & Tiiola M. 2021: Compa ison o epi luo es- cence mic oscopy and low cy ome y in coun ing eshwa e picophy oplank on. Bo eal En . Res. 26: 17–27. The smalle he phy oplank on, he g ea e e o is equi ed o dis inguish indi idual cells by op ics-based me hods. Flow cy ome y is widely applied in ma ine picophy oplank on esea ch, bu in eshwa e esea ch i s ole has emained mino . We compa ed epi luo es- cence mic oscopy and low cy ome y in assessing he composi ion, abundance and cell sizes o au o luo escen picophy oplank on in epilimnia o 46 Finnish lakes. Phycocyanin- ich picocyanobac e ia we e he mos dominan . The wo me hods yielded compa able o al picophy oplank on abundances, bu he de e mina ion o cell sizes, and hus o al bio- masses, we e on a e age an o de o magni ude highe in he mic oscopy esul s. Howe e , low cy ome y yielded highe cell sizes when applied on small-celled cul u ed algae. Ou s udy demons a ed ha bo h epi luo escence mic oscopy and low cy ome y a e use ul me hods in assessing abundances o phycocyanin- ich and phycoe y h in- ich picocyano- bac e ia and euka yo ic picophy oplank on in lakes. Howe e , accu a e de e mina ion o cell size and biomass emain challenges o mic oscopy and especially o low cy ome y. In oduc ion Epi luo escence mic oscopy and low cy ome y a e he mos commonly used me hods o coun picophy oplank on — he smalles (cell diame e < 2 µm) pho osyn he ic p ima y p oduce s o open wa e s ha a e di icul o obse e wi h a- di ional ligh mic oscopy. Bo h epi luo escence mic oscopy and low cy ome y a e based on dis- inguishing au o luo escence o ligh ha es ing pigmen s in picophy oplank on cells. Based on he au o luo escence p o iles, picophy oplank on can be di ided in o phycocyanin- ich picocy- anobac e ia, phycoe y h in- ich picocyanobac e- ia and euka yo ic picophy oplank on — h ee g oups ha h i e in sligh ly di e en en i on- men s (MacIsaac and S ockne 1993, Callie i 2008). Flow cy ome y is ou inely applied in s ud- ies o ma ine en i onmen s whe e he exis ence o small-celled au o ophs is p onounced, and pa allel mic oscopic obse a ions a e made i necessa y (Collie 2000, Veldhuis and K aay 2000, Johnson and Ma iny 2015). Flow cy om- e y is also highly use ul in enume a ing uni- cellula phy oplank on in labo a o y cul u es 18 Salmi e al. • BOREAL ENV. RES. Vol. 26 (Thomas e al. 2018). Ins ead, he possibili ies o au oma ed luo escence de ec ion and cell coun ing emain a he unde exploi ed in esh- wa e s udies (C osbie e al. 2003, Cellama e e al. 2010). The ac ha mic oscopy is applied mo e commonly in eshwa e esea ch han low cy ome y is p obably due o he be e isualiza- ion p ope ies o mic oscopy, which enable he assessmen o mo e di e se communi ies wi h mul icellula and ilamen ous axa (Pee e s e al. 1989, Toepel e al. 2004). Picophy oplank on abundances based on ei he epi luo escence mic oscopy o low cy ome y can be con e ed o biomasses by assuming a ixed cell olume (Be gkempe and Weisse 2018). Wi h epi luo escence mic oscopy, cell sizes can be es ima ed by measu emen s wi h an eyepiece g a icule o image analysis (Callie i 2008). Sca e ing da a yielded by low cy ome y o ma ine picophy oplank on ha e also been u ilized in quan i ying he a ia ion wi hin cell sizes by es ablishing a con e sion be ween diame e s o e e ence pa icles and o wa d o side sca e ing alues (Cha py and Blancho 1998, Veldhuis and K aay 2000). Be gkempe and Weisse (2018) s a ed ha he exclusion o pico-sized algae is a weak spo in Eu opean lake moni o ing p og ammes whe e phy oplank on is coun ed using ligh mic os- copy. Howe e , low cy ome y o eshwa e picophy oplank on has shown good con o mi y wi h o he quan i ica ion me hods, such as he adi ional epi luo escence mic oscopy (C os- bie e al. 2003, Be gkempe and Weisse 2018), DNA-based quan i ica ion and analysis o pho- osyn he ic pigmen concen a ion (Veldhuis and K aay 2000). Since ins umen a ion o epi luo- escence mic oscopy and low cy ome y is com- monly a ailable, and one o hese wo me hods is ou inely applied, hei con o mi y dese es close inspec ion. In his s udy, we es ed he applicabili y o low cy ome y o assess picophy oplank on in 46 bo eal lakes. In addi ion o abundances, we compa ed biomass es ima es de e mined by he low cy ome y and epi luo escence mic oscopy. In mic oscopy, we used eyepiece g a icule o measu e cell sizes; and in low cy ome y, we used la ex beads o a known size as e e ences. To de e mine he accu acy o he cell size es i- ma es, we measu ed cell diame e s o cul u ed small-celled cyanobac e ia by scanning elec on mic oscopy (SEM), epi luo escence mic os- copy and low cy ome y. We hypo hesised ha epi luo escence mic oscopy and low cy ome- y should show cohe en abundances and bio- masses o he h ee picophy oplank on pigmen g oups. We an icipa ed SEM, epi luo escence mic oscopy and low cy ome y o yield consis- en cell diame e es ima es. Ma e ial and me hods Sampling and sample p ocessing Samples we e aken om epilimnia o 46 lakes in sou he n, cen al and eas e n Finland be ween June and Augus 2015 (Table 1). A wa e sample ( olumes: 2.0 L, 2.6 L o 3.5 L; Limnos, Limnos. pl, Poland) was used o in eg a e he wa e column om he su ace o maximum 2 m dep h o un il a he mocline ( empe a u e change > 1°C pe me e ) was me . Tempe a u e was measu ed wi h a he mome e a ached o he sample . In he labo a o y, wa e o picophy oplank on samples was p e-sie ed h ough a 250 µm mesh and hen h ough a s e ile 5 µm po e size sy inge il e o emo e la ge phy oplank on and o he la ge pa icles. Fo epi luo escence mic oscopy, unp e- se ed picophy oplank on we e collec ed on o black polyca bona e il e s (po e size: 0.22 µm, Me ck Millipo e, Ge many). The black il e s we e we moun ed wi h glyce ol and s o ed ozen a –20°C (e.g. Boo h 1993, Salmi e al. 2014) un il coun ed wi h an epi luo escence mic oscope wi hin 6–8 mon hs. Fo low cy ome y, 2.5–4 mL o p e- il e ed wa e was s o ed in 4.5 mL c yo- ials. To enhance he p ese a ion, pa a o mal- dehyde (16%) was added o a inal concen a ion o 1% (MacIsaac and S ockne 1993). C yo ials we e s o ed ozen a –80°C un il coun ed wi h low cy ome y wi hin 8–10 mon hs. Epi luo escence mic oscopy An Axio Ve .A1 epi luo escence mic oscope (Ca l Zeiss, Ge many) equipped wi h blue (470 nm) and g een (530 nm) LED ligh sou ces BOREAL ENV. RES. Vol. 26 • Me hods o coun ing picophy oplank on 19 Table 1. Lis o sampling da es, sampled lakes and coo dina es o he sampling si es. ID Sampling Da e Lake Coo dina es (dd.mm.yyyy) No h Eas 1 03.08.2015 Ah eninen 62.94957 26.86753 2 08.07.2015 Ala-Kei ele 62.67649 25.86500 3 01.07.2015 Au ejä i 62.04546 23.36944 4 06.08.2015 Hiiden esi 60.39003 24.16344 5 03.08.2015 Hi ijä i 62.95176 26.91284 6 20.07.2015 Iloman sinjä i 62.68340 30.90113 7 29.06.2015 Iso Rau a esi 62.06703 25.04697 8 29.07.2015 Juojä i 62.76788 28.56303 9 27.07.2015 Ka ankajä i 62.71111 24.81972 10 29.07.2015 Ke majä i 62.44644 28.67387 11 08.07.2015 Keski-Kei ele 62.83782 26.02901 12 20.07.2015 Koi e e 62.94894 30.62775 13 20.07.2015 Konni esi 61.13990 26.14805 14 03.08.2015 Koskelo esi 62.67619 26.86563 15 08.07.2015 Kuhnamo 62.60993 25.67405 16 29.07.2015 Lanne esi 62.55910 25.44299 17 07.07.2015 Leppä esi 62.23902 25.95099 18 10.08.2015 Lohjanjä i 60.24951 24.03428 19 27.07.2015 Mahlunjä i 62.67585 25.08897 20 13.07.2015 Muu uejä i 63.08189 25.51459 21 03.08.2015 Niini esi 62.73920 26.85005 22 20.07.2015 Nuo ajä i 62.68001 31.14093 23 01.07.2015 Palo esi 61.89591 23.93470 24 31.07.2015 Pankajä i 63.37140 30.20379 25 06.07.2015 Peu unka 62.44579 25.85154 26 24.07.2015 Pieksänjä i 62.32546 27.13805 27 26.07.2015 Pielinen 63.10321 29.98289 28 10.08.2015 Pihlaja esi 62.35952 24.32613 29 29.07.2015 Pyhäjä i 62.72323 25.44316 30 22.07.2015 Re unen 62.94876 28.65307 31 22.07.2015 Rikka esi 62.80724 28.74926 32 01.07.2015 Ruo esi 62.00598 24.10244 33 05.08.2015 Ru ajä i 61.94827 26.07683 34 27.07.2015 Saa ijä i 62.72306 25.18175 35 29.07.2015 Summasjä i 62.64539 25.38419 36 29.07.2015 Su as esi 62.48378 28.22890 37 19.07.2015 Sääksjä i 61.39501 22.40007 38 15.07.2015 Ta janne 62.13989 24.03638 39 15.07.2015 Tois esi 62.28604 23.73901 40 24.07.2015 Unnukka 62.38304 28.03689 41 06.07.2015 Uu ainen 62.48408 26.13479 42 15.07.2015 Vaski esi 62.13944 23.77312 43 06.07.2015 Va ianjä i 62.48346 25.89432 44 01.08.2015 Viekijä i 63.38422 29.73474 45 13.07.2015 Vuosjä i 62.99522 25.52393 46 13.07.2015 Ylä-Kei ele 63.07631 25.77570 20 Salmi e al. • BOREAL ENV. RES. Vol. 26 was used o coun ing picophy oplank on on he black il e s. The blue LED was connec ed o il e se 09 (EX: BP 450-490, beamspli e : 510, EM: LP 515, Ca l Zeiss, Ge many) and he g een LED o il e se 14 (EX: BP 510-560, beamspli e : 580, EM: LP 590, Ca l Zeiss, Ge many). Phyco- cyanin- ich (he ea e PC cells) showed a he weak deep ed au o luo escence wi h he blue se . Ins ead, wi h he g een se hey showed b igh e ed au o luo escence. Wi h he blue se , au o luo- escence o phycoe y h in- ich picocyanobac e ia (he ea e PE cells) was dis inguished as ligh o ange and ha o euka yo ic picophy oplank- on as ed. PE cells showed b igh o ange and euka yo ic picophy oplank on only weak ed au - o luo escence wi h he g een se (MacIsaac and S ockne 1993). Picophy oplank on we e coun ed wi h 1000× o al magni ica ion om a leas en andomly chosen ields o iew ac oss he black memb ane il e . Flow cy ome y A FACSCalibu low cy ome e (Bec on-Dick- inson, USA) equipped wi h 488 nm lase exci a- ion, de ec o s o o wa d (FSC) and side (SSC) sca e s and h ee channels o luo escence de ec ion: g een (FL1, 530/30 nm), o ange (FL2, 585/42 nm) and ed (FL3, 650 nm, LP) was used in his s udy. Picophy oplank on we e di ided in o PC, PE and euka yo ic cells acco ding o he in ensi ies o hei o ange and ed luo es- cen signals (Fig. 1). The applied se ings o pho omul iplie s we e FSC-H E01, SSC-H 350, FL1 600, FL2 600 and FL3 600. All channels we e deployed in loga i hmic mode. The p i- ma y pa ame e o eco ding coun ed e en s was side sca e and seconda y pa ame e ed luo escence. The h eshold alue o bo h was 25. A low low a e was used o all samples and 10 000 e en s we e eco ded. The low a e was e i ied wice e e y day (be o e and a e samples) by weigh ing 1 mL o wa e wi h an analy ical balance (AT21, Me le -Toledo, Aus- alia, eadabili y 1 µg), unning i on low cy ome e o 5–10 minu es and weighing i again o measu e he olume o he wa e ha wen h ough he low cell in a speci ied ime. Daily speci ic low a es (mean 10 µL min–1, SD = 0.9, numbe o wo king days = 9) we e used o con e he coun ed e en s in o cell Fig 1. Picophy oplank on cy og ams selec ed o illus a e he a ia ion in lakes Koi e e, Keski-Kei ele, Muu ue- jä i, Pieksänjä i, Re unen, Rikka esi, Su asjä i and Viekijä i. Phycocyanin- ich (PC) cells a e indica ed by blue do s and euka yo ic cells by ed do s. The second plo o Lake Keski-Kei ele desc ibes he loca ion o phycoe y h in- ich (PE) cells ep esen ed by o ange do s. No e ha bo h axes a e on a loga- i hmic scale. BOREAL ENV. RES. Vol. 26 • Me hods o coun ing picophy oplank on 21 abundances. Cell size es ima es we e ob ained by es ablishing a eg ession model be ween o - wa d sca e and diame e s o e e ence mono- dispe sal la ex beads (Phinney and Cucci 1989). Diame e s o he used la ex beads we e 0.3 µm, 0.6 µm (Sigma-Ald ich, USA), 1.0 µm (Beck- man Coul e , USA), 3.0 µm (Sigma-Ald ich, USA) and 6.0 µm (The mo Fishe Scien i ic, USA) and hei o wa d sca e s we e eco ded by he low cy ome e on 2–5 sepa a e wo king days wi h he same FSC and SSC se ings as o he picophy oplank on (Fig. 2). Flowing So wa e e . 2.5.1 (Uni e si y o Tu ku, Finland) was used o p ocessing he low cy ome y da a. Picophy oplank on pop- ula ions we e delimi ed om he sca e plo s wi hou i s sc u inizing he mic oscopy esul s o a oid bias in he in e p e a ion. The FSC o each cell was con e ed o cell diame e s be o e he a e age cell diame e in he popula ion was calcula ed. Cell size measu emen s o cul u ed cyanobac e ia F esh clonal cul u es o Ch oococcus, Snowella and Synechococcus sp. in liquid Z8 medium we e il e ed h ough 5 µm po e size sy inge il e s simila ly o lake wa e samples. Sam- ples o epi luo escence mic oscopy and low cy ome y we e p epa ed as desc ibed abo e and examined wi h he same ins umen s as he lake samples. Using epi luo escence mic os- copy, he diame e s o cul u ed cyanobac e ia we e measu ed wi h he g een exci a ion il e se . Wi hin low cy ome y, same samples o each cul u e we e un on h ee consecu i e days. Fo scanning elec on mic oscopy (SEM), 1–3 mL o each cul u e was collec ed on a Wha man GF/F il e and ixed wi h 2.5% glu- a aldehyde in 0.1 M phospha e bu e (pH 7.4). A e washing wi h phospha e bu e and insing wi h wa e , samples we e dehyd a ed wi h a g aded e hanol se ies. D ying was done using a c i ical poin d ye (K850, Quo um Technologies, UK) and a e wa ds, samples we e a ached o specimen s ubs, hey we e spu e coa ed (Q150 T ES, Quo um Technol- ogies, UK) wi h a hin laye o pla inum. Sam- ples we e imaged wi h ield-emission scanning elec on mic oscope (Sigma HD VP, Ca l Zeiss AG, Ge many). The p epa a ion and imaging o samples was done in he Elec on Mic oscopy Co e Facili y o Biocen e Oulu, Uni e si y o Oulu. S a is ical me hods To coun picophy oplank on wi h he epi - luo escence mic oscope (see Epi luo escence mic oscopy sec ion), a p op ie a y compu e p og amme was used o calcula e 95% con i- dence in e als o mean abundances and bio- masses in eal ime o op imize coun ing e o (e.g. Salmi e al. 2014). Con idence in e als o mean abundance we e calcula ed as: (1) whe e 0.025 is he 97.5% pe cen ile o he dis- ibu ion wi h n—1 deg ees o eedom, s2 is he sample a iance and n is he numbe o epli- ca e mic oscope iews. Shapes (sphe e, o a ional ellipsoid, cylin- de ) o picophy oplank on cells we e es ima ed and he main dimensions measu ed wi h an eyepiece g a icule a a scale o 1 µm. The bio- mass o picophy oplank on was calcula ed by assuming a cell densi y equal o ha o wa e . Fig 2. Compa ison o mean o wa d sca e ing (FSC) and nominal diame e o la ex beads, eg ession equa- ion and deg ee o explana ion. c lmean %, / . 100 0025 2 s n 22 Salmi e al. • BOREAL ENV. RES. Vol. 26 Con idence in e als o mean biomass o pico- phy oplank on we e calcula ed as: (2) whe e 0.025 is he 97.5% pe cen ile o he dis ibu ion wi h n—1 deg ees o eedom and, de ia ing om he equa ion 1, ∑s2 is sum o he a iances o biomasses o di e en size classes and n is he numbe o eplica es. The equi ed numbe o coun ed iews was es ima ed so ha c l% ≤ 30 o o al biomass was eached, bu a leas 10 iews we e coun ed. Flow cy ome y-based abundances we e compa ed agains mic oscopy-based abun- dances and biomasses by sc u inizing i he low cy ome y esul s all inside he con i- dence in e als o he mic oscopy assessmen s. Co ela ions be ween he wo me hods we e sc u inized using Spea man’s co ela ion and he signi icance o he di e ences we e es ed wi h a ela ed-samples Wilcoxon signed- ank es . The non-pa ame ic app oaches we e chosen because o he ela i ely low numbe o he samples (46 lakes). The s a is ical analyses we e made using SPSS S a is ics e . 26 (IBM, USA). Fig. 3. Picophy oplank on abundance a ios de e mined by epi luo escence mic oscopy and low cy ome y o all picophy oplank on cells (To al picophy oplank on,uppe le panel), phycocyanin- ich cells (PC cells, uppe igh panel), phy oe y h in- ich cells (PE cells, lowe le panel), and euka yo ic picophy oplank on (Euka yo ic cells, lowe igh panel). The ho izon al e o ba s ep esen he 95% con idence in e al o he a e age abundance in he sample coun ed using an epi luo escence mic oscope. No e he di e en axis scales in he uppe and lowe panels. c lmean %, / . 100 0025 2 n s BOREAL ENV. RES. Vol. 26 • Me hods o coun ing picophy oplank on 23 Resul s Ou da ase om 46 Finnish lakes o e ed a good basis o e alua e he consis ency o picophy o- plank on abundance and biomass measu emen s by epi luo escence mic oscopy and low cy om- e y in na u al condi ions. He e, we epo he quan i a i e esul s yielded by he wo me hods as well as he compa isons o cell size measu e- men s based on epi luo escence mic ocopy, low cy ome y and SEM. Abundance Acco ding o mic oscopic obse a ions o he lake samples, mos (mean: 77%, SD = 7) o he picophy oplank on uni s we e soli a y cells and only a ew di iding cells o small colonies we e obse ed in he lake samples. Some o he indi- iduals, howe e , migh ha e o igina ed om colonies, des oyed du ing p e- il a ions. To al picophy oplank on abundances in mic oscopi- cally coun ed samples a ied om 2.0 × 103 o 1.0 × 105 uni s mL–1 (Fig. 3). PC cells co e ed mos (mean: 80%, SD = 17) o ha . The p o- po ion o euka yo ic uni s was no ably lowe (mean: 18%, SD = 17) and only a ew samples con ained PE cells (Fig. 3). Consis en wi h he luo escence mic oscopy samples, PC cells also accoun ed o he majo - i y o low cy ome y samples, and euka yo ic and PE cells we e p esen in small numbe s (Fig. 3). Fo o al picophy oplank on abundance, he median mic oscopy/ low cy ome y a io was 1.0 (mean: 1.2, SD = 0.81). The di e ence be ween o al abundances was no signi ican ( ela ed-samples Wilcoxon signed- ank es , p = 0.9) and 46% o he low cy ome ically de e mined o al abundances ell inside he 95% con idence in e als o co esponding mic o- scopic esul s (Fig. 3; Spea man’s co ela ion, ρ = 0.79, p < 0.001). When PC cells, PE cells and euka yo ic cells we e analysed sepa a ely, he median mic oscopy/ low cy ome y a ios we e 0.92, 1.01 and 0.46, espec i ely (means: 1.0, SD = 0.89; 1.2, SD = 1.22 and 0.49, SD = 0.37, espec i ely). The di e ences o picocyano- bac e ia we e no signi ican ( ela ed-samples Wilcoxon signed- ank es , p = 0.2 o PC and p = 0.06 o PE). Addi ionally, 50% o he low cy ome ically de e mined PC and 75% o he PE cell abundances we e inside he 95% con i- dence in e als o mic oscopic cell coun s. Fo euka yo ic picophy oplank on, low cy ome y ga e signi ican ly highe coun s han mic os- copy did ( ela ed-samples Wilcoxon signed- ank es , p < 0.001), and 22% o he low cy ome y assessmen s we e inside he 95% con idence in e als o mic oscopy-based abundance assess- men s. The a e age ime used o coun ing by low cy ome e was ela i ely long (mean: 9 min- u es SD = 5), since he na u al samples we e a he spa se o low cy ome y. Compa ed wi h mic oscopy whe e he coun ing e o was adjus ed o each he 95% con idence in e - als o he mean biomass, low cy ome y was app oxima ely en old as e . The o al numbe o coun ed picophy oplank on uni s was no a- bly highe in low cy ome ic samples (a e age 1885, SD = 1280) compa ed wi h mic oscopic coun s (mean: 162, SD = 66). Thus, imewise low cy ome y was a mo e cos -e ec i e way o coun picophy oplank on. Cell size and biomass In mic oscopic es ima ions, he mos common cell diame e in he lake samples was 1 µm (70%, SD = 18). Ins ead, a e age diame e based on e e ence beads and o wa d sca e (FSC) was s ikingly di e en : 0.56 µm (SD = 0.39). The e o e, biomass es ima es we e, on a e age, 15- old highe (SD = 23) using mic oscopy, and co ela ion be ween FSC and mic oscopy-based o al biomass es ima es was a he weak (Spea - man’s co ela ion, ρ = 0.51, p < 0.001, Fig. 4). Cell size measu emen s o cul u ed cyanobac e ia Scanning elec on mic oscopy (SEM), epi luo- escence mic oscopy and low cy ome y yielded simila cell diame e s o picosized Synechoc- occus (Table 2). Howe e , o Ch oococcus and Snowella, low cy ome y ga e no ably wide diame e s han SEM o epi luo escence mic os- 24 Salmi e al. • BOREAL ENV. RES. Vol. 26 copy. Addi ionally, he a io o epi luo escence mic oscopy-based cell abundances o low cy ome y-based uni coun s was less han 1 (Table 2). Thus, he wide diame e es ima es we e likely pa ly due o he appea ance o Ch oococcus and Snowella as di iding cells and mic ocolonies a he han as indi idual cells; in low cy ome y pa icles posi ion hemsel es along hei longes diame e s. The p esence o di iding cells was con i med by he mic oscopy me hods. Discussion To sc u inize he o e all cong uence o he abun- dances and biomasses assessed by epi luo es- cence mic oscopy and low cy ome y in ou s udy, se e al sou ces o a ia ion need o be conside ed. Va ia ion accumula es om (a) sam- pling and sample p ocessing; (b) s o age o he samples; and he (c) coun ing p o ocols hem- sel es. Since hese s eps a e a he well ou lined in li e a u e o di e en ypes o mic oscopy and low cy ome y samples, we emphasise in he discussion he coun ing p ocedu es and accu- mula ion o o al a ia ion. Abundance C osbie e al. (2003) compa ed low cy om- e y and epi luo escence mic oscopy in coun ing picophy oplank on om Lake Mondsee. Using a FACSCalibu ins umen , simila o he one we used he e, hey epo ed s ong eg ession be ween he wo me hods when coun ing soli a y picophy oplank on ( 2 = 0.95) o mic ocolonies ( 2 = 0.88). In ou s udy, co ela ion be ween he me hods emained lowe (Spea man’s co ela- ion, ρ = 0.79, p < 0.001). In ou s udy, samples o mic oscopy and low cy ome y we e pa al- lel subsamples ha we e p ocessed and s o ed di e en ly, bu acco ding o he gene al ou - lines o he discipline. Addi ionally, he success o mic oscopic coun ing om il e s depends highly on he dis ibu ion o coun ed uni s on he il e . The dis ibu ion is o en expec ed o be andom, which is seldom he case due o he p ope ies o he il e ing sys em and in e ac ions be ween cells (e.g. McNabb 1960, San o d e al. 1969). The agg ega ion o cells was likely a majo eason o he some imes ela i ely wide con idence in e als o he mean o al abundance (up o 49% o he mean abundance; Fig. 3). In low cy ome y, in addi ion o he dis ibu ion o cells in he sample wa e , e o migh accumu- la e om inconsis en low a e. To o e come his, e e ence beads wi h known concen a ions Fig 4. Rela ionships be ween o al picophy oplank on biomasses (mg m–3) de e mined by epi luo escence mic oscopy and low cy ome y. Table 2. Mean cell diame e o cul u ed cyanobac e ial cells es ima ed by di e en me hods, and he abundance a io assessed by epi luo escence mic oscopy (cells) and low cy ome y (uni s). S anda d de ia ions o scanning elec on mic oscopy (SEM) and low cy ome y (Flow) a e indica ed in pa en heses. In epi luo escence mic oscopy, s anda d de ia ions a e no gi en due o sepa a e, eyepiece g a icule-based measu emen s. Cell diame e s SEM (μm) Epi luo escence (μm) Flow (μm) Ch oococcus 2.71 (0.27) 3 5.0 (4.4) Snowella 2.36 (0.31) 3 4.4 (0.6) Synechococcus 0.73 (0.13) 1 1.0 (0.3)