Vol
. 11
3: 1
85-1
98,
1994
--
MA
RIN
E E
CO
LO
GY
PR
OG
RE
SS
SE
RIE
S
Mar
. E
col.
Pro
g. S
er.
-p-
-
Pub
lish
ed O
ctob
er 1
3
-- -
Dis
trib
uti
on, s
ize
and
bac
teri
al c
olon
izat
ion
of
tra
nsp
aren
t ex
opol
ymer
par
ticl
es (
TE
P)
in t
he
ocea
n
U. P
asso
w l,
A.
L. ~
lld
re
d~
e
Ma
rin
e S
cien
ce I
nsti
tute
, 'B
iolo
gica
l S
cien
ces,
Uni
vers
ity
of C
alif
orni
a, S
anta
Bar
bar
a, C
alif
orni
a 93
10
6, U
SA
AB
STR
AC
T: T
he d
istr
ibut
ion,
ab
un
dan
ce, a
nd
siz
e sp
ectr
a of
tr
ansp
aren
t ex
opol
ymer
par
ticl
es (TEP)
gene
rate
d fr
om e
xcre
tion
pro
duct
s of
phyt
opla
nkto
n w
ere
inve
stig
ated
. T
EP
was
fou
nd a
t al
l st
atio
ns
stud
ied,
but
ab
un
dan
ce a
nd
tota
l qu
anti
ty o
f TEP
vari
ed b
y 4 o
rder
s of
mag
nitu
de a
mon
g co
asta
l, s
lope
an
d o
cean
ic e
nvir
onm
ents
. A
bund
ance
an
d s
ize
dist
ribu
tion
s of
TE
P ap
pea
red
to
be
rela
ted
to p
lank
- to
n co
mpo
siti
on.
Abu
ndan
ce t
ende
d to
be
high
er i
n co
asta
l an
d sh
allo
w w
ater
s co
mpa
red
to o
cean
ic
and
dee
p w
ater
s re
spec
tive
ly.
The
av
erag
e si
ze o
f T
EP
was
app
reci
ably
lar
ger
at s
tati
ons
wh
ere
diat
oms
dom
inat
ed p
hyto
plan
kton
. S
ize
dist
ribu
tion
s of TEP g
ener
ally
fol
low
ed a
pow
er-l
aw f
unct
ion,
ex
cept
wh
en a
ggre
gati
on d
omin
ated
int
erac
tion
of
part
icle
s. The t
ype
01 s
ize
dist
ribu
tion
of
TE
P m
ay
allo
w u
s to
pre
dict
th
e te
nden
cy o
f a
plan
kton
com
mun
ity
to a
gg
reg
ate.
Gen
eral
dif
fere
nces
in
the
shap
e of
siz
e sp
ectr
a of
TE
P i
ndic
ate TEP
gene
rate
d by
dia
tom
s to
be
stic
kier
an
d m
ore
frac
tal
com
- pa
red
to TEP g
ener
ated
by
non-
diat
om p
lank
ton.
All
TE
P w
ere
colo
nize
d by
bac
teri
a, b
ut n
o co
rrel
a-
tion
was
foun
d be
twee
n nu
mbe
rs of
att
ache
d ba
cter
ia o
n in
divi
dual
TEP a
nd
the
resp
ecti
ve s
ize
of TEP.
Bac
teri
a de
nsit
y on
TE
P, h
owev
er, d
ecre
ased
wit
h in
crea
sing
siz
e of
TEP e
xpon
enti
ally
. Bet
wee
n 2 a
nd
25
% o
f th
e to
tal
bact
eria
l po
pula
tion
wer
e at
tach
ed t
o TEP. T
otal
num
ber
of b
acte
ria
wer
e no
t co
rre-
la
ted with
tota
l qua
ntit
y of
TEP, s
ugge
stin
g th
at n
o si
mpl
e ov
eral
l rel
atio
nshi
p ex
ists
bet
wee
n TEP
and
ba
cter
ia
KE
Y W
OR
DS.
TE
P . A
ttac
hed
bact
eria
S
ize
freq
uenc
y di
stri
buti
ons
INT
RO
DU
CT
ION
A
new
cl
ass
of
tran
spar
ent
part
icle
s,
call
ed T
EP
(t
ran
spar
ent e
xopo
lym
er p
arti
cles
), h
as r
ecen
tly
been
di
scov
ered
to
be
abu
nd
ant
in t
he
oce
an (
All
dred
ge e
t al
. 1
99
3).
TE
P ca
n b
e m
ade
visi
ble
by
sta
inin
g w
ith
alci
an b
lue,
a s
tain
spe
cifi
c fo
r po
lysa
ccha
ride
s. T
EP
ar
e pr
esum
ably
pro
duce
d fr
om d
isso
lved
car
bohy
drat
e po
lym
ers
exud
ed b
y ph
ytop
lank
ton
and
bac
teri
a. I
n se
awat
er t
hes
e po
lym
er m
olec
ules
ali
gn a
nd
coa
lesc
e vi
a ca
tion
bri
dgin
g, f
orm
ing
invi
sibl
e pa
rtic
les
rang
ing
from
mic
rons
to
hu
nd
red
s of
mic
rons
in
size
. W
hile
pr
elim
inar
y st
udie
s in
dica
te t
hat
TE
P a
re im
port
dnt.
in
the
aggr
egat
ion
of di
atom
blo
oms,
pro
vide
the
mat
rix
of m
arin
e sn
ow, s
erv
e as
a s
ubst
rate
an
d m
icro
habi
tat
for
atta
ched
bac
teri
a, a
nd
rep
rese
nt
an i
mpo
rtan
t pat
h-
way
for
th
e tr
ansf
orm
atio
n of
clis
solv
ed t
o pa
rtic
ulat
e
mat
ter
(All
dred
ge e
t al
. 19
93, P
asso
w e
t al
. 1
99
4),
the
di
stri
buti
on,
abu
nd
ance
, an
d ch
arac
teri
stic
s of
this
ne
w c
lass
of
part
icle
s re
mai
n la
rgel
y u
nk
no
wn
. In
form
atio
n on
th
e ab
un
dan
ce a
nd
siz
e di
stri
buti
ons
of
TE
P p
rodu
ced
in d
iffe
rent
oce
anog
raph
ic r
egim
es
and
by
diff
eren
t sp
ecie
s of
phy
topl
ankt
on i
s ne
cess
ary
to a
ccur
atel
y as
sess
th
e si
gnif
ican
ce o
f th
is n
ew p
arti
- cl
e cl
ass
to t
he
biol
ogy
and
che
mis
try
of th
e pe
lagi
c zo
ne a
nd
to
iden
tify
tax
a m
ost
like
ly t
o p
rod
uce
TE
P
abun
dant
ly.
In
the
pres
ent
stud
y w
e co
mp
are
the
abu
nd
ance
an
d s
ize
dist
ribu
tion
of
TE
P g
ener
ated
by
diff
eren
t ph
ylop
lank
ton
spec
ies
and
d
eter
min
e th
e si
ze d
istr
ibut
ions
an
d a
bu
nd
ance
s of
T
EP
in
coas
tal
and
oc
eani
c en
viro
nmen
ts
in
rela
tion
to
pl
ankt
on
com
posi
tion
. T
he
role
of
T
EP
as
a si
te f
or b
acte
rial
at
tach
men
t in
di
ffer
ent
mar
ine
envi
ronm
ents
is
also
in
vest
igat
ed.
0 I
nter
-Res
earc
h 19
94
Res
dle
of
full
art
icle
not
pe
rnu
tfe
d
Mar. E
col, Prog S
er. 113: 185-198, 1994
ME
TH
OD
Cu
ltures and
samp
le collection.
Abundance
and
particle size frequency distributions of
TE
P w
ere in-
vestigated in batch cultures and
seawater. B
atch cul- tures ot" C
haetoceros gracili?,, Nitzschia angularis, T
ha
- lassiosira
weissflogij an
d
colonial P
fiaeocystis were
gro
wn
in f/2 m
edium w
ith silica at 12OC
and
100 pE
m-2
S''
light (12
:12
h cycle) in unbubbled Fernbach
flasks. Care w
as taken
to prep
are the media free of
precip
itates, as
salt precipitates
gen
erated
du
ring
autoclaving produce stain
able artefacts w
hich are dif- ficult to distinguish from
TE
P. C
ontrols for precipitates w
ere mad
e for each batch of
media.
Seaw
ater samp
les were collected on several o
cca- sions d
urin
g Ju
ne an
d July
1992 off
Santa B
arbara, C
aliforn
ia, USA
(34"
20
' N,
119' 50
' W),
by scu
ba
divers at 5 to 20 m
dep
ths. S
eawater sam
ples were col-
lected with
Niskin bottles in
August
1992 from 0 to
400 m d
epth
s in Monterey B
ay, California (36' 48' N
, 121Â 5
3' W
and
36' 4
6' N
, 121' 57
' W) an
d in M
arch 1
99
3 fro
m 300 to 1500 m
dep
ths in th
e open Atlantic off
Berm
uda (BA
TS
Stn 54; 3
1" 50' N
, 64' 10' W
]. S
lide p
reparatio
n an
d en
um
eration
. Sem
i-perrna- n
ent slides of T
EP
were p
repared
in duplicate by pass-
ing
l to 20 m
1 of seaw
ater onto a 0.4 pm polycarbonate
filter (Poretics). T
EP w
ere stained
on dam
p filters w
ith an
aqu
eou
s solution of 0
.02
% alcian blue (8 G
X), a
hydrophilic, catonic dy
e specific for polysaccharides, an
d 0
.06
% acetic acid. T
he stain
ing
solution was p
re- filtered th
rou
gh
0.2 pm before use. as th
e stain agg
re- g
ates with itself. T
he filter w
as fully covered with stain
(abo
ut 400 to 500 pi) w
hich was im
mediately sucked
through (as stainin
g is im
med
iate). Since residual stain
reacts with salts in
seawater to create artefacts, funnels
and
base w
ere rinsed with distilled w
ater before use.
Vacuum
pressu
re was k
ept low
and
constant. Filters
were th
en transferred to slides an
d p
repared
accord- in
g
to th
e F
ilter-Transfer-F
reeze (FTF) technique
(Hew
es & H
olm-H
ansen 1
98
3, H
ewes
et al. 1984).
Blanks w
ere prep
ared w
ith 'clean' seaw
ater (filtered rep
eatedly
through 0.2
pm
) or by
dipping the filter into seaw
ater before staining as described abo
ve.
All sam
ples, ex
cept th
ose from
Berm
ud
a, were p
re- p
ared fresh
. Sam
ples from
Berm
uda were preserved in
buffered form
alin (0
.5 to
1%
) and
slides p
repared
w
ithin 2 wk
after fixation. Slides m
ay be stored at room
temp
erature in th
e dark
for up
to 6 mo.
Total
bacteria ab
un
dan
ces w
ere en
um
erated
on black
polycarbonate filters
(Poretics, 0
.2 urn)
after staining w
ith acridine oran
ge according to H
obble et at. (1
977J. S
amples
were d
ou
ble-stain
ed w
ith D
AP
I (4
',6-d
iamid
ino
-2-p
hen
y lindole) an
d
alcian blue
to en
um
erate bacteria
attached
to T
EP
. Repljcate
sea- w
ater samples w
ere first stained
in th
e funnel with
0.5 ml D
AP1 (25 pm
ml"') for 5
to 10 min, filtered onto
0.2
urn polycarbonate filters (Poretics] (P
orter & F
eig 1980, K
ing & P
arker 1988) an
d them
stained with alcian
blu
e as described above. Th
e FT
F technique h
as been reported to yield accurate estim
ates of bacterial ab
un
- d
ances (K
ing & P
arker 1988) which w
e could confirm
(Alldredge el al. 1
993).
Phytoplankton an
d T
EP
(> 3 urn) w
ere counted in 20 to 150 fields on each
slide at 200x magnification w
ith a com
pound light microscope. T
EP (N
) were assigned to
6
size classes by
maxim
um
leng
th, w
ith each
size increm
ent (dl) double th
e length of th
e previous. The
area of TE
P wcis calculated based on m
easurem
ents of
the actual size of 20 representive TE
P in each size class
for each set of
samples. A
lthough the shap
e of TE
P v
ar- ied
and
often was am
orphous, TE
P in an
y o
ne sam
ple an
d size class w
ere of sim
ilar shap
e and
the area of
TE
P on filters can be estim
ated fairly accurately with
this method (P
assow &
AL
ldredge un
pu
bl.).
Total bacteria an
d bacteria attach
ed to T
EP
were
counted using a Zeiss epifluorescence m
icroscope at a m
agnification of 8
00
x. T
en fields were counted fro
m
each of
2 replicate slides for total bacteria counts. Bac-
teria associated with 20 to 30 T
EP per slide w
ere enu
- m
erated by sw
itching betw
een W
and
visible Light
and
individual TE
P w
ere sized. Th
e averag
e numbers
of bacteria per area of
TEP were calculated for each
particle.
Estim
ates of
bacteria attach
ed
to T
EP
are considered conservative because T
EP are presum
ably 3-dim
ensional, making enum
eration on 1 plan
e diffi- cult, an
d thin sh
eets of T
EP
were difficult to discern.
TE
P concentrations in B
ermuda sam
ples were too low
to en
um
erate attached
bacteria, Im
pact of form
alin fixation. C
ontrols sho
wed
that form
alin itself do
es not interfere with th
e stain. H
ow-
ever, since the binding capacity of
alcian blue dep
end
s on
the
pH
(Parker
&
Diboll
1966, Horobin
1988), preservation
techniques w
hich alter
the p
tl of
the
sample are expected to affect quantitative results. W
e fo
un
d how
ever that small ch
ang
es in pH
(8.2 to 8.7) did not ch
ang
e the size or am
ount of stain
ed particles.
We
tested the im
pact of form
alin preservation on
TE
P using sam
ples collected d
urin
g a diatom
bloom
and
from a culture of colonial P
haeocysiis. S
lides from
samp
les collected during a diatom bloom
in Jun
e off S
anta B
arbara were prepared both from
fresh sam
ples an
d from
aliquots which had b
een p
reserved
in forma-
lin (0.5
to 1%
) for 6 mo. C
omparisons of
the particle
size distributions of T
EP
in preserv
ed an
d fresh ali-
quots revealed no significant difference [l-test, df = 4,
p >
0
.1), although
form
alin-p
reserved
samples
ap-
peared
to slightly underestimate total T
EP
abu
nd
ance.
Unpreserved, intact an
d healthy colonies of
Ph
aeo- cystis did not stain w
ith our method as stronger co
n-
centrations and longer staining times are n
eeded
to
Pas
sow
& A
lldre
dge:
Dis
trib
~ili
on of
tra
nsp
iirc
nl cxopolynier ji
rtrl
icle
s
stai
n th
e co
lony
m
atri
x.
How
ever
, fo
rmal
in p
rese
r-
vati
on l
ead
s to
th
e di
srup
tion
of
th
e co
lony
mat
rix
and
sh
red
s of
th
e m
atri
x st
ain,
cr
eati
ng
TE
P-l
ike
arte
fact
s.
Fra
gile
org
anis
ms
like
cil
iate
s or
ath
ecat
e di
nofl
agel
late
s of
ten
disi
nteg
rate
in
for
mal
in a
nd
th
e ce
llul
ar
poly
sacc
hari
des
rele
ased
m
ay
also
cr
eate
ar
tefa
cts.
C
alcu
lati
on o
f th
e fr
acti
on o
f ba
cter
ia a
ttac
hed
to
TE
P. W
e ca
lcul
ated
the
fra
ctio
n of
bact
eria
att
ache
d to
T
EP
in
2 d
iffe
rent
way
s fr
om o
ur c
ount
s. F
irst
, th
e p
er-
cen
tag
e of
att
ach
ed b
acte
ria
was
cal
cula
ted
dire
ctly
for
each
dou
ble-
stai
ned
slid
e bo
rn m
icro
scop
e co
unts
of
tree
ver
sus
atta
ched
ba
cter
ia.
Sec
on
d,
beca
use
the
mag
nifi
cati
on n
eed
ed t
o co
unt
bact
eria
is
som
etim
es
too
high
to s
imul
tane
ousl
y co
unt l
arg
e T
EP
rep
rese
nta
- ti
vely
, we
calc
ulat
ed t
he
frac
tion
of
bact
eria
ass
ocia
ted
wit
h T
EP
fro
m (
1) t
he s
ize
dist
ribu
tion
of
TE
P (TEP
slid
es),
(2
) the
cou
nts
of to
tal
num
ber
of b
acte
ria
(acr
-i-
dine
ora
ng
e sl
ides
) an
d (
3) t
he
gen
eral
rel
atio
nshi
p b
etw
een
bac
teri
al d
ensi
ty o
n T
EP
(n
um
ber
of b
acte
ria
per
area
TE
P)
and
TE
P s
ize
deri
ved
from
all
cou
nts
or
doub
le-s
tain
ed s
lide
s. I
n g
ener
al,
both
ca
lcul
atio
ns
gav
e si
mil
ar
resu
lts,
al
thou
gh
som
etim
es
the
firs
t m
etho
d g
avr
high
er v
alue
s. A
ll v
alue
s gi
ven
in t
he
pre
sen
t p
aper
are
cal
cula
ted
acco
rdin
g to
th
e se
cond
m
etho
d an
d a
re c
onsi
dere
d co
nser
vati
ve.
Par
ticl
e-si
ze
dist
ribu
tion
s. P
arti
cle-
size
spe
ctra
of
susp
end
ed m
atte
r in
th
e oc
ean
can
ofte
n b
e fi
tted
to
a po
wer
-law
dis
trib
utio
n de
rive
d fr
om a
vol
ume
dist
iil.
~n-
li
on h
isto
gram
tha
t re
pre
sen
ts th
e di
stri
buti
on i
n w
hich
th
e vo
lum
e of
par
ticl
es o
f ea
ch s
ize
inte
rval
rem
ains
co
nsta
nt
(Bad
er 1
970,
She
ldon
el
'11.
19
72,
Ler
mii
n 19
79).
Dev
iati
ons
from
suc
h a
dist
ribu
tion
m<
iy g
ive
indi
cati
ons
of p
hysi
cal
or b
iolo
gica
l pr
oces
ses,
suc
h as
ag
gre
gat
ion
, fr
agm
enta
tion
, di
ssol
utio
n, o
r co
nsum
p-
tion
by
anim
als,
whi
ch a
lter
the
siz
e di
stri
buti
on o
f pa
rtic
les.
P
arti
cle-
size
dis
trib
utio
ns o
f T
EP
wer
e d
eter
min
ed
from
mic
rosc
opic
al e
num
erat
ions
. T
he
mea
n m
axim
al
len
gth
val
ue o
f ea
ch s
ize
inte
rval
(l,
in p
m)
and
the
ra
tio
betw
een
TE
P a
bu
nd
ance
per
inte
rval
an
d t
he s
ize
inte
rval
(dN
/di,
in T
EP
ml-* pm
") w
ere
plot
ted
on l
og-
log
coor
dina
te g
rap
hs.
Th
e be
st f
it li
nes
wer
e ca
lcu-
la
ted
bas
ed o
n a
leas
t-sq
uar
es e
stim
atio
n of
pa
rarn
e-
ters
A a
nd
b in
the
fol
low
ing
pow
er-l
aw r
elat
ions
hip
wh
ere
A a
nd
ba
re c
onst
ants
. Th
e sl
ope
of t
he s
ize
dis-
tr
ibut
ions
plo
tted
in l
og-l
og c
oord
inat
es is
giv
en b
y th
e po
wer
ex
po
nen
t b a
nd
des
crib
es
the
shap
e of
th
e cu
rve.
A d
ecre
ase
in b
sig
nifi
es a
red
ucti
on i
n th
e p
er-
cen
tag
e of
sm
all
part
icle
s. A
val
ue o
f b =
4 i
mpl
ies
equa
l pa
rtic
le v
olum
es i
n ea
ch s
ize
clas
s. A
ch
ang
e in
A
wh
en b
rem
ains
con
stan
t si
gnif
ies
a ch
ang
e in
the
to
tal
num
ber
of p
arti
cles
in
all
size
cla
sses
.
Par
ticl
e-si
ze d
istr
ibut
ions
of
TE
P a
re b
ased
on
th
e m
axim
al l
eng
th o
f T
EP
on
fil
ters
. T
he
actu
al s
ize
of
TE
P
wh
en
susp
end
ed
in
seaw
ater
is
pr
esum
ably
sm
alle
r th
an t
he
size
of
TE
P a
fter
th
ey c
olla
pse
on f
il-
ters
(P
asso
w &
All
dred
ge u
npub
l.).
It
is,
how
ever
, as
- su
med
tha
t th
e sh
ape
of
the
size
dis
trib
utio
ns o
f T
EP
, i.
e. t
he r
elat
ive
size
of
TE
P,
rem
ains
th
e sa
me
du
rin
g
filt
rati
on.
For
th
e co
asta
l d
ata,
reg
ress
ion
line
s w
ere
only
cal
- cu
late
d fo
r sa
mp
les
wh
ere
TE
P o
ccur
red
in a
t le
ast
4 di
ffer
ent
size
cla
sses
. All
regr
essi
ons
of
sam
ples
for
w
hich
cal
cula
tion
s w
ere
poss
ible
(28
tota
l) w
ere
stat
is-
tica
lly
sign
ific
ant
(p Ã
0.01
, in
mos
t ca
ses
p 0.
001)
, ex
cept
for
2 s
ampl
es w
hich
did
not
fol
low
a p
ower
-law
di
stri
buti
on a
nd
wer
e co
llec
ted
duri
ng t
he
agg
reg
atio
n
phas
e of
a
diat
om b
loom
. T
EP
con
cent
rati
ons
at t
he
open
oce
an s
tati
on o
ff B
erm
uda
(BA
TS
) wer
e lo
w a
nd
re
gres
sion
s w
ere
calc
ulat
ed if
TE
P w
as f
ound
in
3 or
m
ore
size
cla
sses
(9
out
of
12
dep
ths)
. Sev
en r
egre
s-
sion
s w
ere
sign
ific
ant
at t
he
0.0
5 le
vel.
T
o as
sess
how
ag
gre
gat
ion
ch
ang
es t
he
size
dis
trib
- ut
ion
of
TE
P,
size
dis
trib
utio
ns w
ere
calc
ulat
ed f
rom
a
floc
cula
tion
ex
per
imen
t w
ith
a ba
tch
cult
ure
of
Ch
aeto
cero
s gr
acili
s co
nduc
ted
earl
ier
(Pas
sow
et
al.
1994
). F
or
that
ex
peri
men
t th
e cu
ltur
e ha
d b
een
ro
tate
d fo
r 20
rni
n in
a c
uv
ette
flo
ccul
ator
(L
ogan
&
Kir
chm
ann
19
91
) an
d s
ubsa
mpl
es h
ad b
een
coll
ecte
d af
ter
4, 8
an
d 20
inin
of
rota
tion
at
a sh
ear
of
30 S
-'.
Cal
cula
tion
s w
ere
base
d on
cou
nts
of
TE
P i
n 4
size
c,
lrrm
,s b
etw
een
20
and
500
pm
.
RE
SUL
TS
TEP in
nat
ure
Ab
un
dan
ce, s
ize
and
ver
tica
l di
stri
buti
on
TE
P w
as f
ound
in d
ll na
tura
l se
awat
er s
ampl
es in
ves-
ti
gate
d. T
otal
nu
mb
er a
bu
nd
ance
an
d t
otal
are
a (e
sti-
m
ate
of
quan
tity
) of
TE
P,
how
ever
, va
ried
by
4 o
rder
s of
mag
nit
ud
e (1
0'
to 1
04 T
EP
ml-
' an
d 0
.2 to
200
0 m
m2
1-I
resp
ecti
vely
) (T
able
1).
Th
e le
ng
th o
f in
divi
dual
T
EP
ran
ged
from
3 t
o se
vera
l h
un
dre
d p
m. T
he
aver
- a
ge
dia
met
er o
f T
EP
ran
ged
fro
m 1
2 p
m a
t th
e b
egin
- ni
ng o
f th
e di
atom
blo
om t
o 1
63
pm
du
rin
g fl
occu
lati
on
of t
he b
loom
. A
t M
onte
rey
Bay
, th
e av
erag
e d
iam
eter
of
TE
P r
ang
ed b
etw
een
8 a
nd
13 p
m, e
xcep
t 27
Aug
ust
1992
at
10
m
(wh
ere
diat
oms
do
min
ated
th
e ph
yto-
pl
ankt
on)
wh
en it
was
15
pm
. A
t th
e B
AT
S st
atio
n th
e av
erag
e d
iam
eter
ran
ged
bet
wee
n 9
an
d 1
2 p
m, e
xcep
t at
900
m w
her
e th
e av
erag
e d
iam
eter
was
23
pm.
Exa
mpl
es o
f ve
rtic
al d
istr
ibut
ion
pat
tern
s of
TE
P a
nd
ce
lls
a1 se
vera
l sta
tion
s ar
e de
pict
ed i
n F
ig. 1. T
EP
con
- ce
ntra
tion
s w
ere
gene
rall
y h
igh
er i
n th
e tr
opho
geni
c
Mar, E
col. Prog. S
er. 113: 185-198, 1994
Table 1
. Abundance, lolal area an
d slope of size speclrd of T
EP (b
, cf. Eq
. 1) for some typical sam
ples. MB
: Monterey B
ay; SB
: Santa Ba
rha
ra: B
AT
S: Berm
uda. Da
ta from
diatom bloom
from A
lldredge et al. (19
93
)
Sam
ples
Environm
ent D
ominating plankton
TE
P A
bundance A
rea (no. ml")
(mm
2 l l)
24 Jul. 10 m
C
oastal SB
Clear
25 I
2.8
27 A
ug. 76 111
Slop
e MB
C
lear 500
50 3.0
26 Au
g. 5 m
Coastal M
B
Mixed
4926 500
3.0
26 Au
g. 400 m
BA
TS, 400 m
27 Au
g. 10 ~n
27 Au
g. 5 m
27 Au
g. 20 m
BA
TS. 300 m
B
AT
S, 800 m
BA
TS, 900 m
B
AT
S, 1400 m
Coastal M
B
Ocean
ic
Coastal M
B
Coastal M
B
Coastal M
B
Oceanic
Oceanic
Oceanic
Oceanic
Clear
Clear
Diatom
s D
inoflagellates D
inoflagellates + Ph
aeocystis
Clear
Clear
Diatom
s C
lear
19 Jun, 10 m
Coastal SB
D
iatom bloom
24 Jun. 10 m
C
oastal SB
D
iatom bloom
29 Jun, 10 m
C
oastal SB
Diatom
bloom
(plankton-rich) zon
e than
below (for ex
amp
le: Mon-
terey Bay, 26 A
ugust 1992; F
ig. la, b
]. TE
P concen-
trations decreased
parallel with concentrations of bac-
terio-, phyto- and
protozooplankton at 20 m. H
owever,
concentrations of T
EP
varied appreciably below th
e trophogenic zo
ne an
d a relative high ab
un
dan
ce wa
s o
bserv
ed at 300 m
. Diatom
s were rare at this station.
and
microplankton w
as dominated by dinoflagellates
and
dilates.
Ab
un
dan
ce and
size of T
EP
may also vary
within
the tio
ph
og
enic zone, ap
pearin
g in acco
rdan
ce with
chan
ges in th
e plankton composition (F
ig. {c, d
). In M
onterey Bay (2
7 A
ugust 1992), TE
P ab
un
dan
ce was
lowest at
10
m, th
e dep
th of
the diatom
maxim
um
(105 diatoms l"'),
and
highest at 50 m, th
e lower en
d of
the trophogenic zone. A
t 10 m th
e averag
e size of TE
P
was, how
ever, very large. Dinoflagellates w
are
most
abu
nd
ant at th
e surface (104 1-l) an
d d
ecreased con-
tinuously with d
epth
. A sm
all population of colonial
Phaeocystis w
as presen
t betw
een 1
0 an
d 20 n
l dep
th.
Bacteria concentrations d
ecreased continuously w
ith d
epth
. C
om
pared
to coastal stations, concentrations of TEP w
ere extremely low
at all dep
ths (300 to 1400 m
) of the
op
en ocean station (B
AT
S) n
ear Berm
uda (Fig. Ie, f).
At M
onterey Bay (2
6 A
ugust 1992), for exam
ple, ab
ou
t 2500 T
EP
ml-I w
ere observed at 400 in w
hereas only
2 TE
P m
l"' prevailed at 400 m in th
e op
en A
tlantic. T
otal quantity of TE
P [expressed as total area) was also
3 to 4 orders of mag
nitu
de low
er com
pared
to coastal w
aters. Above 800 m
and
below 1000 m
concentrations
were very uniform
, but a peak
in TEP num-bers w
as observed at 900 m
. Plankton concentrations w
ere low
(c
100 cells I-')
at all dep
ths, except at 900 m
. Several
species of m
ostly Chaetoceros w
ere observed at con- centrations of
1.6 X 105 cells 1"' at 900 m
.
Size distributions
Although
total ab
un
dan
ce of T
EP
was extrem
ely variable, the general sh
apes of th
e size frequ
ency
dis- tributions w
ere similar (F
ig. 2a). All but 2 of
the size
frequency spectra of T
EP
analyzed (n =
33) follow a
Pareto (pow
er-law) distribution (E
q. 1, F
ig. 2a 10 cl].
How
ever, both at Monterey B
ay (27 August) an
d at
BA
TS
the sh
apes (b-values] of the pow
er-law distribu-
tions at different dep
ths differ (F
ig. 2b, c). We classified
all frequency distributions into 2 main g
rou
ps and 1
transition g
rou
p according to their b-value. T
he 2 av
er- ag
e b-values representing the 2 m
ain groups are
sig- nificantly different (S
tudents' t, p < 0.001). E
xamples of
normalized
frequency distributions (dN
/N) a
re given
for each g
rou
p in F
ig. 3. In th
e first type of distribution (Gro
up
1 ) over 95 % of
TE
P occurred in th
e first 2 size classes (c2
5 p
m) and
more th
an 6
0%
were sm
aller than 10 pm. T
his gro
up
includes all sp
ectra fitting a pow
er-law relationship
with a value of
b b
etween
2.9
and
3.9
(averag
e b =
3.3 Â 0
,3, m
edian b = 3
.3). In this g
rou
p, th
e fit of the
regression lines to the data w
ds extremely tig
ht, indi-
cating a good fit to the pow
er-law distribution. A
ll size
Pti
sso
w &
Alld
recl
gc D
istr
ibut
ion
01 t
rans
pare
nt c
xop
olyn
ier
part
icle
s 189
26 A
ugus
t
a
TE
P a
rea
(mm
I"
')
0
300
600
L
EP (
m1
1
sow
8000
TE
P c
once
ntra
tion
26 A
ugus
t
* - I b Bac
teri
a (1
0'm
l "'
1
Phyt
opla
nkt
on (
rnil
>
TEP
are
a (mm ' F
' 1
0
500
,. .,
. . . .
. . , . "
Â¥
Â
....
diat
oms
27 A
ugus
t c
'EP
(m
ll>
5
00
0
-(
13
>8
<(
,
10000
27 A
ugus
t
Ph
yto
pla
nkt
on
(rn
l')
0
100
200
BA
TS s
tatio
n e
BOO
TE
P a
rea
(mm
/V
'
) 0
-3
.
0 5
L
-
TE
P (
rn
l)
8-
TEP
con
cem
raiio
n
BA
TS st
atio
n
f
Fig
. 1.
Ver
tica
l pr
ofile
s of
mic
ro-
and
bac
leri
opla
nkto
n an
d o
f T
EP
abu
nd
ance
an
d t
otal
are
a at
(a,
b)
a de
ep s
tati
on (
26
Au
gust
19
92)
and
(c.
d) a
sha
llow
er s
tati
on (
27
Aug
ust
1992
) in
Mon
tere
y B
ay, a
nd
(e,
f) B
AT
S M
arch
199
3. S
D o
f T
EP
ab
un
dan
ce o
ften
to
o sm
all t
o sh
ow
spec
tra
belo
ngin
g to
thi
s g
rou
p h
ave
in c
omm
on t
hat
they
wer
e ob
serv
ed a
t st
atio
ns w
her
e di
atom
co
nce
n-
trat
ions
wer
e ve
ry l
ow a
nd
nev
er d
omin
atin
g. D
ino-
fl
agel
late
s,
cili
ates
or
Pha
eocy
stis
w
ere
mod
erat
ely
abu
nd
ant
at s
ome
of
thes
e st
atio
ns,
wh
erea
s m
icro
- pl
ankt
on c
once
ntra
tion
s w
ere
extr
emel
y lo
w a
t ot
hers
. S
ampl
es
belo
ngin
g to
th
is
gro
up
(1
6) c
ame
from
co
asta
l S
anta
Bar
bara
, fro
m M
onte
rey
Bay
(F
ig. 2
a, b
) an
d f
rom
the
op
en o
cean
(Fig. 2c
). A
ll fr
eque
ncy
dist
ri-
buti
ons
fro
m s
ampl
es
coll
ecte
d du
ring
su
mm
er
in
Mon
tere
y B
ay (
Fig
. Za,
b)
exce
pt t
he 1
0 m
dep
th o
n 27
Aug
ust
1992
bel
ong
in t
his
grou
p. T
he
size
dis
trib
u-
tion
s ob
serv
ed a
t th
e B
AT
S s
tati
on b
etw
een
300
an
d
800
m c
lear
ly b
elon
g to
Gro
up
1 a
s w
ell
(b = 3
.0 to
3.6),
but
at g
reat
er d
epth
s th
e va
lues
of
b w
ere
too
low
to
qual
ify.
G
rou
p 2
inc
lude
s al
l sa
mp
les
(13)
wit
h a
val
ue
of
b b
etw
een
0.7
an
d 2
.2 (
aver
age
1.6
 0
.45,
med
ian
1.7)
. A
ltho
ugh
sign
ific
ant,
th
ese
regr
essi
ons
ten
ded
to
be
less
tig
ht a
nd
sta
nd
ard
dev
iati
ons
wer
e h
igh
er, i
ndic
at-
ing
devi
atio
ns f
rom
a s
impl
e po
wer
-law
dis
trib
utio
n.
Siz
e di
stri
buti
ons
of th
is t
yp
e w
ere
gene
rall
y o
bse
rved
a
t st
atio
ns
wh
ere
diat
oms
wer
e do
min
atin
g ph
yto-
pl
ankt
on.
Dis
trib
utio
ns c
olle
cted
du
rin
g t
he
earl
y fl
oc-
190 M
ar. Ecol. P
roq. Ser. 1 1 1
: 185- 198, 1994
TE
P (n
o. ml"
urn
") cin
g stag
e of a diato
m bloom
off San
ta Barb
ara (1
9 an
d 24 Ju
ne
1992; Fig
. 2d
) both b
elon
g to
Gro
up
2 (b=
1.6 an
d 2.1 resp
ectively
]. Sp
ectra of T
EP
at BA
TS
at 900 m (F
ig. 2c), w
he
re ab
un
- d
ance of
Ch
aetoceros sop
. was h
igh
, belo
ng
to th
is gro
up
as well (
b = 1.5). In sam
ples collected
at 2 dep
ths off S
anta B
arbara in
July
(tifter sedi-
men
tation
of a bloom) an
d at 1000 m
at B
AT
S,
TE
P ex
hib
ited a size d
istribu
tion
of ty
pe 2 (
b =
1
.4, 1
.5, 2
.2), alth
ou
gh
no
ph
yto
plan
kto
n
was
presen
t. S
ize distrib
utio
ns of T
EP
ob
served
in th
e p
res- ence of d
iatom
s did
not all belo
ng
into
Gro
up
2.
So
me sp
ectra exh
ibited
b-v
alues too lo
w to
qu
al- ify for G
rou
p 1
and
too hig
h for G
rou
p 2 (tran
si- tion g
rou
p). S
amp
les collected
at an early stag
e of
the d
iatom
bloom (b
= 2
.7) a
nd
the sam
ple
from
Mo
nterey
Bay fro
m
10 m
on
27 Au
gu
st 1992 (
b =
2.6; Fig.
2b
) wh
ere diato
ms
were
do
min
ating
a mixed su
mm
er dsscin
blag
e, hav
e sp
ectra of TEP b
elon
gin
g to th
e transitio
n g
rou
p.
In all sam
ples ex
hib
iting
spectra b
elon
gin
g to
the tran
sition
gro
up
diato
ms d
om
inated
ph
yto
- p
lank
ton
, excep
t a
t dep
ths below
1
00
0 m
at B
AT
S, w
here
no
p
hy
top
lank
ton
w
as p
resent
(Fig
, 2c). S
ize distrib
utio
ns of T
EP
ob
served
du
ring
the
late floccing stage of a d
iatom
bloom off
San
ta B
arbara (1
0 m
) had a differen
t sha
pe
(Fig
. 2d
). T
he size d
istribu
tion
s of TEP from
26 and
29 Ju
ne
19
92
, wh
en th
e bloom w
as fully flocced
, d
id n
ot follow
a po
wer-law
distrib
utio
n a
nd
did
not fit in eith
er gro
up
.
D@
BOO
20
0 - 6
00
0
1 150 -
40
0 -
40
00
- 100 -
200 -
20
00
50
-
0-
0
-
0
^ 1
26
August: G
rou
p 1
(M
B. Sml
l.. \ '-. \
- \
'. \
'. \
'. \
', 2
7 A
ugust: Gro
up
1
\ ,
(Me
, 7Çm
, 80
0m
Gro
up
1
1 -.*
Bacteria attach
ed to T
EP
Th
e nu
mb
er of b
acteria attached
to ind
ivid
ual
TE
P w
as no
t correlated
to the size of
the resp
ec- tive T
EP
(Fig
. 4a). M
ost T
EP
we
re sm
aller than
BOO
Fig. 2. S
ize distributions of T
EP
in the field. (a) Total
abundance (sum of
N) ranges from
25 to 5000 TE
P
ml. Three exam
ples from S
anta Barbara (S
B) 24 Ju
ly
at 10 m
(b = 2.8) and M
onterey Bay (M
B) near-shore
26 August at 5 m
(b =
3.0) and
offshore 27 August at
76 m (b
= 3.0
). (b, c) T
he shape of the size distribution
of T
EP m
ay vary with depth. E
xamples from
(b) M
on- terey B
ay (MB
) 27 August at 5
, 10 and 20 m (fa= 3
.0,2
.6
and 3.4 respectively) and (c) BA
TS at 300. 800, 900 and
1400 m (b
= 3.6, 3
.3, 2
.2 and 2.8 respectively). (d
) The
shape of the size distributions of TEP may change dur-
ing the flocculation of diatom
blooms. E
xample from
the S
anta Barbara channel during a C
haetoceros spp.
bloom at 10 m
depth on 19, 24 and 29 Jun
e (b =
1.6, 2
.1 respectively, 29 June does not follow
a power-law
distribution]
Jun
e: Gro
up
2
29
June /
no
gro
up
/
/
,
A' ..
"S
A-" N
T
EP
leng
th (urn)
Pas
sow
& A
Jldr
edcf
e- D
istr
ibut
ion
of tr
an
sp
are
nt
pxop
olym
cr p
drtic
las
L
W
I- -
- Q
26
Au
gu
st 1
0m
0
Q G
rou
p 1
-.
v 29 J
une
10
m '
no gro
up '
Fig
. 3.
Nor
mal
ized
siz
e-fr
eque
ncy
dist
ribu
tion
s of
TE
P. E
xam
ples
of
the
4 di
ffer
ent
type
s of
dist
ribu
tion
s ob
serv
ed.
Th
e di
stri
buti
on
from
26
Aug
ust
belo
ngs
to G
roup
1 (
non-
diat
oms)
, th
at f
rom
24
Jun
e to
Gro
up 2
(fl
occi
ng d
iato
ms)
, th
at f
rom
27
Aug
ust t
o th
e tr
ansi
tion
gro
up (
no
n-
floc
cing
dia
tom
s) a
nd t
hat
from
29
Jun
e do
es n
ot
fit i
nto
any
cate
gory
(fl
eece
d di
atom
s)
10
0
TE
P l
eng
th (u
rn)
50 u
rn a
nd
car
ried
les
s th
an
100
bac
teri
a,
but
up t
o 90
0 ba
cter
ia w
ere
obse
rved
per
pa
rtic
le. S
tain
able
thi
n sh
eets
of
TE
P, e
spe-
ci
ally
ver
y la
rge
on
es (
3 20
0 pm
), c
arri
ed
com
para
bly
low
ab
un
dan
ces
of ba
cter
ia.
How
ever
, th
e ba
cter
ial
abu
nd
ance
per
ar
ea o
f T
EP
is a
sig
nifi
cant
fun
ctio
n of
TE
P
size
(F
ig. 4b)
. Wh
erea
s sm
all T
EP
(c
10
pm
) h
ad 2
bac
teri
a at
tach
ed p
er p
n-i2
, les
s th
an
0.1
bact
eria
wer
e at
tach
ed p
er u
rn2
of T
EP
on
TEP >
100
pm. T
her
e w
as n
o ap
pre
cia-
bl
e di
ffer
ence
in b
acte
rial
den
sity
bet
wee
n
TEP
coll
ecte
d d
uri
ng
th
e di
atom
blo
om a
nd
T
EP
col
lect
ed a
t ot
her
tim
es.
Tot
al a
bu
nd
ance
of
bact
eria
in
a sa
mp
le
was
not
cor
rela
ted
to t
otal
qua
ntit
y of
TE
P (F
ig. 5
a).
Th
e f
ract
ion
of ba
cter
ia a
ttac
hed
to
TE
P (
Fig
. 5b
) gen
eral
ly r
ang
ed b
etw
een
2
and
25
% (
1 e
xcep
tion
on
27 A
ugus
t 19
92
at 5
0 m
: 8
9%
). N
o di
ffer
ence
in
atta
ched
ba
cter
ia
was
obs
erve
d b
etw
een
TE
P c
ol-
lect
ed d
urin
g th
e di
atom
blo
om s
tudy
and
T
EP
col
lect
ed a
t ot
her
tim
es.
NON-
diat
om
0
DIA
TO
M
TEP i
n c
ult
ure
s
Ab
un
dan
ce
TE
P f
ound
in
b
atch
cu
ltur
es o
f di
atom
sp
ecie
s ge
nera
lly
had
a
sim
ilar
ov
eral
l ap
pea
ran
ce t
o TEP f
ound
in
nat
ure
. H
ow-
ever
, ab
un
dan
ce a
nd
siz
e of
TE
P g
ener
ated
by
di
ffer
ent
spec
ies
diff
ered
app
reci
ably
(T
able
2)
. T
EP
o
bse
rved
in
cu
ltur
es of
Ch
ae
loce
ros
grac
ilis
wer
e ve
ry s
mal
l (3
to
5 pm
, av
erag
e 4 pm) a
nd
co
mm
a-sh
aped
TE
P le
ng
th (u
rn)
Fig
. 4
(a) T
otal
num
ber
abun
danc
e of
bac
teri
a pe
r in
divi
dual
TE
P v
ersu
s si
ze of
the
resp
ecti
ve TEP.
[b) B
acte
ria
dens
ity
per
area
TEP v
ersu
s th
e si
ze
of th
e re
spec
tive
TEP
Mar. E
col. Prog. Ser. 113; 185-1 98, 1994
0
No
n-d
iatom
typ
e
A
Diato
m-typ
e
0 N
on
-diatam
typ
e
A
Diato
m-typ
e
0
1 2
To
tal Area o
f TE
P
(mm
2 rnl"')
Fig. 5
. (a) Total bacteria num
ber per sample versus total quantity of
TEP. (b
] Fraction of bacteria attached to TE
P versus the total quantity of TE
P per sam
ple. Regression calculated om
itting the 2 values of the fleeced bloom
and the 89 %
value
du
ced 3500 T
EP
ml"
within th
e first 2 d
(early exp
on
enti~
il ph
ase), but no signi-
ficant ch
ang
e w
as observed
thereafter (F
ig. 6b
). Most
TEP
were larg
e (1
0 to
25 pm
) and
total quantity of TE
P w
as hig
h
com
pared
to C, gracilis. T
EP
gen
erated by
Nitzschia an
gu
laris belonged mostly to th
e sm
allest size
class, but
larger particles
were
abu
nd
ant
(Fig. 6c). TEP
from
N.
ang
ularis w
as circular or squ
are-shap
ed
and
an o
rder of
rndgnitude less abu
nd
ant
than
C. gracitis. T
he d
ecrease of total TE
P ab
un
dan
ce on
Day
5
(late exponential phase) w
as presumably a
n artefact, as N
. angularis form
s mats w
hich contain TE
P
on
flask walls. T
hese could not be sam
pled
quantitatively.
As
maxim
al cell
density an
d cell sizes of
these 3 species differ con- siderably (F
ig. 6), th
e specific quantity of T
EP
g
enerated
(total area
of T
EP
/total volum
e of cells) was calculated (T
able 2).
According to those calculations C
. gracills produced th
e most T
EP
per cell volum
e. C
om
pared
to the larg
e differences in num
- b
er abu
nd
anrn
of T
EP
betw
een species,
the specific quantity of T
EP
produced was,
however, rem
arkab
ly sim
ilar, N
o T
EP
was found in cultures of colonial
Phaeocystis ev
en
in stationary p
hase a
s long as colonies d
id not
disintegrate. In sen
escence w
hen
colonies beg
an to dis-
integ
rate, the fragm
ents of the disrupted
colonies stained
blue. A
lthough fo
rmed
differently from
diatom T
EP
, these parti-
cles may
be classified as T
EP an
d could
play an
important part for th
e formation of
marine
snow
after P
haeocystis bloom
s (P
assow &
Wassm
ann 19
94
).
bu
t slightly larger particles b
egan
to app
ear in late ex
po
nen
tial ph
ase on
Day 9 (F
ig. 6a). T
otal TE
P con- S
ize-frequency distributions centrations
were very high
(up
to 8000 T
EP
ml")
com
pared
to oth
er diatoms a
nd
increased du
ring
the
Most of
the size frequency spectra of
TE
P d
urin
g
wh
ole
growth
cycle. T
halassiosira
weissflogjj
pro- the
growth
cycle of
Chaetoceros
gracilis belong
Table 2. C
haracteristics of TE
P generated in cultures of diatom
s
Species M
aximal abundance
Maxim
al total area Specific production
(no, nil")
(mm2 mr
l) (mm2
TE
P nun-' cell '1
Ch
aetoceros gracilis T
halassiosira w
eissflogii N
itzschia an
gularis
"TE
P formed m
ats on flask walls m
aking quantification of T
EP
after early exponential phase impossible
Pass
ow &
Alk
lrec
lg~:
Dis
trib
utio
n of
tran
spar
ent e
xopo
lym
cr p
arti
cles
to G
rou
p 1.
Init
iall
y, a
t th
e b
egin
nin
g o
f th
e ex
po
nen
tial
ph
ase
(Day
3) t
he
val
ue
of b
was
ex
cept
iona
lly
hig
h (
b =
4.2
), d
ecre
ased
du
r-
ing
exp
onen
tial
gro
wth
(to
3.6
on D
ay g
), b
ut
incr
ease
d sl
ight
ly
at t
he
begi
nnin
g of
th
e st
atio
nary
ph
ase,
bef
ore
it d
ecre
ased
dra
sti-
ca
lly
(Day
1
3,
b =
2.4
) a
l la
te
stat
iona
ry
ph
ase
(Fig
. 7a)
. TEP p
rod
uce
d in
th
e cu
llur
e of
N
Uzs
chia
an
gula
ris
exhi
bite
d at
i!c
ii-Iy
ex
pone
nlia
l p
has
e a
dist
ribu
tion
bel
ongi
ng
10 C
I'I
IL
~~
2
(b
= 2
.1).
Siz
e sp
cdra
of
TE
P
gen
erat
ed b
y T
hala
ssio
sira
w
elss
flog
ii d
id
not
foll
ow a
pow
er-l
aw d
istr
ibut
ion.
-I \
\ la
te e
xpo
nen
tial
Flo
ccul
a ti
on e
xper
imen
t
i ,*,
, A
da
y 0
,
A
,,
0
,
Th
e ch
ang
e in
si
ze
spec
tra
of
TE
P o
bse
rved
du
rin
g a
flo
ccul
atio
n ex
peri
men
t w
ith
Cha
etoc
eros
gr
acil
is
Illu
stra
tes
how
ag
gre
gat
ion
pro
cess
es a
lter
th
e si
ze d
istr
ib-
utio
n of
TE
P (
Fig
. 7b). I
niti
ally
the
siz
e sp
ec-
tra
of
TE
P f
itte
d a
pow
er-l
aw d
istr
ibut
ion
wit
h a
b-va
lue
ol
Gro
up
1 (
b =
3.2)
. bu
t d
ur-
in
g th
e ag
gre
gat
ion
b-v
alue
s d
ecre
ased
an
d
[,he
fit
10
a po
wer
-law
fu
ncti
on b
egan
to
de
teri
orat
e (a
fter
4 a
nd 8
rni
n b
= 2
.3 a
nd
0.9
re
spec
tive
ly).
Alt
er 2
0 m
in r
otat
ion
the
size
di
stri
buti
on o
f T
EP
did
not
fit
a po
wer
-law
fu
ncti
on a
ny
mo
re a
nd
th
e h
igh
est
abu
n-
dan
ce o
f T
EP
was
fou
nd i
n t
he
thir
d si
ze
clas
s, a
t 10
0 pm
. DIS
CU
SSIO
N
Dis
trib
utio
n of
TEP
The
gen
eral
dis
trib
utio
n p
atte
rn o
f T
EP
in
the
oce
an w
as s
imil
ar to
th
at o
f ph
ytop
lank
- to
n: T
EP
was
fou
nd t
o be
mor
e ab
un
dan
t in
co
asta
l ar
eas
than
in
the
op
en o
cean
an
d
mor
e ab
un
dan
t in
the
trop
hoge
nic
zone
tha
n be
low
. T
he
rang
es o
f co
ncen
trat
ions
an
d
size
s of
T
EP
in
the
oce
an w
ere
sim
ilar
to
thos
e of
ph
ytop
lank
ton.
N
ever
thel
ess,
no
si
mpl
e ov
eral
l re
lati
onsh
ip b
etw
een
am
bi-
en
t ph
ytop
lank
ton
and
TE
P c
once
ntra
tion
w
as o
bser
ved.
Phy
topl
ankt
on c
ompo
siti
on,
how
ever
, ap
pea
red
to
be
an i
mpo
rtan
t fa
c-
tor.
Con
cent
rati
ons
of TEP t
end
ed t
,o b
e lo
w
wh
en
diat
oms
do
min
ated
ph
ytop
lank
ton,
b
ut
aver
age
size
an
d to
tal
area
of
TE
P w
ere
larg
e co
mp
ared
to
stat
ions
wh
ere
diat
oms
wer
e sc
arce
.
- /
We
exo
on
entl
al
b
da
y 0
A
h
r ,
l l
10
10
0
- - E 1
00
0
, ea
rly
exp
on
enti
al
3.
7
10
TE
P l
en
gth
(ur
n)
10
0
Fig
. 6
. Si
ze-f
requ
ency
dis
trib
utio
n [d
(N/l
og l
) vs
log
11
of TCP b
efor
e in
ocul
atio
n of
th
e m
edia
(D
ay 0
) an
d du
ring
ear
ly a
nd l
ate
expo
nent
ial
grow
th p
hase
of
the
3 di
atom
cul
ture
s. (
a) C
hae
toce
ros
grac
ilis.
Ear
ly:
Day
3 (
2 X
10
5 cel
ls r
nl")
; la
te: Day 9
(10
" cel
ls ml"'). (
b)
Tha
lass
iosi
ra
wei
ssfl
ogii
. Ear
ly: D
ay 2
(2
X 10
" ce
lls
ml"
'); la
te:
Day
5 (
7.7
X lo
4 cells
ml-l). (
c) N
itzs
chia
ang
ular
is. E
arly
: Day
2 (
7 X
103
cell
s m
l*')
; lat
e: D
ay 5
(2
X 10
'' ce
lls m
l-'l
194 M
ar. Eco
l. Prog. S
er. 113: 185-198, 1994
A
- - a
utions of com
ponent particles (Lam
bert et b
-va
lue
s
cells ml
al. 1981, McC
ave 19
84
). -
a U
nder the assum
ption of constant
pro-
4 -.
&.
'Â¥ 'Â
¥
duction of particles into the sm
allest size *
à - - +
- - Ã
-A'" \
class (to keep
the flux of
material through
- -
the spectrum constant) coagulation d
ue to
0 sh
ear and
loss of larg
e agg
regates d
ue to
3-
sedim
entation will create stead y-state size
spectra which fit a pow
er-ldw distribution
10
with b =
4 (Lerrnan 1979, H
un
t 1980, 1982,
n
McC
ave 1984). The effect of
aggregation
~s
l
on
size distributions dep
end
s, however, on
the m
echanisms (B
rownian m
otion, shear,
,
differential settling) governing coagulation lo5
(Hu
nt1
98
0,1
98
2).
4
10 12
As dissolution is faster for sm
aller than
for 1
2 6
a larger particles the slope of
the resulting In
ocu
lation
d
ay
s
size distribution would d
ecrease (smaller b
) if only dissolution occurred (L
erman
1979). 1
b i
0 o m
inu
tes
0
8 min
utes
A
20 m
inu
tes
c
- 1
0 1
Q.
-
"A
Fr,igm
entation or disaggregation
on the
other hand lead to a residual distribution w
ith ~i
steeper slope (b
larger, m
ore sm
aller particles) (L
erman 1
97
9). F
iltration rnecha- nism
s, like grazing of zooplankton, may not
chan
ge the sh
ape of
the curve (if they are
effective over the w
hole sixe rang
e), bu
t red
uce
the total
nu
mb
er of
particles (ch
ang
e in A
). It is not know
n which processes signifi-
cantly alter abu
nd
ance an
d distribution of
TE
P. T
he sh
apes of
size distributions may
indicate wh
ich processes
are of prim
ary
50
100
300
importance for T
EP
. In the following discus-
sion the size spectra of T
EP
observed in T
EP
length (urn) different
environments w
ill b
e analyzed,
Fig. 7. (a) C
han
ge of b
-valu
e du
ring
gro
wth
of Ch
aciocero
s gracilis. Cell
mak
ing
the foU
owina assum
ptions: co
ncen
tration
w
as determined daily.
(b) C
han
ge in
the sh
ape of
size (1) A
power-law
distribution with
b =
4 spectra of T
EP
du
ring
flocculation of C
. gracilis represents a theoretical steady-state situa- tion, w
here production of
small TEP is bal-
Th
eory
of size spectra an
d assum
ptions an
ced b
y loss of coagulated T
EP
. We w
iEl use this value as a reference point in our discussion.
Particle
size distributions
reflect interactions
be- (2
) For steady-state conditions to occur particles must
tween particles. M
arine particulate matter can g
ener-
be generated continuously into the low
end
of the size
ally be fitted by a power-law
distribution. Th
e shap
e of spectra. T
EP
is formed from
dissolved excretion prod- size distributions of
rndrine susp
end
ed particles b
e- ucts of
phytoplankton via coagulation of m
icrofibrils tw
een l to l00
pm
, or even
up to l00
0 pm, ap
pear to be
(Alldredge et
al. 1993, Passow
et
al. 1994). New
ly essentially constant as reflected in a constant value of
formed T
EP
will accordingly b
e introduced into th
e th
e logarithmic slope of th
e size distributions (b
= 4
) low
end
of th
e size spectrum.
(Sheldon et al. 1
97
2, L
al & L
erman 1
97
5, L
erman et al.
(3) A
s TE
P are relatively larg
e but do
not sink app
re- 1977, H
un
t 1980, 1982, M
cCave
1984). Reasons
tor ciably (u
np
ub
l. data), it is assum
ed thai coagulation of such a constant size distribution of m
arine su
spen
ded
T
EP
is dominated by sh
ear. m
atter are, however, controversial an
d it has been su
g-
(4) Furtherm
ore, it
will
be
assumed
that T
EP
is g
ested that
distributions of individual particle
types prim
arily g
enerated
by phytoplankton an
d that pro-
may not follow
a po
wer-law
distributions, but that the
duction of TE
P ceases in th
e absen
ce of phytoplankton. overall distributions are m
ade u
p of log-norm
al distrib- P
resumably T
EP
is also gen
erated by excretion prod-
Pass
ow &
Alld
redg
e: D
istri
butio
n of
tran
spar
ent e
xopo
lym
er p
artic
les
ucts
01
bac
teri
a, b
ut n
o re
lati
onsh
ip b
etw
een
bac
teri
a an
d T
EP
ab
un
dan
ce w
as o
bse
rved
. (5
) It i
s as
sum
ed th
at T
EP
gen
erat
ed b
y d
iffe
ren
t sp
e-
cies
hav
e di
ffer
ent
prop
erti
es.
TE
P d
iffe
red
cons
ider
- ab
ly i
n si
ze,
shap
e an
d a
bu
nd
ance
bet
wee
n s
peci
es.
Th
e ro
le o
f T
EP
du
rin
g a
gg
reg
atio
n a
lso
diff
ers
for
clif-
fe
ren
t sp
ecie
s (d
oc
ke
r &
Pas
sow
in p
ress
). T
he
stic
ki-
ness
of
TE
P p
rod
uce
d b
y C
haet
ocer
os g
raci
lis
is h
igh
(P
asso
w e
t al
. 19
94).
wh
erea
s T
EP
of
Nit
ztsc
hia
angu
- la
ns
and
Tha
lass
iosi
ra w
eiss
flog
ii d
o no
t ap
pea
r to
be
stic
ky (
Log
an e
t al
. 1994).
Kie
rboe
& H
anse
n (
1993
) fo
und
TE
P f
rom
Rho
dom
onas
hal
tica
an
d S
keie
ton
emfi
co
stat
um n
ot t
o b
e st
icky
, w
her
eas
Cos
cino
disc
us s
p.
gen
erat
ed T
EP
that
co
agu
late
d e
asil
y,
Siz
e sp
ectr
a of
TEP
In c
ultu
re
Siz
e sp
ectr
a o
bse
rved
in
cu
ltu
res
may
hel
p i
nte
rpre
t fi
eld
dat
a, a
s ba
tch
cult
ures
are
rel
ativ
ely
sim
ple
sys-
te
ms
(no
loss
fro
m s
edim
enta
tio
n, c
on
sum
pti
on
). T
otal
qu
anti
ty a
nd
ab
un
dan
ce o
f T
EP
in
crea
sed
du
rin
g t
he
grow
th c
ycle
of
Cha
etoc
eros
gra
cjli
s an
d t
he
syst
em
was
far
from
ste
ady
sta
te. C
han
ges
in
the
size
spe
ctra
of
TE
P d
uri
ng
gro
wth
of
C. g
raci
lis
may
be
expl
aine
d by
the
int
erac
tion
bet
wee
n p
rodu
ctio
n an
d a
gg
reg
a-
tion
of TEP. In
itia
lly
the
b-va
lue
was
ver
y hi
gh (
4.2
) an
d c
lose
to
the
stea
dy
-sta
te e
quil
ibri
um p
ost
ula
ted
if
intr
oduc
tion
of
new
(s
mal
l) p
arti
cles
is
bala
nced
by
co
agul
atio
n an
d l
oss
of la
rge
agg
reg
ates
(H
un
t 1
98
0,
1982
). D
urin
g th
e ex
po
nen
tial
gro
wth
ph
ase
the b-
valu
e of
the
spec
tra
of T
EP
dec
reas
ed c
onti
nuou
sly
(to
3.6
on
Day
g),
sug
ges
tin
g t
hat
coag
ulat
ion
was
sli
ghtl
y fa
ster
tha
n pr
oduc
tion
of
TE
P, A
t th
e on
set
of th
e st
a-
tion
ary
ph
ase
the
form
atio
n ra
te
of T
EP
in
crea
sed
ap
prec
iabl
y, a
pp
aren
tly
now
do
min
atin
g,
as t
he p
er-
cen
tag
e of
sm
all p
arti
cles
in
crea
sed
. Alt
houg
h pr
oduc
- tio
n ra
te
of T
EP
st
ayed
hig
h
duri
ng
the
stat
iona
ry
ph
ase,
th
e b-
valu
e d
rop
ped
dra
stic
ally
, in
dica
ting
tha
t ag
gre
gat
ion
d
om
inat
ed
inte
ract
ions
of
TE
P
at
this
ti
me.
In
the
floc
cula
tor
high
sh
ear
rate
s m
ade
coag
u-
lati
on o
f TEP i
mp
ort
ant
on t
ime
scal
es t
oo s
hort
for
pr
oduc
tion
of
TEP t
o b
e of
sign
ific
ance
. T
he
init
ial
dec
reas
e in
b-v
alu
es a
s fl
occu
lati
on b
egan
an
d t
he
sub
seq
uen
t de
viat
ion
from
a p
ower
-law
dis
trib
utio
n d
uri
ng
th
e fl
occi
ng p
has
e d
escr
ibe
this
ag
gre
gat
ion
do
min
ated
sys
tem
.
In t
he
pre
sen
ce o
f di
atom
s
In t
he
fiel
d, s
ize
spec
tra
01 T
EP
ob
serv
ed in
th
e p
res-
en
ce o
f di
atom
s b
elo
ng
ed e
ith
er to
Gro
up
2, t
he
tran
si-
tion
gro
up
or
did
not
foll
ow a
po
wer
-law
dis
trib
utio
n.
Th
e se
qu
ence
of
spec
tra
of T
EP
col
lect
ed d
uri
ng
th
e fl
occi
ng b
loom
evo
lved
sim
ilar
ly t
o th
ose
obse
rved
fo
r C
ha
eto
cero
s gr
acil
is i
n th
e fl
occu
lato
r, s
ug
ges
tin
g t
hat
co
agul
atio
n of
TE
P d
om
inat
ed p
roce
sses
alt
erin
g s
ize
dist
ribu
tion
s of
TE
P i
n th
ese
sam
ple
s. T
he i
nclu
sion
of
a sp
ecif
ic s
pec
tra
in o
ne
of t
he
gro
up
s al
so a
pp
eare
d to
co
rres
po
nd
to t
he
deg
ree
of fl
occu
lati
on o
f th
e d
iato
ms.
S
ize
dist
ribu
tion
s of
T
EP
bel
ongi
ng t
o th
e tr
ansi
tion
g
rou
p w
ere
ob
serv
ed w
hen
dia
tom
s w
ere
not
floc
cing
. T
he
size
dis
trib
utio
ns o
f T
EP
did
not
fol
low
po
wer
-law
di
stri
buti
ons
wh
en
a bl
oom
w
as
full
y fl
occ
ed.
Siz
e sp
ectr
a be
long
ing
to G
rou
p 2
in
dic
ated
th
at m
acro
- sc
opic
ally
vis
ible
flo
es w
ould
for
m w
ithi
n d
ays.
Thi
s im
plie
s th
at m
icro
scop
ical
ly v
isib
le c
oagu
lati
on o
f T
EP
p
rece
eds
the
mac
rosc
opic
al f
locc
ulat
ion
of a
bloo
m,
sug
ges
tin
g t
hat
siz
e sp
ectr
a of
TE
P m
ay h
ave
pre
dic
- ti
ve v
alu
e.
Dia
tom
ass
emb
lag
es w
ere
mos
tly
Ch
aet
oce
ros
spp
. d
om
inat
ed b
loom
-dia
tom
s. O
ur
exp
erim
ents
as
wel
l as
d
ata
from
oth
ers
(Kia
rboe
& H
anse
n 1
993,
Kia
rboe
et
al.
1994
) in
dic
ate
that
TE
P m
ay b
e le
ss i
mp
ort
ant
dur-
in
g f
locc
ulat
ion
of o
ther
dia
tom
sp
ecie
s.
A
size
sp
ectr
um
nea
r eq
uili
briu
m
cond
itio
ns (
be-
lo
ng
ing
to
Gro
up
1)
was
nev
er o
bser
ved
in t
he
pre
s-
ence
of
diat
oms
in n
atu
re. P
roce
sses
lik
e di
ssol
utio
n or
se
lect
ive
cons
umpt
ion
of sm
alle
r T
EP
may
hav
e ca
use
d
the
ob
serv
ed d
evia
tion
s fr
om a
ste
ady
-sta
te s
itu
atio
n.
How
ever
, as
eff
ects
of
diss
olut
ion
wer
e no
t vi
sibl
e in
cu
ltu
res,
we
do
not
ass
um
e th
is to
be
a pr
oces
s of
maj
or
imp
ort
ance
, S
ize
dist
ribu
tion
s w
ith
b-va
lues
bel
ow 4
w
ould
als
o b
e ex
pec
ted
, if
TE
P w
ere
frac
tal.
In
fact
, u
nd
er t
he
assu
mp
tio
n of
ste
ady
sta
te t
he
frac
tal d
imen
- si
on o
f ag
gre
gat
ed p
arti
cles
may
be
calc
ula
ted
fro
m
b-va
lues
(Ji
ang
& L
ogan
199
1, L
ogan
& K
ilps
199
4). A
s ag
gre
gat
es f
orm
ed i
n m
arin
e sy
stem
s g
ener
ally
hav
e fr
acta
l p
rop
erti
es (
Log
an &
Wil
kins
on 1
990,
Kil
ps e
t al
. 19
94),
we
sug
ges
t th
at t
he
low
b-v
alue
s d
uri
ng
gro
w-
ing
(no
n-fl
occi
ng)
diat
om b
loom
s m
ay r
efle
ct i
n p
art
the
frac
tal
nat
ure
of
TE
P a
gg
reg
ates
.
In t
he
pre
sen
ce o
f no
n-di
atom
phy
topl
ankt
on
Siz
e sp
ectr
a of
TE
P i
n th
e p
rese
nce
of
no
n-d
iato
m
phyt
opla
nkto
n b
elo
ng
ed t
o G
rou
p 1
. Th
ese
dis
trib
u-
tion
s m
ay
rep
rese
nt
cond
itio
ns
nea
r st
ead
y
stat
e,
wh
ere
coag
ulat
ion
of T
EP
is b
alan
ced
by
prod
ucti
on o
f T
EP
. L
arg
e ag
gre
gat
es w
ere
nev
er o
bse
rved
at
thes
e st
atio
ns,
mak
ing
an
ap
pre
ciab
le l
oss
of T
EP
du
e to
si
nkin
g un
like
ly.
If lo
ss w
as l
ow,
both
pro
du
ctio
n a
nd
co
agul
atio
n of
TE
P m
ust
hav
e b
een
low
. Siz
e sp
ectr
a of
TE
P ob
serv
ed i
n th
e p
rese
nce
of
no
n-d
iato
m p
hyto
- p
lan
kto
n n
ever
bel
on
ged
to
Gro
up
2 o
r th
e tr
ansi
tion
g
rou
p,
whi
ch
also
in
dic
ates
th
at
coag
ula
tio
n n
ever
Mar. E
col. Prog. Ser. 113: 185-198, 1994
dominated size distributions of T
EP
in these sam
ples. L
ow
coagulation rates
sug
gest that
TE
P
from n
on
- diatom
phytoplankton may be less sticky than diatom
- T
EP
. Th
e hig
her
b-values may furtherm
ore indicate that non-diatom
TEP is less fractal than TE
P g
enerated
by diatom
s. In the ab
sence of
phytoplankton
Th
e size
distributions of
TE
P
in th
e ab
sence
of phytoplankton w
ere not predictable and
presumably
dep
end
on
the phytoplankton assem
blag
e which g
en-
erated T
EP prior to sam
plin
g. In July, after the sed
i- m
entation of th
e diatom bloom
, the size distribution of
TE
P belonged to G
roup 2 and
probably reflected the
flocced bloom
, alth
ou
gh
phytoplankton
(and
m
ost T
EP
) had
sedirnented. A TEP size spectra belonging to
Group 2 w
as also observed at 1000 m at B
AT
S, below
th
e layer abu
nd
ant in C
haetoceros spp
. Size distribu-
tions of T
EP
from
below the euphotic zone (oceanic
and
coastal) gen
erally belonged to G
roup 1. A
s TE
P
are presumably
not g
enerated
below
the
eup
ho
tic zone, an
d as th
e sinking velocity of individual T
EP
m
ust be very small, it m
ay be concluded that processes ch
ang
ing
distributions of non-diatom
TE
P are slow
.
Co
nclu
sion
s from analysis of size sp
ectra
Our
field d
ata su
gg
ests that
TE
P g
enerated
by
diatoms
differ from
non-diatom
T
EP
. D
iatom
TE
P
app
ear to be more fractal an
d exhibit a h
igh
er slicki- ness. T
EP
pro
du
ced by different diatom
species also
differ. Th
e size spectra of T
EP
gen
erated by diatom
s furtherm
ore differ dep
end
ing
on the d
egree of fioccu-
lation of the diatoms. P
ossibly, the spectra of TE
P g
en-
erated by diatom
s may be u
sed to m
ake predictions
abo
ut th
e deg
ree of flocculation of cells, as th
e coagu- lation of
TE
P ap
pears to p
recede th
e macroscopically
visible flocculation
of the
bloom.
Of
all processes
potentially altering size distributions of TE
P, only coag-
ulation of T
EP
gen
erated by diatom
s was conspicuous.
All other processes seem
ed com
parably slow an
d size
distributions of non-diatom
TE
P ap
peared
to remain
un
chan
ged
over long periods, thus occurring even at g
reat dep
ths. B
acteria associated with T
EP
Th
e abu
nd
ance of
bacteria attached
to TE
P w
as not directly related
to size of TE
P. B
acteria density on TE
P,
however, d
ecreased w
ith increasing size of TE
P. A
sim-
ilar decrease in specific bacterial density w
ith agg
re-
gate size (bacteria per ag
greg
ate weight vs ag
greg
ate size) occurs on m
arine snow (A
lldredge & G
otschalk 1
99
0). T
he d
ecrease of bacterial density with T
EP size
may
be an
artefact of m
easuring TE
P collapsed on fil-
ters (in 2 dimensions). T
EP
are presumably 3-dim
en- sional
and
often
fractal w
hen
su
spen
ded
. B
acteria them
selves may also influence the size of
TE
P, either by degradation or by generation of
alcian blu
e stain-
able m
ucous. Bacteria attaching to surfaces h
ave been
found to excrete extracellular polysaccharides abu
n-
dantly (D
erho 1990, C
owen
1992, Vandevivere
&
Kirchrnan
1993). Th
e high
variability of
bacterial attachm
ent within one size class m
ay be caused by dif-
ferences in deg
ree of colonization, ag
e or composition
of TE
P.
Th
e relationship b
etween
total nu
mb
er of bacteria
and
total quantity of TE
P d
oes not indicate th
at bacte- ria ab
un
dan
ce is increased by th
e presen
ce of T
EP
. T
he larg
e quantity of TE
P during the floccing diatom
bloom w
as not paralleled by a respective increase in
bacteria for exam
ple. F
rom this it m
ay be concluded that T
EP
is not always a preferred en
ergy
source pro- m
oting growth of bacteria an
d no sim
ple relationship exists b
etween
TE
P and bacteria ab
un
dan
ce. Th
e per- cen
tage
of bactcrid
associated w
ith T
EP
, how
ever, ap
pear to d
epen
d
within
certain restrictions on th
e total area of
TE
P. T
his could indicate that
bacteria colonize TEP according to the surface area available, rather than extensively grow
on TE
P.
With one exception, less th
an 2
5%
of all bacteria
were associated w
ith TE
P. Th
ese values are lower than
those reported by us earlier [Alldredge et al. 1993, P
as- sow
et al. 19
94
). This difference is partially d
ue to use
of a different, possibly mo
re accurate, way of
calculat- ing
these p
ercentag
es. Values given in
the present
pap
er should be viewed as low
er estimates, w
hereas
earlier estimates rep
resented
up
per lim
its. In
marine
pelagic environm
ents attached
bacteria arc generally reported to account for less than 5
%, an
d
rarely more than 1
0%
, of the total bacteria assemb
lage
(Hodson et al. 1981, L
inley & F
ield 1982). T
EP
con- tributes appreciably
to m
arine snow
(m
acro-ag
gre-
gates), but as m
arine snow ag
greg
ates are generally rare,
less th
an
5%
of all
bacteria are
attached
to
marine
snow
(Alldredge
et al.
1986, Alldredge
&
Gotschalk
1990). Higher p
ercentag
es (10
to 20
%) of
attached
bacteria were only found in river plum
es or estuaries w
here allochtonous detrital m
aterial is abu
n-
dan
t and turbidity high (Ducklow
& K
irchman
1983, P
ainchaud & T
herriault 1989). W
e believe that
previous studies w
hich h
ave not
considered TE
P-attached bacteria
may
have u
nd
er- estim
ated the percen
tage of
attached
forms. B
acteria attach
ed to T
EP
may have to be ad
ded
to numbers of
attached
bacteria
derived by
most
meth
od
s. Many
Pas
sow
& A
lld
red
ge;
Dis
trib
utio
n of
tran
spar
ent
exop
olym
er p
artic
les
197
met
hods
rel
y on
siz
.e s
epar
atio
n w
ith
filt
ers
(Aza
m &
H
odso
n 19
77).
Pre
lim
inar
y te
sts
show
ed t
hat
TE
P i
s ex
trem
ely
flex
ible
an
d t
hat
TE
P s
ever
al t
ens
of p
m i
n si
ze c
an b
e pa
ssed
thr
ough
a 1
1i1n
filt
er.
Und
er t
he
mic
rosc
ope,
bac
teri
a at
tach
ed t
o T
EP
wou
ld b
e ic
lent
i-
fied
as
free
bac
teri
a w
itho
ut o
ur d
oubl
e-st
aini
ng t
ech-
ni
que.
W
itho
ut
stai
ning
on
ly
TE
P
whi
ch
are
very
de
nsel
y co
loni
zed,
an
d
lyin
g in
ba
cter
ia-p
oor
snr-
ro
undi
ngs,
will
be
reco
gniz
ed a
s pa
rtic
les.
In
a de
tail
ed
mic
rosc
opic
al
inve
stig
atio
n L
inle
y &
F
ield
(1
982)
de
scri
be
the
diff
eren
t ki
nd o
f ba
cter
ial
asso
ciat
ions
th
ey f
ound
on
the
wes
t co
ast
off
Sou
th A
fric
a. A
bout
8
5%
of
all
mic
ro-a
ggre
gati
ons
they
ob
serv
ed
are
desc
ribe
d as
'gel
atin
ous
agg
reg
ates
' whi
ch, a
ccor
ding
to
the
des
crip
tion
giv
en,
may
be
iden
tica
l w
ith
larg
e an
d h
eavi
ly c
olon
ized
TE
P. I
n th
eir
stu
dy
, abo
ut 5
% o
f al
l ba
cter
ia w
ere
atta
ched
an
d 4
% w
ere
atta
ched
to
the
'gel
atin
ous
agg
reg
ates
'. B
ecau
se
of
the
tran
s-
pare
ncy
of TEP, a
larg
e fr
acti
on o
f ba
cter
ia a
ssoc
iate
d w
ith
TE
P,
espe
cial
ly o
n la
rge
web
s, m
ay h
ave
bee
n
over
look
ed e
ven
in t
his
care
ful
inve
stig
atio
n.
Impa
ct o
f T
EP on e
colo
gy
Th
e va
riab
ilit
y in
ab
un
dan
ce, s
ize
and
type
of
TE
P in
m
arin
e sy
stem
s w
ill
resu
lt i
n a
high
var
iabi
lity
in
the
re
lati
ve s
igni
fica
nce
of T
EP
for
ag
gre
gat
ion
, as
a fo
od
sour
ce f
or z
oopl
ankt
on,
in t
he
carb
on c
ycle
, fo
r th
e m
icro
bial
co
mm
unit
y an
d
for
chem
ical
ad
sorp
tion
pr
oces
ses.
A
ll di
atom
s w
e in
vest
igat
ed
gen
erat
ed T
EP
, an
d
Ske
leto
nem
a c
osla
tiim
, C
haet
ocer
os a
ffin
is a
nd
Cos
- ci
nodi
scus
sp
. als
o p
rod
uce
muc
ous
part
icle
s, p
resu
m-
ably
TEP (
Kim
boe
& H
anse
n 1
993)
. Het
eroc
apsa
pyg
- mia
(din
ofla
gell
ate)
for
ms
som
e T
EP
(u
np
ub
l. d
ata)
an
d T
EP
or
TE
P-l
ike
muc
ous
part
icle
s ar
e al
so f
orm
ed
by
Rho
dom
onas
ba
ltica
(K
iorb
oe &
H
anse
n
19
93
).
Col
onia
l P
haeo
cyst
is,
how
ever
, do
not
gen
erat
e TEP
exce
pt w
hen
col
onie
s ar
e di
sint
egra
ting
(P
asso
w &
W
assm
ann
19
94
), I
t th
us a
pp
ears
th
at m
ost
phyt
o-
plan
kton
sp
ecie
s g
ener
ate
TE
P,
and
T
EP
w
ere
obse
rved
at
all
stat
ions
in
nat
ura
l w
ater
s, e
ven
in
the
abse
nce
of
abu
nd
ant
phyt
opla
nkto
n.
Th
e ro
le TEP p
lay
in t
he
form
atio
n of
mar
ine
snow
, ho
wev
er,
vari
es.
Th
e im
port
ance
of
TE
P f
or a
gg
reg
a-
lion
of
som
e di
atom
sp
ecie
s h
as b
een
sho
wn
(Pas
sow
et
al.
19
94
). O
ther
dia
tom
spe
cies
, ho
wev
er,
like
Ske
le-
tone
ma
cost
atu
rn (
Kie
rboe
et
al.
1994
) or
Nit
zsch
ia
angi
ilar
is (
Cro
cker
& P
asso
w i
n pr
ess)
for
m f
loes
wit
h li
ttle
TE
P.
Th
e d
ata
pre
sen
ted
her
e in
dica
te t
hat
TE
P g
ener
ated
by
non
-dia
tom
spe
cies
may
not
con
trib
ute
sign
ific
antl
y to
the
for
mat
ion
of m
arin
e sn
ow.
TE
P, e
spec
iall
y w
hen
cove
red
by b
acte
ria,
may
be
valu
able
foo
d fo
r pr
otoz
oo-
and
zoo
plan
kton
. H
ow-
ever
, ver
y la
rge,
sti
cky
web
s of
TE
P s
uch
as
obse
rved
d
uri
ng
fle
ecin
g bl
oonl
s ar
e m
ore
lik
ely
to i
nhib
it f
eed
- in
g by
cl
oggi
ng
feed
ing
str
uct
ure
s. S
mal
l TE
P, a
s o
bse
rved
mor
e co
mm
only
in
the
abse
nce
of
diat
oms,
o
n t
he
othe
r h
and
, m
ay b
e to
o sm
all
for
man
y f
ilte
r fe
eder
s.
Th
e to
tal
area
of
TE
P w
as r
elat
ivel
y sm
all
at s
tati
ons
wh
ere
diat
oms
wer
e ra
re, a
lth
ou
gh
ab
un
dan
ces
of TEP
wer
e ve
ry h
igh.
Thi
s w
ould
su
gg
est
that
in t
erm
s of
th
e cy
clin
g of
car
bon
TE
P m
ay b
e ap
prec
iabl
y m
ore
iinp
or-
tant
in
conn
ecti
on w
ith
diat
om b
loom
s th
an a
t ot
her
tim
es.
By
serv
ing
as a
sub
stra
te a
nd
pre
sun
~ab
ly an
en
erg
y
sou
rce
for
bact
eria
, TE
P m
ay a
lter
dis
trib
utio
n p
atte
rns
of b
acte
ria
and
the
ir a
vail
abil
ity
as a
foo
d so
urce
. Th
e d
egre
e to
whi
ch T
EP
may
be
util
ized
by
the
mic
robi
al
com
mun
ity
wil
l la
rgel
y d
epen
d u
po
n
the
chem
ical
co
mpo
siti
on o
f T
EP
. A
s ex
cret
ion
prod
ucts
of
phyt
o-
plan
kton
var
y w
ith
spec
ies
(Dec
ho 1
990)
, it
is a
ssu
med
th
at t
he
com
posi
tion
of
TE
P m
ay a
lso
vary
an
d th
at t
he
high
var
iabi
lity
of
ba
cter
ial
colo
niza
tion
of
TE
P m
ay
refl
ect
this
.
Ack
no
wle
dg
emen
ts.
Th
e p
roje
ct
was
fu
nd
ed
by
ON
R
(N00
014-
89-J
3206
). W
e th
ank
M
ark
BrZ
eZin
Ski
an
d
An
ne
Clo
se
for
pro
vid
ing
th
e sa
mp
les
from
B
AT
S an
d
Ch
ris
Go
tsch
alk
tor
hel
p w
ith f
igu
res.
We
furt
her
mo
re t
han
k B
ruce
L
og
an t
or v
alu
able
co
mm
ents
on
an
ear
lier
ver
sion
of
th
is
pap
er,
LIT
ER
AT
UR
E C
ITE
D
All
dre
dg
e, A
. L
., C
ole,
J. J
., C
aro
n, D
, A.
(19
86
). P
rodu
ctio
n of
h
eter
otr
op
hic
b
acte
ria
inh
abit
ing
m
acro
scop
ic
org
anic
ag
gre
gat
es (
mar
ine
sno
w)
from
su
rfac
e w
ater
s. L
imno
l.
Oce
ano
gr.
31:
68-
78
All
dre
dg
e, A
. L., G
ots
chal
k,
C.
C:.
(199
0). T
he
rela
tiv
e co
ntr
i-
buti
on o
f m
arin
e sn
ow
of
dif
fere
nt
ori
gin
s to
bio
logi
cal
pro
cess
es i
n c
oas
tal
wat
ers.
Co
nt.
She
lf Res.
10: 4
1-58
A
lld
red
ge,
A.
L.,
Pas
sow
, U
,, L
oq
an, B
. (1
99
3).
Th
e ex
iste
nce
, ab
un
dan
ce, a
nd
sig
nifi
canc
e of
lar
ge
tran
spar
ent
exo
po
ly-
mer
par
ticl
es i
n th
e o
cean
. D
eep
Sea
Res
. I 4
0:
1131
-1 1
40
Aza
rn,
F..
Ho
dso
n,
R.
E.
(19
77
). S
ize
dis
trib
uti
on
an
d a
ctiv
ity
of
mar
ine
mic
roh
eter
otr
op
hs.
L
imno
l.
Oce
ano
gr.
22
: 49
2-50
1 B
ader
, H
. (1
970)
. T
he
hy
per
bo
lic
dis
trib
uti
on
of
par
ticl
es.
J. g
eop
hy
s. R
es.
75:
2822
-283
0 C
ow
en, J
. P.
(1
99
2).
Mor
phol
ogic
al s
tud
y o
f m
arin
e b
acte
rial
ca
psu
les:
im
plic
atio
ns f
or m
arin
e ag
gre
gat
es.
Mar
. B
iol.
114:
85
-95
C
rock
et,
K.,
Pas
sow
, U
. (in
pre
ss).
Dif
fere
ntia
l ag
gre
gat
ion
of
dia
tom
s. M
ar. E
col.
Pro
g. S
er.
Dec
ho
, A
. W
. (1
99
0).
Mic
robi
al
exo
po
lym
er
secr
etio
ns
in
oce
an e
nv
iro
nm
ents
: th
eir
role
in
food
web
s an
d m
arin
e p
roce
sses
. O
cean
og
r. m
ar. B
iol.
A. R
ev.
28:
73-
153
Duc
klow
, H
. W
., K
irch
man
, D
. L
. (1
983)
. Bac
teri
al d
yn
amic
s an
d d
istr
ibut
ion
du
rin
g s
pri
ng
dia
tom
blo
om i
n H
ud
son
R
iver
plu
me,
US
A. J
. P
lan
kto
n R
es. 5
: 33
3-35
5 H
ewes
, C. D
., H
olm
-Han
sen
, 0. (
1983
). A
met
ho
d f
or r
eco
ver
- in
g na
nopl
ankt
.on
from
fil
ters
for
id
enti
fica
tio
n w
ith
th
e
Mar. E
col. Prog. S
er, 113: 185-198, 1994
microscope:
the
filter-transfer-freeze (F
TF
) tech
niq
ue.
Lim
nol. Ocean
og
r. 28: 389-394 H
ewes, C
. D., R
eid, F. M. H
., Ho
lm-H
ansen
, 0. (1984). T
he
qu
antitativ
e analysis of nan
op
lank
ton
: a study of methods.
J. Plankton R
es. 6: 601-613 H
obbie, J. E., D
aley, R. J., Jasp
er, S. (1977). U
se of Nuclepore
filters for counting bacteria by fluorescence microscopy.
Appl. environ. M
icrobiol. 33: 1225-1228 H
odson, R. E., M
accubbin, A. E
., Pom
eroy, L. R
. (1981). Dis-
solved aden
osin
e triph
osp
hare utilization
by free-living an
d attach
ed bacterioplankton. M
ar. Biol. 64: 43-51
Horobin, R
. W. (1988). U
nderstandi.ng histochemistry: selec-
tion, evaluation and
desig
n o
f biological stains. Biochem
- istry an
d biotechnology. J. W
iley and
Sons, N
ew Y
ork 1-iunt, J. R
. (1980). Prediction of oceanic particle sizedistribu-
lions from coagulation an
d sed
imen
tation
mechanm
isrns. In:
Kavanaugh, M
. C., L
eckie, S. (ed
s.) Particulates in
water -
characterization, fate, effects and
removal. A
dv. C
hemistry 189: 243-257
Hunt,
J . R.
(1982). Self -similar
particle-size disiributions
du
ring
coagulation: theory and
exp
erimen
tal verification. J. F
luid Mech. 122: l(i9-185
Jiang
, Q., L
ogan, B. E
. (1991). Fractal dim
ensions of ag
gre-
gates
determ
ined
fro
m stead
y-state
size distributions.
Environ. S
ci. Tech. 25: 2031-2038
Kilps, J. R
., L
ogan, B. E
., AU
dredge, A
. L.
(1994) Fractal
dimensions of m
arine snow
agg
regates d
etermin
ed from
im
age analysis of
in situ ph
oto
grap
hs. D
eep S
ea Res, (in
press)
King. L. K.. P
arker, B. C
. (1988). A sim
ple, rapid meth
od
for en
um
erating
total viable and
metabohcally active bacteria
in gro
un
dw
ater. Appl. environ. M
icrobiol. 54: 1630-1631 K
iarboe, T., H
ansen
, J. L. S
. (1993). Phytoplankton ag
greg
ate form
ation: observations of p
atterns an
d m
echanisms of
cell sticking and
the. significance of exopolymer m
aterial. J. P
lankton Res. 1
5: 993-1018
Kiarboe, T
., Lu
nd
sgaard
, C., O
lesen, M., Han
sen, J. (1994).
Aggregation an
d sed
imen
tation
processes during a sprin
g
phytoplankton bloom: a field ex
perim
ent to test co
agu
la- lion theory. J. m
ar. Res. 52: 297-323
La), D
., Lerm
an, A. (1975). S
ize spectra of b
iog
enicp
articles in ocean w
aler and
sedim
ent. J. q
eop
hy
s. Res. 80: 423-430
Lam
bert, C. E
., Jehan
no
, C., S
ilverb
erg, N
., B
run-Cottan,
J. C., C
hesselet, R. (1
98
1). Log-norm
al distributions of sus-
pen
ded
particles in the o
pen
ocean
. J. mar. R
es. 39: 77-98
Th
is article was su
bmitted to th
e editor
Lerm
an, A
. (1979). Geochernical processes' w
ater and
sedi- m
ent environments. J. W
iley and
Sons, N
ew
York
Lerm
an, A., C
arder, K
. L., B
etzer, P. R. (1977). E
Jimination of
fine suspensoids
in th
e oceanic w
ater colum
n. E
arth P
ldiiet. Sci. L
ett 37: 61-70 L
inley, E. A
. S., F
ield, J. G. (1982). T
he n
ature an
d ecological
significance of bacterial ag
greg
ation
in a nearsh
ore u
p-
welling system
. Estuar
coast. Shelf S
ci. 14: 1 - 11 L
ogan, B. E
., Kilps, J. R
. (1994). Fractal dim
ensions of agg
re- g
ates formed in different fluid m
echanical environments.
Water R
es. 39: 390-395 L
ogan, B. E
., Kirctunann, D
. L. (1991). Uptake of
dissolved orgarucs by m
arine bacteria as a function of fluid m
otion. M
ar. Biol. Ill: 175-181
Logan, B
. E., P
assow, U
., Alldredge, A
. L. (1994). V
anab
le retention of
diatoms on screen
s du
ring
size separations. L
imnol. O
ceanogr. 39: 390-395 L
ogan, B. E.,
Wilkinson, D
. B.
(1990). F
ractal geometry of
marine snow
and
oth
er biological ag
greg
ates. L
imnol,
Oceanogr. 35: 130-136
McC
ave. I. N. (1984). S
ize spectra and
agg
regatio
n of
sus- p
end
ed particles in
the d
eep ocean. D
eep Sea R
es. 31: 329-352
Parker, B
. C., D
iboll, A. G
. (1966). Alcian stains for histochem
- ical localization of
acid and
sulfciird polysaccharides in alg
ae. Phycologia 6: 36-46
Pain
chau
d, J
,, Therriault, J.-C
. (1989). Relationship betw
een bacteria, phytoplankton an
d particulate organic carbon in
the u
pp
er St. L
awrence estuary. M
ar. Ecol. P
rog. Ser. 56:
301-311 P
assow, U
., Alldredge, A
. L., Logan, 8. E
. (1994). Th
e role of particulate carbohydrate ex
ud
att's in the flocculalion of
diatom bloom
s. Deep
Sea R
es. 4 1: 335-357 P
assow, U,, W
assmann, P. (1994). O
n the trophic fate of Phaeo-
cystis pouch
etii (Hariot): IV
. Th
e formation of m
ann
e sno
w
by P. pouch
etii. Mar. E
col. Prog. S
er. 104: 153-161 P
orter, G. K
., Feig, Y
. S
. (1980). Th
e use of
DA
P1 for identily- ing an
d counting aquatic m
icroflora. Linm
ol. Oceanogr.
25: 943-948 S
heldon, R. W
., Prak
ash, A
, Sutcliff, W
. H. Jr (1972). T
he size
distribution of particles in th
e ocenn. Lim
nol. Ocean
og
r. 17: 327-340
Vandevivere, P., K
irchman, D
. L. 11993). A
ttachment stim
u- lates exopolysaccharide synthesis by a bacterium
. Appl.
environ. Microbiol. 59: 3280-3286
Manuscript first
received: Janu
ary 25, 1994 R
evised version accep
ted: June 8
, 1994
