Content uploaded by Thomas D. Brock
Author content
All content in this area was uploaded by Thomas D. Brock on May 16, 2016
Content may be subject to copyright.
APPLiED
AND
ENVIRONMENTAL
MICROBIOLOGY,
Oct.
1977,
p.
433-452
Copyright
X)
1977
American
Society
for
Microbiology
Vol.
34,
No.
4
Printed
in
U.S.A.
Structure,
Growth,
and
Decomposition
of
Laminated
Algal-
Bacterial
Mats
in
Alkaline
Hot
Springs
W.
N.
DOEMEL'
AND
THOMAS
D.
BROCK2*
Biology
Department,
Wabash
College,
Crawfordsville,
Indiana
47933,1
and
Department
of
Bacteriology,
University
of
Wisconsin,
Madison,
Wisconsin
537062
Received
for
publication
12
April
1977
Laminated
mats
of
unique
character
in
siliceous
alkaline
hot
springs
of
Yellowstone
Park
are
formed
predominantly
by
two
organisms,
a
unicellular
blue-green
alga,
Synechococcus
lividus,
and
a
filamentous,
gliding,
photosyn-
thetic
bacterium,
Chloroflexus
aurantiacus.
The
mats
can
be
divided
approxi-
mately
into
two
major
zones:
an
upper,
aerobic
zone
in
which
sufficient
light
penetrates
for
net
photosynthesis,
and
a
lower,
anaerobic
zone,
where
photosyn-
thesis
does
not
occur
and
decomposition
is
the
dominant
process.
Growth
of
the
mat
was
followed
by
marking
the
mat
surface
with
silicon
carbide
particles.
The
motile
Chloroflexus
migrates
vertically
at
night,
due
to
positive
aerotaxis,
responding
to
reduced
02
levels
induced
by
dark
respiration.
The
growth
rates
of
mats
were
estimated
at
about
50
,um/day.
Observations
of
a
single
mat
at
Octopus
Spring
showed
that
despite
the
rapid
growth
rate,
the
thickness
of
the
mat
remained
essentially
constant,
and
silicon
carbide
layers
placed
on
the
surface
gradually
moved
to
the
bottom
of
the
mat,
showing
that
decomposition
was
taking
place.
There
was
a
rapid
initial
rate
of
decomposition,
with
an
apparent
half-time
of
about
1
month,
followed
by
a
slower
period
of
decompooi-
tion
with
a
half-time
of
about
12
months.
Within
a
year,
complete
decomposition
of
a
mat
of
about
2-cm
thickness
can
occur.
Also,
the
region
in
which
decomposition
occurs
is
strictly
anaerobic,
showing
that
complete
decomposition
of
organic
matter
from
these
organisms
can
occur
in
the
absence
of
02.
Widespread
interest
in
the
paleomicrobiology
of
Precambrian
stromatolites
has
focused
atten-
tion
on
living
laminated
algal
mats
that
might
be
similar
to
those
which
formed
the
Precam-
brian
deposits
(35).
Extensive
work
has
been
done
over
the
past
several
decades
on
stroma-
tolitic
blue-green
algal
mats
in
marine
environ-
ments,
and
a
few
studies
on
freshwater
algal
reefs
have
also
been
carried
out.
Early
work
on
the
algal
mats
of
Yellowstone
National
Park
by
Weed
(38)
focused
geological
attention
on
these
extensive
and
interesting
structures,
but
little
microbiological
and
ecological
work
was
done
until
the
mid-1960s,
when
this
laboratory
began
a
long-term
investigation
of
the
geomi-
crobiology
of
geothermal
habitats.
The
Yellow-
stone
mats
are
of
special
interest
because
they
are
siliceous
and
hence
may
provide
a
model
(or
analog)
for
the
depositional
environment
of
Precambrian
iron
formations
(34).
Indeed,
many
of
the
microorganisms
of
the
Yellowstone
mats
resemble
morphologically
the
fascinatinig
microbiota
of
the
Gunflint
(3).
In
a
brief
study,
Walter
et
al.
(36)
reported
on
living
conophy-
ton-like
structures
formed
by
a
blue-green
alga
(Phormidium
sp.)
that
grew
in
hot
springs
at
temperatures
around
40
to
450C.
More
detailed
studies
on
these
structures
are
reported
by
Walter
et
al.
(37).
The
present
study
deals
with
a
different
type
of
mat
which
occurs
exten-
sively
in
hot
springs
of
neutral
to
alkaline
pH
at
temperatures
of
55
to
650C.
This
mat
is
composed
of
two
quite
different
photosynthetic
organisms,
a
unicellular
blue-green
alga
(Sy-
nechococcus
lividus)
and
a
gliding,
filamen-
tous,
photosynthetic
bacterium
(Chloroflexus
aurantiacus).
These
two
organisms,
and
asso-
ciated
heterotrophic
bacteria,
form
well-defined
laminated
mats
of
considerable
thickness,
in
which
the
photosynthetic
bacterium
is
respon-
sible
for
the
structural
integrity.
These
mats
are
of
special
interest
because
they
occur
at
temperatures
too
high
for
the
development
of
grazing
animals
(or
any
other
eucaryotic
orga-
nisms)
and
so
provide
some
insight
into
how
mats
might
have
developed
in
the
Precam-
brian,
when
metazoans
were
absent.
As
Awra-
mik
(2)
has
shown,
stromatolite
diversity
shows
a
marked
decrease
in
the
Late
Precambrian,
at
a
time
when
metazoan
life
first
appeared.
However,
decomposition
processes
in
the
Yel-
lowstone
mats
occur
readily
despite
the
absence
433
434
DOEMEL
AND
BROCK
of
metazoans,
so
a
complete
procaryotic
food
chain
leading
to
mineralization
of
organic
mat-
ter
can
occur.
STUDY
AREAS
The
locations
of
Octopus
Spring
and
Ravine
Spring
have
been
given
by
Madigan
and
Brock
(24).
Octopus
Spring
(Pool
A)
is
a
typical
alkaline
spring
(pH
8.5
to
8.7). It
is
located
about
0.15
km
SSE
of
Great
Fountain
Geyser
and
the
White
Creek
Valley.
An
extensive
microbial
mat,
about
2
to
4
cm
deep,
exists
in
a
shallow
extension
of
this
pool
(Fig.
1).
This
backwater
region
is
separated
from
the
main
pool
by
an
incomplete
barrier
of
siliceous
sinter
(5).
Although
many
hot
springs
have
short
life
spans,
Octopus Spring
seems
to
be
relatively
stable,
since
it
can be
readily
identified
from
Peale's
1883
descrip-
tion
(25)
as
his
no.
10
(p.
165),
with
temperature
and
flow
conditions
similar
to
those
now
existing.
Springs
74-6
(source:
80°C,
pH
7.7),
74-7
(source:
68°C,
pH
9.2),
and
74-3
(source:
78°C,
pH
8.0
to
8.2)
are
near
Ravine
Spring.
Buffalo
Pool
is
an
intercon-
nected
group
of
three
springs
(source:
91°C,
pH
9.3)
located
on
the
east
bank
of
White
Creek
upstream
from
pool
74-3.
Toadstool
Spring
is
located
about
5
m
from
Mushroom
Spring
(4).
This
spring
first
started
to
flow
as
a
geyser
during
the
summer
of
1969
and
then
subsided
to
a
spring.
During
the
summer
of
1973,
the
spring
had
a
source
tempera-
ture
of
about
85°C,
a
good
flow,
and
a
single
effluent
channel
with
a
gradient
from
80°C
to
ambient
tem-
perature.
By
30
May
1974,
the
source
temperature
had
decreased
to
75°C
and
the
minimal
flow
rate
from
the
source
no
longer
supported
an
extensive
mat.
In
June
1975,
the
spring
no
longer
flowed
and
the
algal
mat
had
dried.
Sulfide
Spring
(source:
86°C,
pH
6.6;
sulfide
level,
1.3
mg/liter)
is
one
of
the
two
springs
located
about
0.2
km
ENE
of
the
Fire-
hole
Lake
parking
area,
50
m
south
of
the
Firehole
Lake
Loop
Road.
MATERIALS
AND
METHODS
Measurement
of
light
intensity.
Measurements
of
light
intensity
in
the
field
were
made
with
a
Gossen
Pilot
light
meter
equipped
with
an
incident
light
attachment.
Measurements
of
the
reduction
of
light
by
microbial
cores
were
made
with
a
Kipp
and
Zonen
solarimeter
connected
to
a Keithley
model
601
electrometer
(Keithley
Instruments,
Inc.).
The
solarimeter
was
modified
by
replacing
the
protective
FIG.
1.
Octopus
Spring
mat.
The
source
is
in
the
background
and
the
mat
is
in
the
foreground,
in
the
backwater
behind
the
sinter
dike.
APPL.
ENVIRON.
MICROBIOL.
ALGAL-BACTERIAL
MATS
IN
ALKALINE
HOT
SPRINGS
435
globe
with
a
cardboard
mask
having
a
5.8-mm-
diameter
hole.
Light
from
a
12-V,
60-W
ribbon
filament
lamp
(Zeiss
microscope
illuminator)
was
focused
with
a
lens
so
that
a
circle
covered
the
hole
in
the
mask.
To
calculate
the
radiation
intensity,
the
calibration
value
supplied
by
Kipp
and
Zonen
was
used
(1
g
cal/cm2
per
min
=
8.8
mV).
Measurement
of
photosynthesis.
The
procedures
for
measuring
photosynthesis
have
been
described
elsewhere
(9,
24).
To
distinguish
photosynthesis
by
Chloroflexus
from
photosynthesis
by
blue-green
al-
gae,
0.1
ml
of
5
x
10-4
M
3-(3,4-dichlorophenyl)-1,1-
dimethylurea
(DCMU)
was
added
to
appropriate
vials
(4).
Thin
sections
of
algal
cores.
Cores
(3.5-mm
ID)
were
prepared
for
thin
sectioning
by
dehydration
through
an
acetone
series
and
then
infiltration
with
Durcupan
A/M
(Polyscience,
Inc.,
Warrington,
Pa.).
Sections
1
um
thick
were
prepared
with
a
Porter-
Blum
ultramicrotome,
placed
onto
clean
slides,
and
dried
at
70
to
80°C
for
15
to
30
min.
After
cooling,
the
sections
were
rinsed
with
three
changes
of
distilled
water
and
then
dried.
One
to
two
drops
of
Richardson's
strain
(0.5%
azure
II,
0.5%
methylene
blue,
0.5%
borax)
was
placed
onto
each
section,
and
the
slide
was
warmed
at
60
to
65°C
for
5
min.
The
excess
stain
was
removed,
and
the
slide
was
rinsed
in
distilled
water.
After
drying,
a
drop
of
resin
was
added,
followed
by
a
cover
slip,
and
the
slides
were
dried
at
55
to
60°C
for
48
h.
Measurement
of
mat
growth.
The
mat
was
marked
periodically
with
15.5-cm
circles
of
120-grit
silicon
carbide
(Buehler
Ltd.,
Evanston,
Ill.),
placed
on
the
surface
with
a
metal
template.
About
1
teaspoon
of
silicon
carbide
sprinkled
onto
the
mat
gave
a
thickness
of
0.3
to
0.7
mm.
The
center
of
each
circle
was
marked
with
an
iron
nail.
These
nails
appeared
not
to
oxidize
in
the
spring
and
were
intact
after
7
years.
Two
to
three
cores
were
collected
from
each
of
these
stations
with
a
brass
cork
borer
(8.4
or
9.4
mm
in
diameter).
The
cores
were
immediately
placed
into
30-mm
petri
dishes
containing
2%
molten
agar,
which
then
was
allowed
to
harden.
The
agar
stabilized
the
core
during
transport
and
storage
and
also
enabled
the
core
to
be
sectioned.
The
agar
disk
around
the
core
was
removed
from
the
petri
dish
and
trimmed
so
that
the
core
was
encased
in
a
rectangle
of
agar.
This
agar
block
was
then
halved
longitudinally
with
a
razor
blade.
One
of
the
halves
was
placed
on
a
glass
slide
so
that
the
surface
of
the
core
was
exposed.
With
a
Zeiss
Photomicroscope
I
equipped
with
a
Optivar,
using
either
a
x2.5
or
x6.3
objective
and
an
optical
micrometer,
the
height
of
mat
above
the
silicon
carbide
was
estimated.
The
magnification
was
adjusted
to
minimize
measuring
errors,
and
at
least
10
measurements
of
the
height
above
the
silicon
carbide
were
made
on
each
core.
After
the
height
of
the
mat
above
the
carborun-
dum
was
measured,
this
portion
of
the
core
was
sectioned
from
both
halves
of
the
core
with
a
razor
blade,
and
the
protein
was
measured
as
described
below.
Measurement
of
biomass.
Chlorophyll
a
and
bac-
teriochlorophyll
c
were
used
as
measures
of
photo-
synthetic
biomass.
Although
in
vivo
absorption
spectra
differ,
in
solvent
extracts
bacteriochloro-
phyll
c
and
chlorophyll
a
have
similar
maxima,
between
662
and
668
nm.
To
distinguish
these
two
chlorophylls,
an
in
vivo
procedure
(27)
was
used
which
involves
eliminating
most
of
the
light
scatter-
ing
caused
by
the
cells
by
suspending
material
in
a
saturated
sucrose
solution.
The
refractive
index
of
this
solution
approximates
that
of
the
cells,
so
the
in
vivo
peaks
of
chlorophyll
a,
680
nm,
and
bacte-
riochlorophyll
c,
740
nm,
can
be
resolved.
Since
some
scattering
still
remains, thus
raising
the
base
line,
the
high
base
lines
were
corrected
by
subtracting
the
absorbance
at
640
and
700
nm,
respectively.
The
concentration
of
both
chlorophylls
was
also
determined
in
methanolic
extracts
if
two
distinct
peaks
could
be
discerned
(4).
In
some
of
the
later
work,
a
thin-layer
chromatographic
procedure
(23)
was
used
to
separate
and
quantify
the
two
chloro-
phylls.
Protein
was
determined
by
the
Lowry
method
as
modified
by
Brock
and
Brock
(8).
Before
protein
was
extracted
from
the
cores
in
agar,
the
samples
were
boiled
for
30
min
in
0.5
N
perchloric
acid,
which
solubilized
the
agar
and
precipitated
the
protein.
Oxygen
measurements.
Dissolved
oxygen
in
the
water
immediately
above
the
surface
of
the
mat
was
measured
by
modifying
the
standard
azide
modification
of
the
Winkler
method
for
dissolved
oxygen
(1).
Serum
bottles
(60-ml
Vitro
400;
Wheaton
Glass
Co.)
with
sleeved
rubber
serum
stoppers
were
flushed
with
nitrogen
gas
having
less
than
0.05%
oxygen
(Matheson
Gas
Products
Co.)
for
5
min.
After
flushing
and
removal
of
the
needles
so
that
a
slight
positive
pressure
remained,
the
stoppers
were
sealed
with
a
thin
film
of
silicone
sealant
(Dow
Chemical
Co.).
A
30-ml
disposable
syringe
with
a
2-
inch
needle
was
then
flushed
with
nitrogen
and
used
to
withdraw
30
ml
of
gas
from
the
vials.
This
syringe
then
was
slowly
filled
with
water
from
the
spring,
and
the
water
was
injected
into
the
vial;
again
a
positive
pressure
was
maintained.
To
each
vial,
0.2
ml
of
alkaline
potassium
iodide
solution
was
injected,
followed
by
0.2
ml
of
manganous
sulfate
solution.
The
vials
were
vigorously
mixed
for
several
minutes;
the
floc
was
allowed
to
settle;
and
0.2
ml
of
concentrated
sulfuric
acid
was
injected
into
each
vial.
The
iodine
freed
was
then
titrated
with
a
0.0025
N
thiosulfate
solution,
with
a
soluble
starch
solution
added
toward
the
end
of
the
titration
to
increase
sensitivity.
RESULTS
Morphology
of
Synechococcus-Chloroflexus
Mats.
Although
there
are
several
classes
of
microbial
mats
in
alkaline
thermal
springs,
only
the
microbial
mats
containing
S.
lividus,
a
blue-green
alga
(cyanobacterium),
and
Chlo-
roflexus,
a
filamentous,
photosynthetic
bacte-
rium,
are
considered
here.
A
typical
Synecho-
coccus-Chloroflexus
microbial
mat
is
in
a
tide
pool
of
Octopus
Spring
(Fig.
1)
and
in
the
VOL.
34,
1977
436
DOEMEL
AND
BROCK
stream
flowing
from
this
pool.
Although
the
surface
of
this
mat
appeared
at
first
glance
to
be
smooth
to
slightly
rough
with
no
distinctive
structure,
a
closer
examination
revealed
the
presence
of
several
distinctive
types
of
struc-
tures.
(i)
On
the
mat
in
the
pool
and
the
stream,
clusters
of
1-
to
2-mm
conical
projec-
tions,
called
nodules,
were
present
at
tempera-
tures
up
to
about
650C
(Fig.
2).
Often
adjacent
nodes
were
interconnected
by
short,
raised
ridges,
producing
a
mat
consisting
of
a
fine
network
of
nodes
and
ridges
(Fig.
3).
(ii)
Also
present
on
these
mats,
but
less
frequently,
were
2-
to
10-mm
circular
structures,
called
colonies
(Fig.
4).
These
colonies
usually
were
orange
with
yellow-green
centers,
and
in
the
orange
region
there
sometimes
were
several
concentric
ridges
and/or
nodules.
(iii)
On
the
mats
in
flowing
water
there
were
yet
other
distinct
structures,
called
streamers
(Fig.
5
shows
streamers
from
Toadstool
Spring).
These
streamers
extended
several
centimeters
from
their
attachment
to
the
mat
and
appeared
to
be
more
abundant
where
the
flow
was
more
turbulent.
The
typical
laminated
mat,
similar
to
a
stro-
matolite
(43),
was
beneath
these
surface
struc-
tures
(Fig.
6).
To
better
understand
the
distri-
bution
of
microorganisms
within
the
laminated
mat,
microscopy
was
done
on
verticle
sections
(prepared
with
a
razor
blade)
of
fresh
cores
embedded
in
2%
agar.
Because
of
the
opacity
of
such
sections,
visualization
was
done
with
vertical
fluorescence
illumination.
The
Syne-
chococcus
could
be
observed
directly
with
this
microscopic
technique
because
of
the
autofluo-
rescence
of
chlorophyll
a.
The
sections
were
then
stained
with
a
0.1%
solution
of
the
fluores-
cent
dye
acridine
orange;
with
this
stain,
bac-
teria
fluoresce
either
red
or
green.
Also,
sam-
ples
were
removed
from
the
various
lamina-
tions
and
were
examined
directly
with
phase
and
phase-fluorescence
illumination.
These
mi-
croscopic
observations
are
summarized
in
Table
1.
In
the
upper
layer,
0.2
to
1
mm
in
thickness,
Synechococcus
and
Chloroflexus
predominated
(Fig.
7).
Often
there
was
a
second
layer,
0.1
to
0.6
mm
thick,
immediately
beneath
this
upper
layer
which
was
similar
except that
it
was
a
dark
blue-green
and
the
autofluorescence
of
the
Synechococcus
was
much
more
intense.
Beneath
this
region,
Synechococcus
was
rarely
observed,
even
as
empty
cells,
but
there
was
an
abundance
of
Chloroflexus
in
this
next
layer,
0.8
to
1.4
mm
thick.
Although
the
major-
ity
of
filaments
in
this
region
did
not
autoflu-
oresce
and
although
other
organisms
were
usu-
ally
absent,
in
some
thin
sections
there
were
APPL.
ENVIRON.
MICROBIOL.
1.1-,m-diameter
filaments
that
autofluoresced
and
were
similar
to
the
filamentous
blue-green
alga
Pseudanabaena
(S.
Golubic,
personal
communication).
Below
about
3
to
5
mm,
most
of
the
filamentous
bacteria
appeared
moribund
(Fig.
8),
and
unicellular
rods
were
concentrated
in
bands
having
sulfide-forming
activity,
as
revealed
by
the
ferrous
ammonium
sulfate
technique
(see
Materials
and
Methods),
sug-
gesting
that
at
least
some
are
sulfate-reducing
bacteria.
Often
in
the
lower
regions,
refractile
spherical
bodies
resembling
myxobacterial
cysts
were
observed
that
neither
autofluoresced
nor
fluoresced
when
stained
with
acridine
or-
ange.
Also,
in
a
number
of
samples,
chains
of
spherical
cells
were
seen
that
were
similar
to
an
organism
described
by
Geitler
(19)
as
the
blue-green
alga
Isocystis.
However,
these
cells
never
autofluoresced
and
hence
either
are
not
blue-green
algae
or
are
not
viable.
Bauld and
Brock
(4)
demonstrated
that
bac-
teriochlorophyll
c
distribution
correlated
with
the
distribution
of
the
bacterial
filaments
and
isolated
a
number
of
strains
of
the
filamentous
photosynthetic
bacterium
Chloroflexus.
Since
microscopy
of
isolated
nodes,
colonies,
and
streamers
revealed
that
all
of
these
structures
were
composed
predominantly
of
bacterial
fila-
ments
with
some
Synechococcus,
the
composi-
tion
of
these
structures
was
further
investi-
gated
with
[14C]bicarbonate
incorporation,
chlorophyll
analysis,
and
autoradiography.
Colonies,
nodules,
and
streamers
were
sam-
pled
from
mats
at
Octopus
Spring
and
Toad-
stool
Geyser.
Colonies
were
removed
with
a
cork
borer
having
a
larger
diameter
than
that
of
the
colony.
With
a
dissecting
microscope,
the
cores
were
separated
in
the
field
into
sev-
eral
serial
thin
sections,
approximately
0.5
mm
thick.
Each
of
these
sections
was
gently
homog-
enized
in
5
ml
of
water,
and
the
photosynthetic
activity
and
chlorophyll
content
of
the
homoge-
nate
was
measured.
Although
the
chlorophyll
concentrations
and
proportions
were
variable
in
different
cores,
the
colonies
were
enriched
in
bacteriochlorophyll
c
in
the
upper
layers.
In
contrast,
the
regions
surrounding
the
colo-
nies
were
high
in
chlorophyll
a.
In
the
colonies,
a
high
percentage
of
incorporation
of
[14C]bi-
carbonate
was
insensitive
to
DCMU.
Similar
experiments
were
done
with
nodules
and
streamers.
Both
were
removed
from
the
microbial
mat
by
aspiration
with
a
Pasteur
pipette
and
rubber
bulb.
In
nodules,
bacterio-
chlorophyll
c
was
present
in
considerably
higher
amounts
than
in
the
surface
layers
of
surrounding
mat.
Microscopy
suggested
that
nodules
are
composed
of
a
dense
mass
of
Chlo-
roflexus
filaments
arranged
in
a
conical
shape,
FIG.
2.
Nodules
on
the
Octopus
Spring
mat.
The
surface
of
the
microbial
mat
is
often
covered
with
1-
to
2-mm
conical
projections.
Diameter
of
core,
about
3.5
mm.
FIG.
3.
Surface
of
a
core
similar
to
that
in
Fig.
2.
The
nodules
extend
1
to
2
mm
above
the
surface
and
are
often
connected
by
ridges
forming
a
fine
lattice.
FIG.
4.
Orange
colonies
on
the
Octopus
Spring
mat.
Photo
taken
directly
through
the
water.
The
center
of
the
colony
is
often
green,
shown
here
by
a
small
dark
dot.
Distance
on
the
photograph
about
20
cm.
FIG.
5.
Streamers.
Filamentous
streamers
(St)
in
the
effluent
of
Toadstool
Spring.
The
light
circles
are
silicon
carbide
(C),
15
cm
in
diameter,
and
serve
as
markers.
The
streamers
are
present
only
in
flowing
water.
The
current
flows
from
left
to
right.
FIG.
6.
Macrophotograph
of
a
complete
core,
showing
laminations.
This
is
one
from
the
Octopus
Spring
mat.
Note
the
surface
nodules
(N),
the
upper
dark
layer
(D),
and
the
laminae
(L)
below
this
layer.
Often
there
are
bands
of
sinter
(Si)
in
the
core
(coarse
white
zones).
437
TABLE
1.
Topography
and
composition
of
a
mat
at
Octopus
Springa
Core
section
Phase-contrast
and
fluorescence
microscopy
of
macer-
Core
section
Direct
microscopy
of
whole
cores,
(mm)
ations
of
core
sections
(mm)
using
acridine
orange
fluores-
cence
by
incident
light
0-0.2
Synechococcus
>
Chloroflexus
0-1.2
Synechococcus,
filaments,
'Is-
ocystis"
0.55-1.6
Highly
granulated
Chloroflexus
>
rods,
no
Sy-
1.2-2.5
Filaments
and
debris
nechococcus,
refractile
spheres,
some
algal
filaments
1.6-2.1
Moribund
Synechococcus,
refractile
Chloro-
2.5-3.2
Amorphous
debris
flexus
and
short
rods
2.9-3.65
Chloroflexus
and
"Isocystis,"
some
algal
fila-
3.2-4.2
Few
Synechococcus,
with
large
ments
numbers
of
rods
5-7.5
Moribund
filaments
4.2-6.5
Amorphous
material
8.7-10.2
Rods,
few
moribund
filaments,
no
Synechococ-
6.5-7.5
Bacterial
rods
and
moribund
cus,
sulfate
reduction
filaments,
some
Synecho-
coccus
10.2-11.7
Gelatinous
matrix,
primarily
unicellular
rods,
7.5-7.8
Dense
rods
moribtnd
filaments
11.7-12.55
Unicellular
rods,
few
moribund
filaments
7.8-8.9
Dense
rods
and
moribund
fil-
aments
12.55-15.2
Unicellular
rods,
sulfate
reduction
8.9-11
Dense
rods
15.2-15.4
Unicellular
rods
(0.4
,um),
no
filaments,
pine
11-17
Amorphous
material
pollen
-
a
This
is
a
composite
of
a
series
of
cores
incubated
in
0.4%
ferrous
ammonium
sulfate
agar
(2%)
to
detect
regions
of
sulfide
production,
and
then
split
vertically
with
a
razor
blade
and
examined
microscopically.
Regions
of
sulfate
reduction
are
indicated
in
column
two.
The
designation
"bacterial
filaments"
includes
all
filamentous
organisms
lacking
algal
chlorophyll
a
(as
revealed
by
absence
of
red
autofluorescence).
They
were
0.5
to
1.5
,im
in
diameter
and
were
of
indeterminate
length;
they
were
probably
Chloroflexus.
Algal
filaments
showed
red
autofluorescence
and
were
1.1
Aum
in
diameter;
they
resembled
Pseudanabaena.
The
bacterial
rods
were
0.7
Am
in
diameter
and
4.0
,um
in
length
and
were
probably
not
photosynthetic.
In
addition
to
direc'
visualization
by
phase-contrast
and
fluorescence
microscopy
of
macerated
samples,
the
vertical
distribution
of
organisms
were
visualized
without
disturbance
by
adding
100
,ug
of
acridine
orange
per
ml
to
the
flat
surface
of
split
cores
and
observing
with
incident
light
by
fluorescence
microscopy.
with
the
Synechococcus
concentrated
in
pock-
ets
of
filaments.
In
streamers,
chlorophylls
were
extracted
and
isolated
with
thin-layer
chromatography
(23).
About
40%
of
the
total
chlorophyll
absorbing
at
665
nm
was
bacter-
iochlorophyll
c,
the
remainder
being
chloro-
phyll
a.
Results
with
[14C]bicarbonate
incorpo-
ration
support
the
observation
of
high
Chloro-
flexus
concentrations
in
the
streamers.
These
observations
with
colonies,
nodules,
and
streamers
demonstrate
that
a
large
proportion
of
the
population
in
these
structures
is
photo-
synthetic
bacteria.
Since
bacterial
filaments
are
abundant
whereas
other
bacteria
are
few,
and
since
Chloroflexus
has
been
isolated
from
these
structures,
this
is
indirect
evidence
sup-
porting
the
structural
role
of
Chloroflexus.
If
the
mat
is
held
together
by
Chloroflexus,
then
removal
of
the
Synechococcus
should
not
significantly
alter
the
structure
of
the
mat.
Preliminary
experiments
demonstrated
that
when
a
microbial
mat
was
darkened,
within
24
h
the
mat
turned
orange
in
color
and
this
orange
layer
was
enriched
in
Chloroflexus.
Also,
as
shown
by
Madigan
and
Brock
(24),
Chloroflexus
is
able
to
photosynthesize
at
con-
siderably
lower
light
intensities
than
Synecho-
coccus.
This
observation
implied
that
at
low
light
intensities,
Synechococcus
would
not
be
able
to
maintain
a
population
and
Chloroflexus
would
predominate.
To
test
this,
an
area
of
the
mat
at
56°C
in
the
stream
flowing
from
Toad-
stool
Geyser
was
covered
by
two
layers
of
a
dense
nylon
cloth
held
on
a
frame,
which
re-
duced
the
incident
light
by
98%.
At
the
same
time,
a
circle
of
120-mesh
silicon
carbide
was
FIG.
7.
Organisms
of
the
upper
layer
of
the
mat,
as
seen
by
phase-contrast
microscopy.
The
filamentous
cells
are
Chloroflexus,
and
the
curved
rod-shaped
cells
are
Synechococcus.
Bar
=
10
,um.
FIG.
8.
Photomicrograph
by
phase
contrast
of
a
smear
from
a
lower
region
of
the
mat.
The
Chloroflexus
filaments
are
moribund.
Mostly
the
material
is
a
formless
debris.
Though
not
obvious
in
the
micrograph,
bacterial
rods
were
often
present.
Same
magnification
as
Fig.
7.
438
DOEMEL
AND
BROCK
APPL.
ENVIRON.
MICROBIOL.
4
tO
"A
439
440
DOEMEL
AND
BROCK
placed
on
the
mat
as
a
marker
and
substratum.
Within
4
h
small
nodules
had
formed
on
the
surface
of
the
mat
and
on
the
surface
of
the
silicon
carbide,
and
within
24
h
a
number
of
pinkish-orange
nodules
were
present
(Fig.
9).
Microscopy
showed
that
Synechococcus
was
ab-
sent
from
the
nodules
and
internodular
regions,
yet
the
nodules
were
similar
to
normal
nodules.
Uptake
by
homogenates
of
these
nodules
of
[14C]bicarbonate
incubated
with
and
without
DCMU
had
equivalent
radioactivity,
suggest-
ing
that
the
algae
were
not
contributing
signif-
icantly
to
the
primary
production
of
these
nodes.
Further,
autoradiograms
prepared
of
this
material
indicated
that
all
of
the
incorpo-
ration
of
['4C]bicarbonate
was
by
bacterial
fila-
ments
(Table
2).
Together
these
observations
add
direct
evidence
to
support
the
notion
that
Chloroflexus
is
the
structural
component
of
the
mat
and
furthermore
that
it
accounts
for
the
growth
of
the
mat.
The
apparent
involvement
of
Chloroflexus
in
the
structure
and
growth
of
the
mat
and
in
the
development
of
the
differentiated
nodules,
streamers,
and
colonies
implies
that
there
may
be
a
nonrandom
orientation
of
the
filaments
within
the
mat.
To
determine
whether
the
orientation
of
filaments
within
the
microbial
mat
is
nonrandom,
1-,um
thin
sections
were
APPL.
ENVIRON.
MICROBIOL.
prepared
from
resin-embedded
corings
and
after
staining
were
examined
with
a
light
mi-
croscope
(see
Materials
and
Methods).
Obser-
vations
of
cores
from
Octopus
Spring
are
re-
ported
here.
Two
patterns
of
orientation
were
observed.
In
some
samples,
in
the
uppermost
region,
the
unicellular
blue-green
alga
was
present
but
showed
no
preferred
orientation,
and
the
photosynthetic
bacterial
filaments
were
oriented
horizontally
(Fig.
10).
In
the
lower
regions,
most
of
the
filamentous
bacteria
were
oriented
vertically
and
Synechococcus
was
not
apparent
(Fig.
11).
In
some
regions,
vertically
oriented
filaments
extended
into
the
upper
region
(Fig.
12).
TABLE
2.
Photosynthesis
by
nodules
under
low
light,
as
seen
by
autoradiographya
Condition
Chloroflexus
Synechococcus
(grains/25
Am3)
(grains/25
Am3)
Light
.
23.3
0
Light
with
DCMU
...
15.3
0
Dark
.
4.6 0
a
Nodules
on
the
surface
of
a
mat
under
a
98%
light
reduction
cover
were
harvested
by
aspiration
with a
Pasteur
pipette,
homogenized,
and
incubated
with
['4C]bicarbonate
for
4
h.
Autoradiograms
were
prepared,
and
grains
over
Chloroflexus
and
Syne-
chococcus
were
counted.
FIG.
9.
Pinkish-orange
nodules
of
Chloroflexus,
which
developed
in
24
h
on
top
of
a
silicon
carbide
circle
laid
down
in
Toadstool
Geyser.
A
neutral
density
filter
over
the
mat
had
reduced
the
light
intensity
to
98%
of
full
sunlight.
Diameter
of
circle
on
photograph
about
85
mm.
FIG.
10.
Photomicrograph
of
a
thin
section
of
the
upper
layer
of
Octopus
Spring
from
resin-embedded
material.
The
surface
of
the
mat
is
toward
the
top
of
the
page.
The
cells
are
predominantly
Synechococcus.
No
preferred
orientation
is
observed.
Bar
=
10
,um.
FIG.
11.
Photomicrograph
of
a
thin
section
of
the
lower
region
of
the
same
mat
as
Fig.
10,
showing
vertical
orientation
of
Chloroflexus
filaments.
No
Synechococcus
is
present
at
this
depth.
Bar
=
10
,um.
FIG.
12.
Photomicrograph
of
a
thin
section
showing
vertical
filaments
extending
into
the
upper
region
through
a
zone
of
dense
Synechococcus.
Bar
=
10
gm.
FIG.
13.
Photomicrograph
of
nodular
mat,
showing
alternating
regions
with
horizontal
Chloroflexus
filaments
at
right
angles.
The
top
of
the
mat
is
in
the
upper
right-hand
corner.
Bar
=
10
gm.
FIG.
14.
Vertical
section
of
a
core
from
Toadstool
Geyser,
showing
a
silicon
carbide
(C)
layer.
The
silicon
carbide
had
been
placed
on
the
surface
of
the
mat
29
days
before
and
is
now
at
a
depth
of
4
mm
into
the
mat.
The
silicon
carbide
layer
is
about
0.2
mm
thick.
441
442
DOEMEL
AND
BROCK
A
different
pattern
was
observed
in
a
portion
of
the
mat
having
abundant
nodes.
Again,
in
the
upper
regions
of
the
mat,
about
0.4
mm
thick,
the
blue-green
alga
was
present
and
the
filamentous
bacteria
were
horizontally
posi-
tioned.
Below
this
region,
the
alga
was
missing
and
the
bacterial
filaments
were
again
horizon-
tally
positioned,
but
in
different
directions.
In
the
section
shown
in
Fig.
13,
in
the
upper
regions
most
filaments
are
in
cross-section.
Immediately
below
this,
there
is
a
region
0.15
mm
thick
in
which
most
filaments
are
in
tan-
gential
section.
Below
this
there
is
another
region,
0.23
mm
thick,
in
which
the
filaments
are
again
in
cross-section.
These
regions
of
alternating
stacks
of
filaments
do
not
extend
through
the
entire
core.
In
one
portion
of
the
core,
the
filaments
are
still
horizontal
through-
out,
but
there
is
no
obvious
stacking
of
fila-
ments.
In
all
sections
and
cores
examined,
the
algae
were
restricted
to
the
upper
0.3
to
0.4
mm
of
the
mat.
Response
of the
mat
components
to
light.
The
depth
of
light
penetration
into
the
mat
was
estimated
directly
by
use
of
an
artificial
light
source
and
a
solarimeter
(see
Materials
and
Methods).
The
top
green
layer
and
an
underlying
orange
layer
of
similar
thickness
were
used.
These
sections
were
placed
on
glass
slides,
and
the
reduction
of
light
energy
by
each
of
the
sections
was
determined.
The
re-
sults
of
a
typical
core
are
shown
in
Table
3A.
The
green
layer
reduced
the
radiation
by
94%,
and
the
orange
layer
reduced
it
by
about
91%.
To
check
for
attenuation
by
photosynthetic
bac-
teria,
in
some
experiments
the
white
light
was
passed
through
an
infrared
(IR)
filter
(Tiffen
no.
88A)
before
being
passed
through
the
mat.
(This
filter
passes
no
radiation
below
730
nm.)
Although
the
reduction
of
IR
radiation
was
greater
in
some
cores,
the
difference
was
small
so
that
the
reduction
of
both
tungsten
and
IR
radiation
was
approximately
the
same.
Thus,
in
these
compact
mats,
light
attenuation
is
not
absorption
only
by
pigments,
but
also
by
scat-
ter.
Indeed,
because
there
was
no
difference
in
attenuation
by
tungsten
and IR
radiation,
it
seems
likely
that
scattering
causes
more
atten-
uation
than
does
pigment
absorption.
On
a
bright
cloudless
day,
light
intensities
may
reach
about
1.5
g
cal/cm2
per
min
(equiva-
lent
to
about
8,400
lx)
at
noon.
Since
the
thin
layer
of
water
overlying
the
mats,
1
to
5
cm
at
the
Octopus
Spring
mat,
does
not
significantly
reduce
the
light
intensity,
the
light
energy
passing
through
the
green
layer
and
reaching
the
orange
layer
can
be
calculated
(Table
3B).
About
2,500
to
3,500
lx
passes
through
the
green
layer,
which
is
probably
sufficient
to
support
autotrophic
growth,
but
only
180
to
450
lx
passes
through
both
layers,
about
1
mm
thick.
Because
these
are
mid-day
values,
most
of
the
time
the
light
intensity
is
much
lower.
Thus,
below
the
green
and
the
top
orange
layers,
the
mat
is
essentially
dark,
implying
that
the
light
penetrates
only
about
1
to
2
mm
into
the
mat.
Although
some
light
does
reach
the
orange
layer
of
the
mat,
Synechococcus
is
restricted
to
the
green
layer
(10),
and
even
cell
ghosts
of
TABLE
3.
Reduction
of
radiant
energy
by
the
green
and
orange
layers
of
a
microbial
mata
Light
intensity
(g
cal/cm2)
(A)
Light
source
Percent
reduc-
Percent
reduc-
Light
alone
With
green
layer
tion
Light
alone
With
orange
layer
tion
White
..............
1.0
0.056
94
1.0
0.097
91
Infrared
............
0.76
0.025
97
0.89
0.080
91
Calculated
light
intensity
(lx)
at
the
bottom
of:
(B)
Location
Green
layer
Orange
layer
Mat
(50C)
.........................................
3,440
450
Mat
(490)
.........................................
2,580
230
Mat
(54C)
.........................................
2,580
460
Outflow
(59-65°C)
...............
...................
4,620
420
Outflow
(50-54°C)
...............
...................
2,580
180
Mat
(580)
.........................................
2,580
360
a
(A)
Samples from
Octopus
Spring
(5000).
The
values
for
(B)
were
calculated
by
assuming
a
light
in-
tensity
of
84,000
lx
(maximum
intensity
at
mid-day).
The
light
intensity
passing
through
the
green
layer
(0.2
to
0.5
mm
thick)
was
determined
by
subtracting
the
radiant
energy
attenuated
by
the
green
layer
from
the
total
energy.
The
light
intensity
at
the
bottom
was
determined
by
multiplying
the
light
passing
through
the
green
layer
by
the
percentage
reduction
by
the
orange
layer
and
subtracting
this
from
the
light
passing
through
the
green
layer
(total
distance
=
0.4
to
1
mm).
APPL.
ENVIRON.
MICROBIOL.
ALGAL-BACTERIAL
MATS
IN
ALKALINE
HOT
SPRINGS
443
Synechococcus
are
not
present
in
the
orange
layer.
If
the
green
layer
of
a
core
is
homoge-
nized
and
the
rate
of
photosynthesis
as
a
func-
tion
of
light
intensity
is
determined,
the
photo-
synthetic
rate
of
Synechococcus
is
significantly
reduced
below
that
of
Chloroflexus
at
intensi-
ties
less
than
10,000
lx
(24).
There
is
also
some
evidence
of
high-light
inhibition
of
photosyn-
thesis
by
Synechococcus.
Probably
only
a
thin
upper
layer
of
cells
is
adapted
to
high
light,
since
there
is
a
distinct
darker
blue-green
band
immediately
below
the
surface
and
also
since
Synechococcus
cells
below
the thin
upper
layer
autofluoresce
more
strongly
when
excited
with
blue
light.
On
the
other
hand,
photosynthesis
by
Chloroflexus
is
light
saturated
at
all
inten-
sities
above
about
1,000
lx.
Synechococcus
may
be
restricted
to
the
upper
layer
simply
because
it
cannot
efficiently
utilize
low
light
intensities.
Chloroflexus
is
present
in
both
the
green
and
orange
layers
of
the
mat
(4).
However,
the
peak
concentration
of
bacteriochlorophyll
c
is
apparently
highest
in
the
upper
region
of
the
mat,
where
Synechococcus
is
also
present
(4).
The
Chloroflexus
in
these
upper
regions
also
appears
to
incorporate
['4C]bicarbonate
more
efficiently
(Table
4)
than
the
Chloroflexus
in
the
lower
regions.
On
a
protein
basis,
57%
of
the
[14C]bicarbonate
incorporated
by
Chloro-
flexus
is
in
the
nodes
and
the
upper
region
of
the
mat.
Furthermore,
if
it
is
conservatively
assumed
that
50%
of
the
protein
in
these
re-
gions
is
derived
from
Synechococcus,
the
con-
tribution
of
Chloroflexus
in
these
regions
in-
creases
to
73%
of
['4C]bicarbonate
incorpora-
tion
by
Chloroflexus
in
all
layers.
These
obser-
vations,
together
with
the
observations
on
light
attenuation
in
the
mats
and
the
earlier
obser-
vations
of
Brock
(7),
suggest
that
the
primary
photic
zone
is
restricted
to
the
upper
0.5
to
1
mm
of
mat.
Since
Chloroflexus
apparently
is
the
structural
component
of
the
mat
and
ac-
counts
for
the
growth
of
the
mat,
the
growth
region
probably
coincides
with
the
region
where
Chloroflexus
is
most
active.
Diurnal
growth
and
response
of
Chloro-
flexus
to
reduce
02
concentrations.
Vertical
growth
of
the
microbial
mats
may
be
partly
the
result
of
the
upward
migration
of
Chloro-
flexus.
Evidence
has
been
presented
suggesting
that
Chloroflexus
migrates
vertically
at
night
or
when
light
is
reduced
or
eliminated
artifi-
cially
by
filters
or
covers
(16).
The
upward
migration
could
be
a
response
to
low
light
or
to
reduced
02,
since
under
the
low-light
conditions
the
algae
are
not
photosynthesizing
and
hence
will
not
produce
02.
These
alternative
hy-
potheses
were
tested
by
incubating
cores
col-
lected
from
the
Octopus
Spring
mat
in
vials
containing
either
spring
water
or
medium
D
(11)
to
which
in
some
instances
a
suspension
of
DCMU
had
been
added
to
a
final
concentration
of
10-5
M.
Vials
containing
cores
were
also
incu-
bated
without
DCMU
in
the
light
and
dark.
Since
DCMU
inhibits
algal
production
of
02,
this
experiment
should
test
the
response
of
Chloroflexus
to
reduced
02.
The
cores
were
incubated
for
from
1
to
several
days
at
56°C
in
a
stream
from
Octopus
Spring.
Within
1
day,
the
surfaces
of
the
cores
incubated
in
dark
vials
and
in
light
or
dark
with
DCMU
were
orange-red,
due
to
the
surface
accumulation
of
Chloroflexus,
whereas
the
surface
of
cores
in-
cubated
in
light
vials
without
DCMU
remained
a
light
green.
The
oxygen
concentration
in
the
DCMU
vials
incubated
in
the
light
was
about
0.7
mg/liter,
compared
with
1.5
mg/liter
in
the
controls
without
DCMU.
Algal
incorporation
of
[14C]bicarbonate
in
the
DCMU
vials
did
not
occur.
It
is
apparent
from
these
observations
that
the
absence
of
algal
02
production
rather
than
the
light
stimulates
the
migration
of
Chlo-
roflexus.
The
production
of
oxygen
was
tested
indi-
TABLE
4.
Photosynthetic
efficiency
of
Synechococcus
and
Chloroflexus
in
the
various
layers
of
a
microbial
mata
Uptake
of
['4C]bicarbonate
Layer
Algal
photosynthesis
Bacterial
photosynthesis
cpm/U
of
Chl
a
cpm/gg
of
protein
cpm/U
of
Bchl
c
cpm/.ug
of
protein
Nodules
on
mat
surface
................
53,500
50.3
32,100
6.55
Blue-green
upper
layer
.92,300
60.1
34,100
14.2
Orange
layer
below
blue-green
layer
None
1.11
6,780
11.4
Pale
orange
layer
.None
None
1,580
4.4
a
Nodules,
0.2-
to
0.5-mm
projections
from
the
mat
surface,
were
harvested
from
the
Octopus
Spring
mat
by
aspiration
with
a
Pasteur
pipette
fitted
with
a
rubber
bulb.
The
upper
1.5-mm
portion
of
a
microbial
core,
lacking
nodules,
was
sectioned
into
three
0.5-mm
layers:
a
blue-green
layer,
an
orange
layer,
and
a
pale
orange
layer.
Photosynthesis
was
measured
on
homogenates
of
each
layer
as
described
in
the
text.
Abbreviations:
Chl
a,
Chlorophyll
a;
Bchl
c,
bacteriochlorophyll
c.
VOL.
34,
1977
444
DOEMEL
AND
BROCK
rectly
by
measuring
the
oxygen
in
the
water
over
the
algal
mat
at
6:00
a.m.
and
at
10:00
a.m.
At
6:00
a.m.,
0.14
to
0.17
mg
of
02
per
liter
was
detected,
and
at
10:00
a.m.,
0.30
mg/
liter
was
detected.
Clearly
oxygen
levels
are
low,
lower
than
in
most
aquatic
habitats
at
normal
temperatures.
However,
although
02
was
low
in
the
water,
in
a
dense
algal
mat,
oxygen
levels
in
the
microenvironment
at
the
surface
may
be
quite
high.
Glass
slides
placed
vertically
into
the
mat
were
rapidly
colonized
and
provided
a
means
for
estimating
the
migration
rate
of
Chloro-
flexus.
On
slides
present
in
the
spring
for
5
days,
vertically
oriented
filaments
were
pres-
ent
to
about
13.5
mm
above
the
surface
of
the
mat.
Assuming
that
cells
present
at
that
level
migrated
from
the
mat,
the
migration
rate,
0.11
mm/h,
is
similar
to
the
rate
of
gliding
by
Chloroflexus
on
agar
(28).
Mechanism
of
mat
growth.
Observations
of
the
Octopus
Spring
mat
over
many
years
have
shown
that
the
thickness
of
the
mat
has
re-
mained
constant.
However,
if
the
mat
is
in
steady
state,
the
growth
rate
will
equal
the
death
rate,
and
no
change
will
be
obvious.
To
study
the
growth
of
the
microbial
mat,
the
growth
region
must
be
isolated
in
some
fashion
from
the
remainder
of
the
mat.
After
a
number
of
preliminary
studies,
it
was
found
that
the
most
suitable
substratum
for
the
study
of
mat
growth
was
silicon
carbide
(the
use
of
silicon
carbide
was
suggested
by
M.
Walter).
When
placed
onto
the
mat,
silicon
carbide
was
rapidly
colonized
by
Chloroflexus
from
beneath.
Within
4
to
5
h,
orange
material
could
be
seen
on
the
surface.
The
only
observable
effect
of
the
silicon
carbide
was
that
the
mat
surface
sometimes
appeared
to
be
darker
in
color
than
the
surrounding
mat
and
could
be
readily
dis-
tinguished
from
the
surrounding
mat
for
sev-
eral
weeks.
However,
the
mats
above
the
sili-
con
carbide
were
level
with
the
surrounding
mat,
suggesting
that
growth
was
not
stimu-
lated.
As
a
result
of
continued
growth,
the
silicon
carbide
became
buried
deep within
the
mat
(Fig.
14)
and
served
as
an
excellent
marker
of
mat
growth
and
decomposition
(see
below).
In
many
stromatolitic
systems,
the
active or
passive
deposition
of
inorganic
material,
partic-
ularly
silicates
or
calcium
carbonate,
contrib-
utes
significantly
to
the
increase
in
mat
thick-
ness
(20).
Although
siliceous
sinter
is
always
a
component
of
the
microbial
mat
at
Octopus
Spring,
it
appears
to
be
a
minor
one
(see
Fig.
6).
The
absence
of
a
significant
inorganic
con-
tribution
would
be
suggested
if
the
height
were
a
function
of
the
organic
matter
in
the
mat.
In
the
upper
2
mm,
there
was
a
direct
correlation
between
mat
height
and
protein
concentration;
in
mats
thicker
than
2
mm,
there
no
longer
was
a
correlation
(Fig.
15).
Not
only
does
this
observation
confirm
the
essential
organic
na-
ture
of
these
mats,
but
it
also
points
again
to
the
limits
of
the
growth
and
photic
zone,
since
height
would
not
necessarily
correlate
with
protein
in
the
lower
(decomposing)
system.
It
was
shown
earlier
that
essentially
no
light
penetrated
below
2
mm
in
the
mat.
Growth
rates.
The
rate
of
increase
of
the
mat
on
the
surface
of
silicon
carbide
should
provide
an
estimate
of
the
actual
growth
rate
if
washout
of
cells,
decomposition,
and
settle-
ment
are
minimal.
Silicon
carbide
was
placed
at
a
number
of
different
stations
in
the
Octopus
Spring
mat,
and
cores
were
collected
at
inter-
vals
(Fig.
16).
The
apparent
doubling
time
ranged
from
5
to 13
days,
with
an
apparent
optimum
temperature
for
growth
at
54°C.
The
relatively
rapid
increase
of
the
54
and
56°C
stations
followed
by
a
slow
period
of
increase
suggests
that
these
may
be
approaching
a
steady-state
level
where
growth
is
balanced
by
decomposition.
Longer-term
studies
extending
over
a
year
support
this
hypothesis
(Table
5).
Protein
con-
centration
appeared
to
be
maintained
at
a
rela-
tively
stable
level,
although
height
above
sili-
con
carbide
continued
to
increase.
A
similar
conclusion
can
also
be
drawn
from
Fig.
15,
which
shows
that
above
2
mm,
although
height
continued
to
increase,
the
protein
concentra-
tion
remained
relatively
stable.
These
observa-
tions
suggest
that
at
steady-state,
the
growth
rate
must
be
balanced
by
an
equivalent
rate
of
washout,
decomposition,
or
predator
consump-
cr
2000
Z
z
1000
0
0
1.
2.0
3.0
MAT
THICKNESS,
mm
4.0
FIG.
15.
Relationship
between
protein
concentra-
tion
and
thickness
of
mat.
Measurements
were
made
of
the
thickness
and
protein
concentration
of
the
microbial
mat
which
accumulated
on
the
surface
of
silicon
carbide
layers
in
various
mats
in
the
White
Creek
area
after
31
days.
Diameter
of
cores,
8.4
mm.
0
00
0
0
0
0
0
4?
1
APPL.
ENVIRON.
MICROBIOL.
ALGAL-BACTERIAL
MATS
IN
ALKALINE
HOT
SPRINGS
445
tion.
In
this
mat,
there
is
little
washout
and
eucaryotic
organisms
are
absent,
so
in
this
system
growth
can
only
be
balanced
by
bacte-
rial
decomposition.
The
optimal
growth
temperature
of
54°C
sug-
gested
by
Fig.
16
was
confirmed
by
short-tern
studies
at
more
temperatures
(Fig.
17).
The
data
points
were
obtained
by
measuring
the
height
over
silicon
carbide
after
31
days;
then
a
second
layer
of
silicon
carbide
was
placed,
and
the
height
over
this
layer
was
measured
after
31
days.
Assuming
that
the
rate
of
in-
crease
is
constant
during
this
period,
the
aver-
age
growth
rate
(in
micrometers
per
day)
was
calculated.
Despite
considerable
variability
in
growth
rates,
it
is
clear
that
the
optimum
rested
between
52
and
56°C.
In
other
Chloroflexus-Synechococcus
micro-
bial
mats
in
springs
and
streams
in
the
White
Creek
Valley
and
in
the
vicinity
of
the
Firehole
Lake
Loop
Road,
although
the
apparent
aver-
age
growth
was
sometimes
less
than
at
Octopus
Spring,
there
was
a
similar
temperature
opti-
DAYS
FIG.
16.
Growth
of
the
Octopus
Spring
mat.
Time
after
silicon
carbide
added.
Data
for
several
stations
at
each
temperature
have
been
pooled.
Apparent
doubling
times
as
calculated
from
the
graphs:
60°C,
10
to
11
days;
58°C,
13
days;
56°C,
6
days;
54°C,
5
days;
49°C,
10
days.
mum
(Table
6).
The
mats
at
Ravine
Spring,
Buffalo
Pool,
and
Sulfide
Spring
all
had
appar-
ent
growth
rates
considerably
less
than
those
at
either
Octopus
Spring
or
Twin
Butte
Vista.
In
the
Ravine
Springs
group,
the
efficiency
of
photosynthesis
was
also
reduced
compared
with
that
of
Octopus
Spring.
The
uptake
per
unit
of
chlorophyll
a
or
bacteriochlorophyll
c
at
90k
80k
70O
60k
ia50
E
:3
40
30
20O
10
o
40
70
50
60
TEMPERATURE
°C
FIG.
17.
Temperature
optimum
for
mat
growth.
Growth
rates
at
16
stations
in
Octopus
Spring
were
determined
for
two
growth
periods,
4
June
through
5
July
and
5
July
through
5
August
1974.
The
average
height
above
the
silicon
carbide
layer
was
determined
for
three
cores
from
each
site,
and
these
three
measurements
were
averaged
to
give
an
average
height
for
the
period.
A
constant
growth
rate
was
assumed,
and
the
increase
in
mat
height
(microme-
ters
per
day)
was
determined
for
each
period.
Each
point
on
this
graph
is
an
average
of
this
determina-
tion
for
the
two
periods.
The
bars
indicate
the
range
of
the
two
measurements.
The
temperatures
are
the
average
at
each
site
during
these
two
periods
and
varied
by
1
to
2°C.
TABLE
5.
Long-term
growth
of
the
Octopus
Spring
mata
Ht
above
silicon
Ag
of
protein/core
Ht
increas
Protein
increase
Station
Date
Days
carbide
(m)
(above
silicon
car-
(,n/day)
(pg/day)
carbide
(mm)
bide)
45-7
7/5/74-8/5/74
31
1.0
1,540
32
50
8/5/74-5/26/75
295
5.3
1,790
18
6.1
5
7/5/74-8/5/74
31
1.4
1,220
45
39
8/5/74-5/26/75
295
5.3
1,750
18
5.9
a
A
silicon
carbide
layer
was
placed
at
each
station
on
5
July
1974.
On
5
August
1974,
the
height
and
protein
of
the
microbial
mat
above
this
layer
of
silicon
carbide
was
measured
on
a
core
8.4
mm
in
diameter.
A
second
layer
of
silicon
carbide
was
placed
at
each
station
on
5
August
1974,
and
the
height
and
protein
were
determined
on
6
May
1975.
I
-
.
VOL.
34,
1977
4
1.--
..-
""I
0
I
i
TABLE
6.
Growth
rate
of
microbial
mats
in
alkaline
thermal
springs,
as
determined
by
silicon
carbide
marking
Spring
pH
of
pool
Temp
at
Range
of
growth
rate
(m/day)
Avg
growth
rate
(Am/
tion
(TC)
Ragdfarwhrt
(mdyy)
Buffalo
Pool
9.3
59
14-19
17
Spring
74-3
8.0
Main
pool
64
9.2-32
22
Main
pool
66
3.9-9.0
6.4
Spring
74-4
Main
pool
9.2-9.5
62
3.4-18
11
Ravine
Spring
8.0
Main
pool
63
15-28
23
Effluent
67
11-48
30
Effluent
62
17-26
23
Spring
74-6
7.7
Main
pool
65
18-23
21
Effluent
70
3.6
3.6
Effluent
62
13-14
14
Effluent
57
23
23
Twin
Butte
Vista
pool
8.0
60
15-29
22
Effluent
61
9.5-12
11
Effluent
57
33-55
44
Effluent
56
88-110
94
Serendipity
Spring
9.4
66
5.6
5.6
Effluent
58
13
13
Effluent
58.9
14
14
Effluent
60
27
27
Sulfide
Spring
6.8
Pool
46
(only
1
determination)
14
Pool
56
12
Pool
47
14
Effluent
48
14
Octopus
Spring
was
about
20
times
greater
than
at
spring
74-6
at
the
same
temperature.
Similar
lower
photosynthetic
efficiencies
were
observed
in
the
mats
in
other
of
the
Ravine
Springs
group.
Since
the
chemistry
of
these
springs
has
not
been
studied,
it
is
not
clear
what
factors
are
reducing
the
rate
of
growth.
Decomposition
of
the
microbial
mats.
The
in
situ
rates
of
decomposition
of
microbial
mats
were
determined
with
a
variation
of
the
proce-
dure
used
to
measure
the
growth
of
the
mat.
At
the
completion
of
a
study
of
mat
growth,
another
layer
of
silicon
carbide
was
placed
onto
the
mat
immediately
over
the
earlier
layer.
At
intervals,
further
cores
were
collected,
and
measurements
of
height
and
protein
were
made
both
on
the
mat
that
had
grown
on
the
surface
of
the
top
layer
of
silicon
carbide
and
on
the
mat
between
the
first
and
second
layers
of
silicon
carbide.
Decreases
in
protein
and
height
of
the
material
between
the
two
silicon
carbide
layers
provided
a
measure
of
decomposition
rate.
In
some
cases,
additional
layers
of
silicon
carbide
were
made.
Thus,
at
one
station,
growth
of
the
upper
microbial
mat
and
decom-
position
of
several
lower
layers
could
be
mea-
sured.
Figure
18
is
a
diagram
of
a
core
from
one
of
these
stations.
The
bands
of
microbial
mat
between
layers
1,
2,
and
3
have
decreased
in
size
considerably.
The
rate
of
decomposition
of
mat
over
a
period
of
1
year
is
illustrated
in
Fig.
19.
Layer
A
decreased
to
about
30%
of
its
original thick-
ness
in
about
a
year.
Initially
there
was
a
rapid
decrease
of
material,
with
an
apparent
half-time
of
about
1
month,
followed
by
a
slower
period
of
decompcisition,
with
an
appar-
ent
half-time
of
12
months.
This
same
biphasic
process
of
decomposition
was
observed
on
all
of
the
mats
examined:
a
rapid
initial
rate
of
decomposition
followed
by
a
slower
rate.
Note
also
that
two
thinner
layers
of
mat,
B
and
C,
no
longer
could
be
separately
distinguished
within
9
months.
Thus,
although
the
rate
of
decomposition
for
at
least
a
portion
of
the
material
was
quite
slow,
eventually
decompo-
sition
was
complete.
The
decomposition
of
the
Octopus
Spring
microbial
mat
appeared
to
be
a
function
of
temperature
(Fig.
20).
In
this
study,
the
decom-
position
rates
of
bands
of
mat
that
had
accu-
mulated
between
June
and
July
1974
were
measured
31
days
after
the
placement
of
a
covering
layer
of
silicon
carbide.
Fourteen
dif-
ferent
sites
on
the
Octopus
Spring
mat
with
temperatures
ranging
between
42
and
70°C
446
DOEMEL
AND
BROCK
APPL.
ENVIRON.
MICROBIOL.
ALGAL-BACTERIAL
MATS
IN
ALKALINE
HOT
SPRINGS
447
SAMPLED
AUGUST
1974
JULY
1974
16
14
JUNE
1974
12
E
10
UJ
8
z
I6
Imm
41
AUGUST
1973
2
APRIL
1972
AUGUST
1971
6
5
4
:3
2
IA
2
4
6
8
10
1214
MONTHS
B
A
B
A^
3
D
F
D
2
A
l
_-
4
5
5
26
7
JUNE
JULY
AUG
MAY
JULY
1974
1975
SAMPLING
DATES
FIG.
18.
Long-term
silicon
carbide
marking
of
a
mat.
This
is
a
diagram
of
a
core
collected
from
station
45-7,
Octopus
Spring,
on
5
August
1974.
Silicon
carbide
layers
(shaded
portions)
were
placed
on
the
mat
at
various
dates
as
indicated.
Bands
of
microbial
mat
(unshaded)
are
lettered.
The
upper
layer,
E,
accumulated
between
July
and
August
1974.
The
bands
of
silicon
carbide
were
about
0.7
mm
in
thickness.
The
thickness
of
all
layers
is
to
scale.
were
studied.
Although
there
was
considerable
scatter
of
points,
the
optimum
temperature
for
decomposition
appeared
to
be
between
52
and
560C.
In
all
of
these
studies
of
cores,
the
height
and
protein
content
of
the
mat
over
the
under-
lying
siliceous
sinter
appeared
to
remain
con-
stant,
at
least
during
a
year.
This
suggests
that
the
growth
rate
approximates
the
decom-
position
rate.
Table
7
summarizes
observations
of
decomposition
in
a
number
of
microbial
mats.
Considering
the
inherent
variability
of
the
mat,
it
does
appear
that
in
general,
growth
rates
approximate
decomposition
rates.
DISCUSSION
Model
of
mat
growth
and
formation.
Figure
21
summarizes
our
current
understanding
of
FIG.
19.
Long-term
decomposition
of
a
microbial
mat.
Silicon
carbide
layer
no.
1
was
placed
onto
the
Octopus
Spring
mat
in
August
1973;
layer
no.
2
was
placed
on
4
June
1974;
layer
no.
3
on
5
July
1974;
layer
no.
4
on
5
August
1974;
layer
no.
5
on
26
May
1975;
and
layer
no.
6
on
10
June
1975.
The
layers
of
mat
are
indicated
by
letters.
The
dates
on
the
graph
are
the
times
when
cores
were
taken
for
analysis.
The
rate
of
decrease
in
thickness
of
layer
A
is
shown
in
the
inset.
i
*
"I50-
20
U
h
40
40
50
60
70
TEMPERATURE
°C
FIG.
20.
Temperature
optimum
for
decomposition
rate
in
the
Octopus
Spring
mat.
The
decomposition
of
a
mat
layer
that
accumulated
between
5
June
1974
and
5
July
1974
was
measured
31
days
after
another
layer
of
silicon
carbide
was
added.
Assum-
ing
the
decomposition
rate
to
be
constant,
the
de-
crease
per
day
was
calculated.
6
c
VOL.
34,
1977
5
448
DOEMEL
AND
BROCK
TABLE
7.
Growth
and
decomposition
rates
for
a
large
number
of
stations,
measured
over
a
31-day
sampling
period
(pug
of
protein/core)
Growth
(j.g
of
protein/
Decomposition
(;g
of
pro-
core)
tein/core)
Net
growth
or
de-
Spring
Station
Temp
(°C)
composition
per
Increase
at
Increase/
Decrease
at
Decrease/
day
31
days
day
31
days
day
74-3
1-3
64
547
18
-356
-11
+7
Ravine
Spring
1-5
63
222
7
-236
-8
-1
2-7
62
388
13
-415
-13
0
2-8
51
354
11
-170
-6
+5
74-6
2-1
65
241
8
-520
-17
-9
2-10
62
203
7
-288
-9
-2
2-11
57
919
30
-419
-16
+14
Octopus
Spring
3-4
42
701
23
-399
-13
+10
5-1
65
905
29
-1,010
-33
-4
3-2
52
1,460
47
-1,130
-36
+11
45-8
54
4,290
138
-1,120
-36
+102
45-7
52
1,535
50
-955
-31
+20
45-6
49
1,020
33
-1,190
-38
-5
45-5
54
1,330
43
-1,350
-44
-1
3-1
56
970
31
-1,670
-54
-23
12
57
868 28
-955
-31
-3
5
54
1,220
39
-356
-11
+28
53
51
766
25
-917
-30
-5
6
52
1,350
43
-1,680
-54
-11
45-11
53
1,410
45
-120
-40
+5
3-5
54
1,540
50
-1,820
-59
-9
Twin
Butte
Vista
48-4
468
23
-841
-27
-4
48-2
255
13
-1,040
-52
-39
4-2
1,485
74
-1,380
-69
-5
3
965
48
-1,140
-57
-9
0
1,670
83
-1,070
-54
-29
4-3
1,640
82
-2,280
-117
-35
a
At
day
0,
8.4-mm-diameter
cores
were
taken,
and
a
new
layer
of
silicon
carbide
was
placed
onto
the
surface
of
mats
that
had
had
another
layer
applied
31
days
before.
After
another
31
days,
cores
were
taken,
and
protein
was
measured
on
the
mat
that
had
developed
on
top
of
the
second
silicon
carbide
layer,
to
provide
a
measure
of
growth
(protein
gain
column).
Protein
was
also
measured
on
the
first
layer
that
was
present
between
the
first
and
second
silicon
carbide
horizons.
The
decrease
in
protein
in
this
layer
from
day
0
gave
a
measure
of
decomposition
rate
(protein
loss
column).
the
manner
of
formation
of
these
alkaline
hot
spring
algal-bacterial
mats.
Beginning
with
a
new
surface,
there
is
an
initial
colonization
by
both
the
Chloroflexus
and
Synechococcus,
with
the
formation
of
a
thin,
microscopically
visible
layer.
It
is
possible
that
the
initial
colonization
of
a
new
substratum
by
Synechococcus
requires
the
presence
of
Chloroflexus
to
form
a
filamen-
tous
matrix,
as
suggested
by
Brock
(7),
but
in
relatively
quiet
waters
the
unicellular
alga
can
probably
develop
by
itself.
However,
soon
after
initial
colonization,
both
organisms
are
pres-
ent,
and
they
remain
together
throughout
the
growth
of
the
mat.
There
is
a
gradual
increase
in
thickness
of
the
mat
over
the
first
several
weeks,
until
the
thickness
of
the
mat
has
built
up
to
the
point
that
self-shading
of
the
Synecho-
coccus
occurs.
From
this
point
on,
the
thickness
of
the
Synechococcus
layer
remains
approxi-
mately
fixed.
In
light
too
dim
for
net
algal
photosynthesis,
the
Synechococcus
cells
die
and
lyse,
so
that
no
microscopically
recognizable
forms
are
present.
This
lysis
may
be
due
either
to
autolysis
or
to
attack
by
lytic
bacteria
pres-
ent
in
the
mat.
At
the
depths
where
light
limits
the
growth
of
Synechococcus,
there
is
still
sufficient
light
for
net
photosynthesis
by
Chloroflexus,
so
that
a
pure
Chloroflexus
un-
dermat
develops.
The
boundary
between
the
Synechococcus-Chloroflexus
upper
mat
and
the
pure
Chloroflexus
undermat
is
quite
distinct,
as
shown
by
the
sharp
color
transition
from
green
to
orange.
As
shown
by
Bauld
and
Brock
(4),
the
Chloroflexus
undermat
is
photosynthet-
ically
active,
and
the
ability
of
Chloroflexus
to
develop
at
lower
light
intensities
than
the
alga
leads
to
the
development
of
pure,
photosynthet-
ically
active
Chloroflexus
populations
at
appro-
priate
depths
in
the
mat.
Below
about
3
mm,
even
the
Chloroflexus
undermat
is
inactive
APPL.
ENVIRON.
MICROBIOL.
ALGAL-BACTERIAL
MATS
IN
ALKALINE
HOT
SPRINGS
449
E
E
E
E
F~~~~~~~~~~~~~~~~~~~~~~~
oa@j9
ii',"
d.ue0o
|
s~~~~~~~~~~~~~~~~~~
o
ussiD
efuejo
souejO
GIq,Mj
ID
c5
3:
0
B~~~~~~~~~~~~~~~~~~
0
*Ej
*
'E
SO.
c
E
o
J)
Bej
BUJ
1
M
w
UOOJD
e6ueJO
Bled~Ble
:~~~~~~~\\I
/:
.
.CNC
2
Z4
UJDe6uejo
UBBJ
E
cc
CoE
0:
VOL.
34,
1977
450
DOEMEL
AND
BROCK
photosynthetically
and
is
presumably
dead.
Mi-
croscopic
examination
of
the
material
below
3
mm
reveals
primarily
moribund
cells
and
empty
filaments.
Thus,
the
growth
of
the
Chlo-
roflexus
population
also
leads
eventually
to
self-shading.
(It
is
possible
to
obtain
pure
Chlo-
roflexus
populations
at
the
surface
of
mats
by
experimentally
reducing
the
light
intensity
to
values
low
enough
that
the
alga
cannot
develop
[16,
24].)
Eventually
a
mat
is
obtained
that
is
several
centimeters
thick
(Fig.
21),
and
further
change
is
not
seen.
At
this
point,
the
mat
is
in
steady
state,
with
continued
growth
being
balanced
by
decomposition.
Thus,
despite
the
dynamic
nature
of
the
mat,
thickness
remains
essen-
tially
constant
over
many
years.
Diurnal
changes
in
the
mat
are
dramatic.
During
the
day,
Synechococcus
and
Chloro-
flexus
develop
together,
but
at
night
Chloro-
flexus
shows
positive
aerotaxis
and
glides
up
on
top,
so
that
surface
accumulations
of
pure
Chloroflexus
occur.
During
the
following
day,
rapid
growth
by
the
Synechococcus
results
in
repopulation
by
the
alga
of
the
surface
of
the
mat.
We
know
from
the
culture
work
of
Dyer
and
Gafford
(18)
and
Peary
and
Castenholz
(26)
and
field
studies
by
Brock
and
Brock
(9)
that
the
Synechococcus
of
hot
springs
is
able
to
divide
very
rapidly,
up
to
eight
doublings
a
day.
The
doubling
time
of
Chloroflexus
has
not
been
measured
in
nature,
but
was
reported
by
Pierson
and
Castenholz
(28)
in
culture
to
be
seven
to
eight
doublings
per
day.
Thus,
the
organisms
maintain
continuously
growing
mixed
populations.
It
may
be
important
in
this
regard
that
Synechococcus
is
an
obligate
pho-
toautotroph
and
hence
is
able
to
produce
or-
ganic
carbon
only
during
the
day,
whereas
(at
least
in
culture)
Chloroflexus
can
grow
by
three
modes
of
nutrition,
photoautotrophic,
photohet-
erotrophic,
and
heterotrophic
(the
latter
only
when
02
is
present),
and
can
assimilate
organic
compounds
excreted
by
Synechococcus
(5).
Storms
frequently
wash
detrital
minerals
onto
the
mat.
As
shown
by
the
silicon
carbide
studies,
when
the
surface
is
covered
with
a
detrital
mineral,
within
hours
it
is
partially
or
completely
covered
with
Chloroflexus.
Soon
after,
the
presence
of
Synechococcus
is
noted.
Since
the
Synechococcus
of
these
hot
springs
is
not
motile,
it
is
likely
that
the
inoculum
for
the
Synechococcus
that
develops
on
top
of
the
silicon
carbide
layer
is
derived
from
the
small
number
of
cells
that
are
continually
present
in
the
water
over
the
mat.
Within
a
few
days
or
a
week,
the
mat
on
top
of
a
silicon
carbide
layer
resembles
a
normal
mat,
and
the
presence
of
the
mineral
cannot
be
discerned
without
cor-
ing.
Geological
relevance.
Although
it
is
un-
likely
that
precisely
similar
mats
developed
in
the
Precambrian,
we
can
view
the
Yellowstone
mats
as
models
for
the
kinds
of
events
that
might
have
occurred.
Two
major
conclusions
of
geological
relevance
derive
from
the
present
work.
(i)
Laminated
mats
can
be
formed
by
pho-
tosynthetic
bacteria
as
well
as
by
blue-green
algae.
It
has
been
conventionally
assumed
(3,
14,
15,
30-32)
that
the
presence
of
stromatolitic
rocks
in
Precambrian
formations
provides
evi-
dence
for
the
existence
of
blue-green
algae.
In
the
absence
of
microfossil
evidence,
this
conclu-
sion
seems
unwarranted.
Although
some
of
the
Precambrian
formations
possess
well-preserved
fossils
that
are
almost
certainly
blue-green
algae
(29),
most
of
the
formations
with
un-
equivocal
blue-green
microfossils
are
109
years
old
or
younger.
Of
the
older
formations,
includ-
ing the
now
famous
Gunflint
chert,
the
evi-
dence
that
the
microfossils
are
blue-green
algae
is
minimal.
Without
knowing
of
the
existence
of
Chloroflexus,
Cloud
(13)
first
suggested
that
some
of
the
microfossils
of
the
Gunflint
could
be
filamentous
bacteria.
Indeed,
the
filamen-
tous
organisms
from
the
Gunflint
and
similar
formations
(3,
13,
14)
are
strikingly
similar
to
Chloroflexus.
We
have
shown
in
the
present
work
that
Chloroflexus
is
responsible
for
holding
the
structure
of
these
laminated
mats
together.
Although
in
most
cases
this
filamentous
bacte-
rium
lives
only
together
with
the
blue-green
alga,
it
is
possible
by
experimentally
reducing
the
light
intensity
to
induce
the
formation
of
mats
consisting
of
pure
Chloroflexus.
Further,
Castenholz
(12)
has
shown
that
in
high-sulfide
springs,
blue-green
algal
development
is
sup-
pressed
by
sulfide,
and
pure
Chloroflexus
pop-
ulations
exist
naturally.
Thus,
mats
can
form
in
the
absence
of
blue-green
algae
and
could
have
formed
in
the
absence
of
blue-green
algae
in
the
Precambrian.
One
of
the
reasons
for
postulating
that
stromatolitic
rocks
had
a
pho-
tosynthetic
origin
was
that
the
vertical
orienta-
tion,
nodular
and
laminar
appearance,
and
other
features
of
these
structures
suggested
a
phototrophic
growth
process.
When
it
was
thought
that
the
only
photosynthetic
filamen-
tous
phototrophs
were
blue-green
algae,
it
was
natural
to
attribute
formation
of
stromatolites
to
them.
The
existence
of
Chloroflexus
and
the
demonstration
that
it
can
form
stromatolitic
structures
alters
the
need
for
this
interpreta-
tion.
APPL.
ENVIRON.
MICROBIOL.
ALGAL-BACTERIAL
MATS
IN
ALKALINE
HOT
SPRINGS
451
Our
results
emphasize
the
critical
impor-
tance
of
developing
criteria
for
evaluating
the
microbiology
of
Precambrian
fossiliferous
rocks.
Simply
relating
observations
on
such
rocks
to
presumed
evolutionary
sequences,
without
objective
evidence,
is
of
little
value
and
may
introduce
many
sources
of
confusion
into
the
interpretation
of
the
fossil
record.
We
do
not
necessarily
feel
that
Chloroflexus
itself
was
responsible
for
the
formation
of
Precam-
brian
stromatolitic
formations.
Its
existence,
however,
shows
that
some
sort
of
photosyn-
thetic
bacterium
could
have
been
the
prime
agent
in
the
formation
of
stromatolites.
This
itself
should
provide
sufficient
reason
for
reex-
aIining
present
and
previous
interpretations.
Since
photosynthetic
bacteria
are
anoxygenic,
our
data
suggest
caution
in
concluding
from
biological
evidence
alone
the
time
at
which
photosynthetic
oxygen
evolution
first
occurred.
(ii)
Complete
decomposition
of
organic
matter
can
occur
under
anaerobic
conditions
and
in
the
absence
of
grazing
animals.
Many
of
the
interpretations
of
how
organic
matter
accumulates
in
the
geosphere
assume
that
de-
composition
is
inhibited
under
anaerobic
con-
ditions.
This
is
standard
reasoning
when
dis-
cussing
the
origin
of
oil
and
coal
(22).
Grazing
animals
have
been
assumed
to
play
an
impor-
tant
role
in
destruction
of
stromatolitic
struc-
tures
and
mats
(2,
33).
Our
data
provide
the
first
clear
evidence
of
complete
decomposition
of
cellular
organic
matter
under
anaerobic
con-
ditions.
Thus,
grazing
animals
are
not
required
for
decomposition,
although
they
greatly
speed
up
the
rate
(7).
The
technique
of
silicon
carbide
marking
has
permitted
long-term
measurements
of
in
situ
decomposition
rates.
The
thickness
and
protein
content
between
successive
silicon
carbide
lay-
ers
decrease
with
time,
and
the
rate
of
decrease
is
a
measure
of
the
decomposition
rate.
Ulti-
mately,
two
silicon
carbide
layers
merge
into
one.
We
have
also
used
a
pine
pollen
horizon,
added
to
the
mats
around
the
end
of
June
each
year,
as
a
further
measure
of
decomposition
(see
lower
stratum
of
Table
1).
The
decomposition
rates
measured
in
this
study
always
followed
the
same
pattern:
ini-
tially
there
was
a
rapid
rate
of
decomposition
for
the
first
2
to
4
week,
followed
by
a
consider-
ably
slower
rate
that
lasted
through
the
subse-
quent
year.
Presumably,
the
initial
rapid
rate
reflects
the
breakdown
of
the
most
readily
decomposable
materials,
and
the
slower
subse-
quent
rate
is
due
to
the
breakdown
of
more
stable
substances.
As
far
as
we
can
tell,
there
is
no
organic
material
remaining
after
about
a
year
of
decomposition,
and
so
the
process
goes
essentially
to
completion.
This
does
not
neces-
sarily
mean
that
all
of
the organic
carbon
of
the
mats
has
been
mineralized,
since
some
diffusion
of
partially
degraded
materials
from
the
mat
into
the
water
may
occur.
However,
because
of
the
compact
nature
of
these
mats,
diffusion
is
probably
minimal,
so
that
to
a
first
approximation
we
can
consider
that
decompo-
sition
occurs
completely
in
situ.
It
is
likely
that
decomposition
is
completely
an
anaerobic
process.
Our
data
on
the
micro-
distribution
of
sulfide
within
these
mats
(17)
show
that
sulfide
is
present
below
about
3
mm
from
the
surface
of
the
mat.
Since
sulfide
is
not
stable
in
the
presence
of
oxygen,
we
can
conclude
that
conditions
are
anaerobic
below
this
point,
and
since
3
mm
is
just
about
the
end
of
the
photic
zone,
it
can
be
concluded
that
virtually
all
of
the
decomposition
that
occurs
is
anaerobic.
Because
of
the
high
temperature
of
these
mats,
all
eucaryotic
organisms
are
ab-
sent,
including
grazing
animals,
so
that
decom-
position
must
occur
as
a
result
of
an
anaerobic
food
chain.
Since
methane
is
produced
in
these
mats
and
thermophilic
methanogenic
bacteria
have
been
isolated
(J.
G.
Zeikus
and
David
Ward,
personal
communications),
complete
mineralization
of
the
organic
carbon
to
meth-
ane
and
carbon
dioxide
probably
occurs.
It
is
true
that
anaerobic
food
chains
are
well
known
in
sewage
digestion
and
rumen
fermen-
tation
(21),
but
in
neither
of
these
habitats
does
the
process
go
to
complete
mineralization.
In
sludge
digestors,
there
is
a
vast
residue
of
undigested
material
which
must
be
disposed
of,
and
in
the
rumen
the
fermentation
acids
formed
are
removed
into
the
blood
stream
and
transported
to
aerobic
tissues
for
oxidation.
Input-output
balances
have
not
been
carried
out
in
marine
or
freshwater
sediments,
where
it
may
be
presumed
that
anaerobic
food
chains
also
exist.
A
number
of
organic
materials
formed
by
living
organisms
are
not
completely
mineral-
ized
anaerobically.
Thus,
lignin,
polyaromatic
rings,
porphyrins,
aliphatic
hydrocarbons,
and
other
constituents
almost
certainly
do
not
de-
compose
completely
under
anaerobic
conditions
(6).
Any
of
these
substances
formed
by
the
organisms
in
the
mats
we
have
studied
must
be
in
such
small
amounts
that
they
do
not
contribute
to
the
mass
of
the
mat.
Unfortu-
nately,
we
were
not
able
to
carry
out
specific
assays
for
individual
compounds
of
these
types,
so
we
cannot
state
that
they
are
not
present.
But
from
the
viewpoint
of
the
formation
of
an
organic
structure,
which
might
become
pre-
VOL.
34,
1977
452
DOEMEL
AND
BROCK
served
and
fossilized,
our
work
shows
that
anaerobic
conditions
alone
are
not
sufficient.
Further
work
on
the
biochemistry
and
biogeo-
chemistry
of
decomposition
in
these
hot
spring
mats
would
be
very
desirable.
ACKNOWLEDGMENTS
This
work
could
not
have
been
done
without
the
diligent
assistance
of
the
following
undergraduate
students:
Donald
Weller,
James
Cook,
and
Timothy
Parkin,
all
of
Wabash
College,
and
Patrick
Remington
of
the
University
of
Wis-
consin.
The
work
was
supported
by
the
College
of
Agricultural
and
Life
Sciences,
by
a
research
grant
from
the
National
Science
Foundation
(GB-35046),
and
by
a
research
contract
from
The
Energy
Research
and
Development
Agency
(COO-
2161-32).
Discussions
with
Malcolm
Walter
provided
the
impetus
for
the
silicon
carbide
marking
experiments,
and
we
also
thank
him
for
making
the
markings
in
1971
and
1972.
LITERATURE
CITED
1.
American
Public
Health
Association.
1971.
Standard
methods
for
the
examination
of
water
and
wastewa-
ter,
13th
ed.
American
Public
Health
Association,
Inc.,
New
York.
2.
Awramik,
S.
M.
1971.
Precambrian
columnar
stroma-
tolite
diversity:
reflection
of
metazoan
appearance.
Science
174:825-827.
3.
Barghoorn,
E.
S.,
and
S.
A.
Tyler.
1965.
Microorga-
nisms
from
the
Gunflint
chert.
Science
147:563-577.
4.
Bauld,
J.,
and
T.
D.
Brock.
1973.
Ecological
studies
on
Chloroflexus,
a
gliding
photosynthetic
bacterium.
Arch.
Mikrobiol.
92:267-284.
5.
Bauld,
J.,
and
T.
D.
Brock.
1974.
Algal
excretion
and
bacterial
assimilation
in
hot
spring
algal
mats.
J.
Phycol.
10:101-106.
6.
Breger,
I.
A.
(ed.).
1963.
Organic
geochemistry.
Mac-
millan,
New
York.
7.
Brock,
T.
D.
1969.
Vertical
zonation
in
hot
spring
algal
mats.
Phycologia
8:201-205.
8.
Brock,
T.
D.,
and
M.
L.
Brock.
1967.
The
measurement
of
chlorophyll,
primary
productivity,
photophospho-
rylation,
and
macromolecules
in
benthic
algal
mats.
Limnol.
Oceanogr.
12:600-605.
9.
Brock,
T.
D.,
and
M.
L.
Brock.
1968.
Measurement
of
steady-state
growth
rates
of
a
thermophilic
alga
di-
rectly
in
nature.
J.
Bacteriol.
95:811-815.
10.
Brock,
T.
D.,
and
M.
L.
Brock.
1969.
Effect
of
light
intensity
on
photosynthesis
by
thermal
algae
adapted
to
natural
and
reduced
sunlight.
Limnol.
Oceanogr.
14:334-341.
11.
Castenholz,
R.
W.
1969.
Thermophilic
blue-green
algae
and
the
thermal
environment.
Bacteriol.
Rev.
33:476-
504.
12.
Castenholz,
R.
W.
1973.
The
possible
photosynthetic
use
of
sulfide
by
the
filamentous
phototrophic
bacte-
ria
of
hot
springs.
Limnol.
Oceanogr.
18:863-876.
13.
Cloud,
P.
E.
1965.
Significance
of
the
Gunflint
(Pre-
cambrian)
microflora.
Science
148:27-35.
14.
Cloud,
P.,
and
G.
R.
Licari.
1972.
Ultrastructure
and
geologic
relations
of
some
two-aeon
old
nostocacean
algae
from
northeastern
Minnesota.
Am.
J.
Sci.
272:138-149.
15.
Cloud,
P.
E.,
and
M.
A.
Semikhatov.
1969.
Proterozoic
stromatolite
zonation.
Am.
J.
Sci.
267:1017-1061.
16.
Doemel,
W.
N.,
and
T.
D.
Brock.
1974.
Bacterial
stromatolites:
origin
of
laminations.
Science
184:1083-
1085.
17.
Doemel,
W.
N.,
and
T.
D.
Brock.
1976.
Vertical
distri-
bution
of
sulfur
species
in
benthic
algal
mats.
Lim-
nol.
Oceanogr.
21:237-244.
18.
Dyer,
D.
L.,
and
T.
D.
Gafford.
1961.
Some
character-
istics
of
a
thermophilic
blue-green
alga.
Science
134:616-617.
19.
Geitler,
L.
1932.
Cyanophyceae.
Akadamie
Verlagsge-
sellschaft
m.b.H.,
Leipzig
(Johnson
Reprint
Corp.,
N.Y.,
1971).
20.
Golubic,
S.
1973.
The
relationship
between
blue-green
algae
and
carbonate
deposits,
p.
434-472.
In N.
G.
Carr
and
B.
A.
Whitton
(ed.),
The
biology
of
blue-
green
algae.
Blackwell
Scientific
Publications,
Ox-
ford.
21.
Hobson,
P.
N.,
S.
Bousfield,
and
R.
Summers.
1974.
Anaerobic
digestion
of
organic
matter.
Crit.
Rev.
Environ.
Control
4:131-191.
22.
Krauskopf,
K. B.
1967.
Introduction
to
geochemistry.
McGraw-Hill,
New
York.
23.
Madigan,
M.
T.,
and
T.
D.
Brock.
1976.
Quantitative
estimation
of
bacteriochlorophyll
c
in
the
presence
of
chlorophyll
a
in
aquatic
environments.
Limnol.
Oceanogr.
21:462-466.
24.
Madigan,
M.
T.,
and
T.
D.
Brock.
1977.
Adaptation
of
hot
spring
phototrophs
to
reduce
light
intensity.
Arch.
Microbiol.
113:111-120.
25.
Peale,
A.
C.
1883.
The
thermal
springs
of
Yellowstone
National
Park,
p.
63-386.
In
12th
Annual
Report
of
the
U.S.
Geological
and
Geographical
Survey
of
the
Territories
for
the
Year
1878.
Part
II.
U.S.
Govern-
ment
Printing
Office,
Washington,
D.C.
26.
Peary,
J.
A.,
and
R.
W.
Castenholz.
1964.
Temperature
strains
of
a
thermophilic
blue-green
alga.
Nature
(London)
220:720-721.
27.
Pfennig,
N.
1969.
Rhodopseudomonas
acidophila,
sp.
nov.,
a
new
species
of
the
budding
purple
nonsulfur
bacteria.
J.
Bacteriol.
99:597-602.
28.
Pierson,
B.
K.,
and
R.
W.
Castenholz.
1974.
A
photo-
trophic
gliding
filamentous
bacterium
of
hot
springs,
Chloroflexus
aurantiacus
gen.
and
sp.
nov.
Arch.
Microbiol.
100:5-24.
29.
Schopf,
J.
W.
1968.
Microflora
of
the
Bitter
Springs
formation,
late
Precambrian,
Central
Australia.
J.
Paleontol.
42:651-688.
30.
Schopf,
J.
W.
1970.
Precambrian
micro-organisms
and
evolutionary
events
prior
to
the
origin
of
vascular
plants.
Biol.
Rev.
45:319-352.
31.
Schopf,
J.
W.,
and
E.
S.
Barghoorn.
1967.
Alga-like
fossils
from
the
early
Precambrian
of
South
Africa.
Science
156:508-512.
32.
Schopf,
J.
W.,
D.
Z.
Oehler,
R.
J.
Horodyski,
and
K.
A.
Kvenvolden.
1971.
Biogenicity
and
significance
of
the
oldest
known
stromatolites.
J.
Paleontol.
45:477-
485.
33.
Stanely,
S.
M.
1973.
An
ecological
theory
for
the
sudden
origin
of
multicellular
life
in
the
late
Precambrian.
Proc.
Natl.
Acad.
Sci.
U.S.A.
70:1486-1489.
34.
Walter,
M.
R.
1972.
A
hot
spring
analog
for
the
deposi-
tional
environment
of
Precambrian
iron
formations
of
the
Lake
Superior
region.
Econ.
Geol.
67:965-980.
35.
Walter,
M.
R.
(ed.).
1976.
Stromatolites.
Elsevier,
Am-
sterdam.
36.
Walter,
M.
R.,
J.
Bauld,
and
T.
D.
Brock.
1972.
Siliceous
algal
and
bacterial
stromatolites
in
hot
spring
and
geyser
effluents
of
Yellowstone
National
Park.
Science
178:402-405.
37.
Walter,
M.
R.,
J.
Bauld,
and
T.
D.
Brock.
1976.
Microbiology
and
morphogenesis
of
columnar
stro-
matolites
(Conophyton,
Vacerrilla)
from
hot
springs
in
Yellowstone
National
Park,
p.
273-310.
In
M.
R.
Walter
(ed.),
Stromatolites.
Elsevier,
Amsterdam.
38.
Weed,
W.
H.
1889.
Formation
of
travertine
and
sili-
ceous
sinter
by
the
vegetation
of
hot
springs.
U.S.
Geol.
Survey
Rep.
9:613-676.
APPL.
ENVIRON.
MICROBIOL.
