GROWTH
OF MARINE
BACTERIA
AT LIMITING
CONCENTRATIONS
OF ORGANIC
CARBON IN SEAWATER1
Holger
Woods Hole Oceanographic
W. Jannasch
Institution,
Woods Hole, Massachusetts
02543
ABSTRACT
Growth responses of several species of heterotrophic
marine bacteria to limiting
conccntrations
of lactate, glycerol, and glucose in seawater were determined
in a chemostat.
In all casts, threshold concentrations
of the limiting
substrates were found below which
the organisms were unable to grow. This phenomenon
is explained
on the basis of a
positive feedback mechanism,
which is abolished below a certain minimum
population
density.
The resulting inhibitory
effect on growth leaves a corresponding
concentration
of the limiting substrate unattacked.
According to their growth parameters, two types of spccics could bc distinguished,
one
adapted to the marine environment
by its ability to grow at low substrate concentrations
and one inactive in natural seawater but surviving.
The implications
of these results on the rates of microbial
transformations
and on the
occurrence and concentrations
OF dissolved organic material in the sea are discussed.
INTRODUCTION
Rates of microbial activities in the sea
arc of essential interest in biological oceanography. Although detailed information on
physiological and biochemical properties of
various marine microorganisms under Zubom&ory conditions is available, there are no
quantitative
data describing
growth in
nutrient
concentrations
characteristic
of
their nutuml environment.
While recent
studies on microbial growth in the sea have
been focused largely on stimulating or inhibiting substances, this study is concerned
with populatioa behavior based on nutritional growth limitation.
Natural seawater is characterized by extremely low nutrient concentrations, compared with media commonly used for
studies of bacterial growth. Most results
obtained by the classical techniques for
assessing microbial activities in such impoverished environments are not amenable
to a meaningful interpretation.
This study
represents an attempt to consider two
major problems :
1) The turnover rates of organic matter
in seawater are usually too low to permit
direct measurements of specific microbial
activities. When substrates are added in
* Contribution
No. 1882 from the Woods Hole
Oceanographic
Institution.
This work was supported by National Science Foundation
Grants GB
5199 and GB 861.
measurable concentrations,
the resulting
turnover rates cannot be extrapolated to
values presumably valid at lower substrate
concentrations.
This has been shown by
studies on regulatory mechanisms in microbial cells. Kjeldgaard (1963) has shown
that biosynthetic abilities diminish as the
growth rate decreases. In turn, minimum
rcquiremcnts
of the bacterial
ccl1 are
known to vary with the synthetic ability
( IIcrbert 1961cl). Holme ( 1958) and Herbert (1961n) reported that under pronounced
growth limitation the ratio between respiratory and biosynthetic
metabolism
is
strongly affected by the nature of the limiting substrate, such as its function as a
source of energy or as an essential nutrient.
Thus, microbial activities measured after
the addition of a nutrient can be interpreted
only in terms of potentialities.
2) Incubating water samples of limited
volume changes drastically
the environmental conditions for growth of microorganisms. The natural habitat represents
an open system where the environmental
conditions do not alter with time. Immediatcly after sampling, the population
becomes affected by the inherent properties
of the closed growth system (batch culturc). Depending upon the volume of the
sample, the natural population will react
to the decrease of the limiting substrate, to
shifts from one limiting factor to another,
GROWTH7
OF BACTERIA
AT LOW
and to the increase of metabolic products.
Characteristic
of this situation is the immediate change of colony numbers as determined on a complex agar medium and of
the species composition as determined on
specific media (ZoBcll and Anderson 1936;
Potter 1960). Th ese effects are less pronounccd in natural waters with high turnover rates and high population densities.
In natural waters, exhaustion of the
limiting
nutrient
and accumulation
of
metabolic products are assumed to be
eliminated largely by metabolic intcractions between species of different metabolic types within the natural population.
These interactions, which contribute to the
more or less steady state of the population,
are destroyed when the sample is transferred to a closed system of relatively
small volume. The technique of employing
plastic bags submerged in the natural
habitat ( McAllister et al. 1961) represents
an interesting compromise. It considers the
delicate balance of natural populations
while still taking advantage of conducting
quantitative measurements in a closed systcm.
In short, most techniques for the detcrmination of a microbial reaction in samples
of natural water suffer from disregarding
the mixed microbial population as a dynamic sys tern. Under experimental steadystate conditions of a chemostat, these difficultics are eliminated to a certain degree
by the use of time-independent relationships
between culture conditions
and growth
propertics of the organisms. Comparative
studies of microbial growth in both closed
and open culture systems (Herbert, Elsworth, and Telling 1956; Pfennig and Jannasch 1962; Jannasch 1965b ) demonstrate
their applicability
in microbial ecology. In
this paper the properties of a chemostat
culture are used to, assess growth responses
of some marinc bacteria to extremely low
concentrations of the limiting substrate in
seawater.
THKORETICAL
The chemostat is a refined continuous
culture system designed to meet the tech-
SUBSlXATE
265
LEVELS
nical requirements for the establishment of
a steady state. The steady state is generally
defined as a time-independent
state where
growth is counterbalanced by the removal
of cells; that is, the exponential growth rate
,I.L( in hr-l ) is equal to the exponential dilutio,n rate D. Consequently, the population
density x ( in mg dry wt/liter ) and
the concentration
of the limiting
substrate in the culture S (in mg/liter)
remain
constant. The mathematical
relationship
between culture conditions,
growth response, and growth parameters ( Herbert
1961b) apply during steady state only.
Mixed populations in the chemostat undergo selective processes. The species
achieving the fastest growth rate under the
particular conditions will approach a steadystate population, while the competing species are displaced (Powell 1958; Jannasch
1965a). Thus, a chemostat population cannot bc considered a model of a natural,
mixed population.
Contois and Yango
( 1964) and Bungay ( 1966) discussed experiments where steady states in mixed
populations of two species could bc obtaincd.
For the current study, it is of importance
that during steady state the substrate conccntration in the chemostat S is dependent
on the growth rate ,L but independent of
the concentration of the limiting substrate
in the reservoir so. The population density
3 is in proportion to so:
so=++
(G-&)> (1)
an d
s = K,D/(,um-
D),
(2)
where pnt is the maximum growth rate at
excess concentration
of the limiting substrate; K, is the concentration of the limiting substrate that gives rise to one-half of
the maximum growth rate (saturation constant); and y is the weight of limiting substratc consumed/weight
of cell substance
produced ( yield coefficient ) .
The thcorctical
relationship
between
population density and substrate concentration during steady state is shown as a
266
HOLGER
W. JANNASCH
so= 1.0
0.5
:
I
2
> 0.25 0
1%
:.*...
s,=o.5
!’
_._.-.
-.-.-.-./
0.5
.A’
1.0
D lhr-‘/
FIG.
1.
Theoretical
steady-state relationship
between population
density
(solid lint)
and concentration
of limiting
substrate (broken line) in
relationship
to the dilution rate and at two concentrations
of the limiting
substrate in the reservoir,
according
to equation
( 1).
Crowth
parameters are : pna = 1.0 hr-‘, K, = 0.04 g/liter,
y = 0.5 (after Herbert et al. 1956).
function of the dilution rate in Fig. 1 for
two substrate concentrations in the rcservoir. It illustrates that, theoretically,
s is
independent of so, while the 3 is proportional to so. This relationship was checked
in the following experiments.
EXPERIMENTAL
The technical requirement
for growing
microorganisms at steady state is maintaining the constancy of: medium flow, volume
of culture, medium composition, temperature, and the degree of mixing and aeration.
For this study, low and accurate flow rates
for long periods of time were especially
important. This was achieved by a gravity
flow system ( Fig. 2)) which proved to be
superior to all pump systems tested. This
rather simple arrangement permitted feeding a number of chemostats from one mcdium reservoir and facilitated
series of
The pressure
simultaneous experiments.
heads were counterbalanced by a capillary
resistance of Teflon tubing (diameter OS0.8 mm, length 2-8 m) placed in a water
bath of constant temperature.
This resistance damped slight pressure changes in
the chemostat due to the continuous rcmoval of culture medium by the acrating
gas mixture. Positive pressure in the culture vessel provided a rapid removal of the
medium through a capillary and prevented
The inlet tube was
back-contaminations.
heated to 80-9OC with a 0.1 ohm/cm ni-
FIG. 2. Chemostat as used in this study (up to
four of thcsc units were used simultaneously).
Symbols (seauence following
the medium flow) :
i, reservoir (-10 to 20 liters ); P, pump for keeping
overflow
vessels, Ov, (pressure
heads)
filled;
P’, peristaltic
pump, used for short time experiments and high flow rates only; Ov’, second overflow vessel for second chemostat unit: D, drop
counting device: L. caoillarv resistance in constant
temperature
w&e,’ baih; d, and pH, measuring
and recording device for dissolved oxygen and ‘pH;
Tc, thermoregulator;
Va, variac; I-1, submersible
heater (quartz);
Pr, temperature
probe; Th, thermometcr;
A, aeration device; T, Telfon-covered
lid; C, culture vessel ( 150 to 1,200 ml); C?, tube
connection to second culture vessel; M, magnetic
stirrer; St, stirring bar; h, niveau difference.
chrome resistance wire 15 cm long to prevent contamination from the culture.
The culture vessel was closed with a
Teflon-covered
stainless steel lid 2.5 cm
thick. Tapered openings received silicone
stoppers with bores for tubing and electrodes. In order to avoid zones of stationary growth, the air space above the surface
of the culture ( splash zone) was kept as
GROWTH
OF BACTERIA
AT LOW
small as possible and self-rinsing. According to the analytical procedures to follow,
Dry Ice or a disinfectant was used in the
receiving vessel.
The dilution
rates of the chemostats
could be independently
adjusted by a)
lifting or lowering the overflow vessel, b )
choice of culture volume, and c) changing
the length of the capillary resistance. The
dilution rate was measured periodically in
the receiving vessel. Its constancy was
checked occasionally in the drop counting
device. Rubber membrane pumps and ball
flowmcters were used for aeration. The
air was sterilized in a series of dry filters
(cotton and glass wool) and then water
saturated.
As the medium, prefiltered offshore seawater was filter-sterilized
and supplemented
with either Na-lactate, glycerol, or D-ghcase in amounts of 0.5 to 100 mg/liter.
Phosphate buffer (7.8 pH) and ammonium
chloride were added to provide an initial
C : N : P ratio of 10 : 4 : 1 by weight in
order to secure growth limitation by the
carbon and energy source All experiments
were done at 20C.
When B could not be measured directly,
the growth parameters were determined as
described by Jannasch (1963a) for the indirect calculation of S. Enzymatic analyses
were applied whenever known procedures
( Bergmeyer 1963) could be adapted to
seawater. Oxygen saturation was controlled
by using a stationary gold amalgam-zinc
electrode system (Tijdt 1958) which was
periodically exchanged when instability occurred due to calcium carbonate prccipitations.
The following strains were isolated either
from chemostat enrichments
(Jannasch
1965u) or from batch culture enrichments
( Vaccaro and Jannasch 1966) : Achromobatter aquamarinus ( 208)) Achromobacter
sp. (317), Pseudomonas sp. (201)) Spirillum lunatum ( 102)) Spirillum sp. ( 101))
and Vibrio sp. (204). All strains grew in
synthetic media and required no vitamins,
The purified cultures were kept in unsupplcmentcd seawater for 2 to 6 months during which period plate counts decreased
by 10 to 60%. For inoculation, the popula-
SUBSTRATE
LEVELS
267
tions were filtered, washed, and resuspcndcd in the culture vcsscl. As soon as
plate counts indicated a distinct increase of
cell numbers, flow was started at the
chosen rate.
Dilution rates ranged from 0.4 to O.Ol/hr,
corresponding to rctcntion times of 2.5 to
100 hr. Longer retention times did not result in steady-state populations due to wall
growth or aggregation of cells.
Twice during a retention time, the population density x and, if possible, the concentration of the limiting
substrate in the
When a
chemostat s were detcrmincd.
steady state was established (X + 3, s + S),
a new reservoir with a lower concentration
of the limiting substrate so was connected
to the chemostat. The so was decreased
stepwisc (100, 50, 20, 10, 5, 2, 1, and 0.5
mg/liter)
until washout of the population
occurred. This procedure was carried out
at the following
relative dilution
rates:
D/run%= 0.5, 0.3, 0.2, 0.1, and in some cases,
0.05. The specific maximum growth rates
for each strain and medium were determined in stationary culture.
Cells were counted as units of equal size
in counting chambers or on membrane filters ( Jannasch 1953) and were calibrated
on a dry weight basis at high population
densities (Jannasch 1963cc). These values
were used for extrapolations at low po,pulation densities,
This method fails with
coccoid organisms and in the case of an
abundant production of “involution’‘-forms.
Best results were obtained with Spirilla,
where cell diameter and amplitude did not
change under starving conditions in the
presence of a delayed division rate.
Death rate determinations must be done
on the basis of viable counts. Here, exponential growth is no longer described by
the e-function but by the base 2, because
each cell, regardless of size, is assumed to
give rise to one colony (doubling time of
cell units rcgardlcss of size ). The growth
rate ,U then becomes the division rate 1/=
p/in 2 ( Pfennig and Jannasch 1962).
RESULTS
The broken line in Fig. 3’ represents the
theoretical function given in equation ( 1)
268
IIOLGER
W.
r
[ANNASCH
1. Threshold concentrations
of three growthlimiting substrates (in mg/liter)
in seawater at several relative growth rates of six strains of marine
bacteria (see p. 267) and the corresponding
maximum growth rates (in hr-‘)
TABLE
~Strain
D/h
Lactate
Glycerol
Glucose
208
0.5
0.5
1.0
0.1
0.05
0.5
1.0
0.15
i*:
0:20
0.5
0.5
1.0
0.34
0.3
0.1
0.05
0.5
1.0
1.0
0.45
fi m
102
Eltn
101
0.5
0.3
0.2
0.1
krl
204
0.3
0.1
0.05
(S, )mg. Lactate/t
FIG. 3. Steady-state
population
density plotted
vs. concentration
of the limiting
substrate in the
reservoir
at four different
growth
rates (strain
101).
&n
317
201
5
10
20
50
0.45
5
10
100
5
5
10
0.35
10
0.15
20
50
100
> 100
0.85
50
50
> 100
0.5
0.3
0.2
0.1
20
50
100
> 100
0.80
50
50
> 100
fi ?I%
no
growth
0.60
:
0.5
0.3
0.1
0.05
&n
for D//L, = 0.5. When strain 101 was
grown at this particular
growth rate at
various lactate concentrations in the reservoir, the decrease of the steady-state population density deviated from the theoretical
curve. This deviation became larger with
decreasing growth rates ( Fig. 3).
Furthermore, below a certain minimum
lactate concentration in the reservoir, no
steady-state population could be maintained
and outwash occurred, The resulting minimum population densities appear as dots
in Fig. 3. The minimum value of so will be
almost identical with the threshold value
of so at which outwash occurred (X + 0)
and at which B becomes equal to so.
Threshold
concentrations
for lactate,
glycerol, and glucose and the maximum
growth rates are given in Table 1 for six
strains of marine bacteria.
According to equation (2) S is equal to
K, at D/pm = 0.5. Therefore, if outwash
occurs at this particular
growth rate on
account of a decreasing so, the resulting
threshold concentration cannot bc smaller
than K,. If it is considerably larger, it must
0.5
5.0
10.0
0.25
gri%h
1
5
20
0.40
dza
0.70
20
50
> 100
0.65
0.80
be assumed that growth is affected. As
shown in Table 2 for two strains, the
threshold concentrations of lactate exceed
the corresponding &-values that were obtained in continuous culture at high popuStrain 317 ( Table 1)
lation densities.
represents a similarly “inefficient”
species
as far as growth at low substrate concenTATSLE
2. Growth constants {determined
at high
population
densities) of two strains of marine bacteria (see p. 267) compared with threshold concentrations
of lactate in seawater at two relative
growth rates
-~
Threshold
Strain
101
201
K*
(mg/liter)
3.0
8.0
P nl.
(hr-I)
0.45
0.80
D//L,
concn. (mg/liter)
= 0.5
5
20
D//A,
= 0.1
50
100
at
GROWTII
OF BACTERIA
AT LOW
SUBSTRATE
LEVELS
269
FIG. 4. Steady-state
concentrations
of limiting
substrate plotted vs. relative growth rate at Eour different substrate concentrations
in the reservoir for strain 101. Horizontal
parts of the curves indicate
washout of the culture (s’ = so). The broken line indicates the theoretical
relationship
( A = K, ).
trations is concerned. In all cases, growth
efficiency decreased with decreasing growth
rate. This phenomenon can be expressed
by a change of the yield coefficient in
equation ( 2) ( Jannasch 1963b ) .
The effect of starving on bacterial cultures often has been interpreted as an increase of a proposed death rate (Postgate
and Hunter 1963). My experiments show
that the ratios between viable count and
direct count does not change appreciably
with decreasing so. In other words, if a
death rate existed, it did not ch,ange and it
could not account for the phenomenon of
premature washout.
DISCUSSION
The washout of the population from the
chemostat at a certain minimum value of
so clearly indicates an uncxpccted change
of growth limitation.
The only factor
changing during the experimental decrease
of so is the population density. Therefore,
since no direct inhibitory
action of the
limiting substrate upon dilution is conccivable, the inhibitory effect, indicated by the
increase of S towards the ordinate (Fig. 4),
must be related to a population effect. In
other words, whatever factor is limiting, the
inhibitory
effect will occur as a corollary
of decreasing population density.
A comparison between Figs. 4 and 1 shows
that, in the present case, the theoretical
relationship bctwecn S and so does not hold
at relatively low growth rates and substrate
concentrations.
If the population density
were plotted in Fig. 4, it would show a
mirror image of the function of 8, as in
Fig. 1.
An explanation of the phenomenon is
suggested by the role of starter populations
in batch cultures. In initially
suboptimal
media, metabolic products may act as conditioning agents for growth (Lodge and
Hinshclwood 1943; Meyrath and McIntosh
1963 ) . This positive feedback mechanism
270
HOLGER
W. JANNASCH
depends upon population density and on
metabolic activity ( growth rate), The current study shows that the phenomenon observed varies with both of these factors
( Fig. 3). When so and, consequently, 3
reach their minimum value, the assumed
feedback action is abolished. At this point,
the production of a proposed metabolite,
for instance, does not meet the demand for
maintenance of the present growth rate,
The resulting drop of the growth rate at a
constant dilution rate leads to complete
washout. When the experiment is done at
a lower dilution rate, the washout occurs
in the presence of a higher population density. Evidence for such a phenomenon and
its mathematical description has been given
earlier by cultivating Spidum
serpens in
a mineral medium with lactate as the limiting carbon and energy source (Jannasch
1963a). It was shown that the inhibitory
effect at low population densities could be
partly eliminated by external adjustment of
the media. At higher population densities,
this adjustment took place internally.
The threshold values reported in Table 1
are relative, of course, and not directly applicablc to seawater. They depend upon
1) the species, 2) the experimental growth
rate, and 3) the given growth conditions.
1) All species studied showed marked
threshold concentrations of the limiting substrates. The enrichment processes were
selective for species of high growth efficiency at low substrate concentrations, so
most of them exhibited relatively low K,values. The two exceptions, strains 201
(see Table 2) and 317, were isolated from
2% lactate agar. Their high &-values, measured directly for all carbon sources tested,
indicated low growth efficiency at low
substrate concentrations.
It has been observed repeatedly (Jannasch 1965a) that in a given species high KSvalues always stem to be combined with
high pm-values or low &values
with low
pullz-values (Table 2). If this is a fact, it
follows that species efficiently growing in
seawater will escape detection and isolation
by the usual techniques of batch culture
enrichments, plating, and so forth, and that
species isolated from seawater on nutrient
agar of the usual strength are not reprcsentative for the active marine microbial flora.
Thus, Winogradsky’s
early concept of
“autochthonous” and “zymogcnous” types of
microorganisms
reappears in terms of
growth parameters.
2) As shown in Fig. 3, the threshold concentrations reported depend on the growth
rate of the organism. This latter parameter
is not directly assessable in the natural
environment. It could be indefinitely
low.
Theoretically,
s’ approaches zero with decreasing growth rate (Fig. 1). However,
the plot of the actual data shows that s
reaches a minimum around 0.3 to 0.5 D/,u~
( Fig. 4). In other words, threshold concentrations can only be larger at lower
growth rates.
3 ) The change of growth limitation below
a certain population density must be attributed to the general growth conditions.
In this sense, seawater must be considered
a suboptimal medium.
The threshold concentrations reported are
at least 1 to, 2 orders of magnitude too high
to account for endogenous respiration or
maintenance metabolisms ( Lamanna 1963).
Furthermore, the inhibitory
effect of the
abolished feedback interaction
is clearly
indicated by the increase of s at low growth
rates (Fig. 4). Thus, the phenomenon observed can be described as truly environmental.
The smallest threshold
concentrations
found in this study are still larger than the
concentrations
of similar compounds reported for natural seawater (Duursma
1965). This discrepancy must be explained
by the fact that the steady-state experiments had to be done with pure cultures.
The models made it possible to study and
to exemplify a principle rather than to reproduce data of a more complex system.
The relatively high concentrations of dissolved organic carbon in the sea and their
constancy (Menzel, in prep.; Walsh and
Douglass 1966) are primarily explained by
an unavailability
of the material to microbial breakdown. The current study suggests
that this unavailability
must not be identical
with chemical resistance against biological
GROWTH
OF BACTERIA
AT LOW
The standing (steady-state
decomposition.
or equilibrium ) concentration
of organic
substrates not utilized by bacterial oxidation may be characteristic for seawater as a
suboptimal environment for microbial life.
REFERENCES
BERGMEYER, II. U. 1963. Methods of enzymatic
analysis.
[Transl. from German.]
Academic,
New York, N.Y. 1064 p.
BUNGAY, II. R. 1966. Population
oscillations
in
symbiotic
continuous
cultures.
Am. Chem.
Sot., 152nd Meeting, Q36.
CONTOIS, D. E., AND L. D. YANGO. 1964. Studies
of mixed steady state microbial
populations.
Am. Chcm. Sot., 148th Meeting, 417.
DUURSMA, E. K. 1965. The dissolved
organic
constituents of sea water, p. 433-475. In J. I’.
Rilcy and G. Skirrow [eds.], Chemical oceanography, v. 1. Academic, New York, N.Y.
HERBERT, D.
1961a.
The chemical composition
of micro-organisms
as a function of their cnvironment.
Symp. Sot. Gen. Microbial.,
11:
391416.
-.
196lb.
A theoretical
analysis of continuous
culture
systems.
Sot. Chcm. Ind,
(London),
Monograph
12, p, 21-53.
1956.
-,
R. ELSWORTH, AND R. C. TELLING,
The continuous culture of bacteria:
a theoretical and experimental
study.
J. Gen. Microbiol., 14: 601-622.
IIOLME, T. 1958. Glycogen
formation
in continuous culture of Esche&hiu
co.& B, p. 6774. In Symp. Continuous
Cult. Microorg.
Prag.
JANNASCH, 11. W.
1953. Zur Methodik
der
quantitativen
Untersuchung
von
Bakterienkulturen
in
fliissigen
Medien.
Arch.
Mikrobiol.,
3 1: 114-124.
-,
1963a.
Bakterielles Wachstum bei geringen Substratkonzentrationen.
Arch,
Mikrobiol., 45: 323-342.
-.
1963b.
Bacterial growth at low population densities ( II ) . Nature, 197: 1322.
-.
1965a. Eine Notiz iiber die Anreicherung
von Mikroorganismen
in Chemostaten, p. 498502. In Anreicherungskultur
und Mutantcnauslese. Zcntr. Bakteriol., v. 1. (Suppl. )
SUBSTRATE
LEVELS
271
1965b.
Continuous culture in microbial
ecology.
Lab. Pratt., 14: 1162-1167.
1963. Dynamics of bacterial
KJELDGAARD, N. 0.
growth.
Nyt Nerd. Arnold Busck, Copenhagcn.
LAMANNA,
C. 1963. Endogenous
metabolism
Ann. N.Y.
with special reference to bacteria.
Acad. Sci., 102: 515-793.
LODGE, R. M., AND C. N. I~INSHELWOOD. 1943.
Physicochemical
aspects of bacterial
growth.
The lag phase of Bact. Zuctis
Part IX.
J, Chem. Sot., 1943 : 213-219.
uerogenes.
MCALLISTER, C. D., T. R. PARSONS, K. STEPHENS,
AND J. D. II. STRICKLAND. 1961. Measurcments of primary
production
in coastal sea
water using a large-volume
plastic sphere.
Limnol. Oceanog., 6: 237-258.
MEYRATII, J,, AND A. F. MCINTOSH.
1963. Size
of inoculum and carbon metabolism in some
J. Gen. Microbial.,
33:
Aspergillus
species.
-.
47-56.
PFENNIG, N., AND II. W. JANNASCH. 1962. Biologisch Grundfragen
bci der homokontinuierlichen Kultur von Mikroorganismen.
Ergeb.
Biol., 25: 93-135.
POSTGATE, J. R., AND J. R. HUNTER.
1963. The
starved
bacteria.
J.
Appl.
survival
of
Bactcriol.,
26: 295-306.
POTTER, L. F. 1960. The effect of pH on the
development
of bacteria in water stored in
Can. J, Microbial.,
6: 257glass containers.
263.
1958. Criteria
for growth
of
POWELL, E. 0.
contaminants
and mutants in continuous
culture.
J. Gen. Microbial.,
18: 259-268.
T~DT, F. 1958. Elektrochemische
Sauerstoffmessungen.
de Gruyter, Berlin. 212 p,
VACCARO, R. F., AND II. W. JANNASCI-I. 1966.
Studies on heterotrophic
activity in seawater
assimilation.
Limnol.
glucose
based
on
Occanog., 11: 596-607.
WALSH, G. E., AND J. DOUGLASS. 1966. Vcrtical distribution
of dissolved carbohydrate
in
Sea off
Bermuda.
Limnol.
the Sargasso
Oceanog., 11: 406407.
ZOBELL, C. E., AND D. Q. ANDERSON. 1936. Observations on the multiplication
of bacteria in
different volumes of stored sea water and the
influence
of oxygen tension and solid surfaces. Biol. Bull., 71: 324-342.