Limnol. Oceanogr.

, 37(6), 1992, 1179-l 192

0 1992, by the American Society of Limnology and Oceanography, Inc.

Nutrient limitation of bacterioplankton growth in

Lake Dillon, Colorado
Donald P. Morris’ and William M. Lewis, Jr.
Center for Limnology, Department of Environmental, Population, and Organismic Biology and Cooperative
Institute for Research in Environmental Science, University of Colorado, Boulder 80309-0334

Abstract
Bacterioplankton biomass, production, and growth rate were measured over a 2-yr period in
Lake Dillon, a mesotrophic Colorado reservoir. In addition, a multivariate statistical analysis and
nutrient addition experiments were used to analyze the regulation of bacterioplankton growth in
situ. Biomass ranged between 170 (winter) and 2,200 mg C m-* (summer); production ranged from
10 (winter) to 625 mg C m-* d-* (summer); annual bacterioplankton production was 47 g C m-2
yr-l in 1987 and 67 in 1988. Population growth rates ranged between 0.001 and 0.08 1 h I. Growth
rates in the epilimnion were substantially below estimated potential rates, suggesting severe nutrient
limitation during the period of stratification. Population growth rates were highly correlated with
P concentrations but not with dissolved organic C concentrations. Bacterioplankton growth in the
summer epilimnion responded strongly to the addition of P alone or in combination with N or
labile organic C. The field data show that bacterioplankton arc frequently without suflicient nutrients
to sustain maximum growth; the experimental and statistical analysts indicate that P, rather than
organic C, is the critical nutrient for bacterioplankton growth in this lake.

Organic carbon of algal origin is readily disproportionately large amount of P (Vad-

used by bacterioplankton (Brock and Clyne stein et al. 1988; Glide 1989; Jiirgens and
1984). In addition, several workers have Giide 1990) and N (Wheeler and Kirchman
demonstrated a relationship between bac- 1986; Kirchman et al. 1990) compared to
terioplankton abundance or production and other planktonic organisms. Although these
phytoplankton production or biomass (e.g. studies support the presumption that bac-
Riemann and Sondergaard 1984; Bird and teria compete favorably for inorganic nu-
Kalff 1984; Cole et al. 1988). Although these trients, they do not show conclusively that
observations suggest strong regulation of planktonic bacteria grow under conditions
bacterioplankton by organic C, there is little of nutrient sufficiency. In fact, cellular P
experimental evidence to demonstrate that contents for natural populations suggest that
the flux of algal C directly constrains bac- bacterioplankton are frequently subsatur-
terioplankton production. ated with P and thus potentially limited by
Inorganic nutrients may also regulate bac- P deficiency (Vadstein et al. 1988; Vadstein
terial activity, but have received less atten- and Olsen 1989).
tion than organic C. The high affinities of Activity or biomass of freshwater plank-
bacteria for inorganic nutrients provides a tonic bacteria can sometimes be increased
significant competitive advantage (Currie by adding inorganic P (Jones 1977; Cooncy
and Kalff 1984a, b,c) and would seem to prc- et al. 1985; Peters et al. 1987; Toolan et al.
elude the possibility of inorganic nutrient 199 1). In addition, statistical evidence sug-
limitation in planktonic bacteria under most gests that P may play an important role in
circumstances. Also, bacteria sequester a determining differences in standing stocks
of bacterioplankton across a wide variety of
’ Present address: Department of Earth and Envi- lakes (Bird and Kalff 1984; Shortreed and
ronmental Sciences, Lehigh University, Bethlehem, Stockner 1986; Currie 1990). However, be-
Pennsylvania 180 15. cause primary production is often stimu-
Acknowledgments lated by P (Schindler 1978), P may influence
Bo Riemann, Russell Bell, and an anonymous re- bacterioplankton indirectly by stimulating
viewer provided suggestions for the improvement of
this manuscript. the rate at which C substrate is produced.
This work was supported by NSF grant BSR 90- Definitive identification of nutrient limi-
20684. tation requires an experimental approach
1179
1180 Morris and Lewis

that will segregate the roles of C and inor- lyzed for particulate organic C (POC) with
ganic nutrients. an elemental analyzer. Particulate P (PP)
During 1987 and 1988, we studied bac- was analyzed according to the method of
terioplankton growth and nutrient limita- Solorzano and Sharp (1980).
tion in Lake Dillon, a mesotrophic Colo- Phytoplankton were collected on GF/C
rado mountain reservoir (elevation, 2,750 filters and chlorophyll a and pheophytin a
m; surface area, 1,330 ha; Morris and Lewis were measured by the spectrophotometric
1988; Lewis et al. 1984). Our purpose was method of Marker et al. (1980) and Nusch
to quantify the role of bacterioplankton in (1980), which is based on extraction of pig-
the C cycle of the lake and to determine the ment with hot 90% ethanol (vol : vol) and
relative importance of C, P, and N in reg- subsequent acidification. Results were pe-
ulating the growth of bacteria. The study riodically validated by HPLC.
relies on data from field populations, nu- Samples collected for determination of
trient limitation experiments, and multi- dissolved organic C (DOC) were filtered
variate statistical analysis to provide insight gently (< 130 mm of Hg) through ashed Gel-
into the seasonal dynamics of bacterio- man A/E filters. The filtrate was frozen
plankton in the lake. pending analysis which was usually within
2 weeks. The sample was digested with per-
Methods sulfate and phosphoric acid in sealed am-
Analytical methods - Samples were col- pules (Wetzel and Likens 1979); CO, from
lected from Lake Dillon at the point of max- the resulting decomposition of DOC was
imum depth (60 m) at biweekly intervals quantified by gas chromatography.
from May through September 1987 and Primary production was determined in
1988 and at monthly intervals during the situ by the 14C method described by Wetzel
rest of the year. On each date, samples were and Likens (1979). Samples collected from
obtained near the surface (0.5, 1.O, 2.0, and seven depths in the photic zone were spiked
5.0 m) with a 7-liter Van Dorn bottle; ad- with 74 kBq of NaH14C0, per sample and
ditional samples were taken with a 5-m in- suspended in 130-ml Pyrex bottles at the
tegrating sampler at 5-m increments from depth from which they were collected. Sam-
7.5 m to the bottom. The samplers were ple incubation (2-4 h) was usually con-
soaked in ethanol and thoroughly rinsed with ducted near midday. Volumetric productiv-
lake water before use. ity values were converted to areal estimates
Samples collected for chemical analysis as suggested by Wetzel and Likens.
were kept cold and dark until they were an- Bacterioplankton were preserved with
alyzed. Particulate and dissolved fractions buffered formaldehyde or glutaraldehyde
were separated by vacuum filtration through (2% final concn), filtered onto 0.2~pm Nu-
ashed and tared Whatman GFK filters. Sol- clepore filters, and counted by fluorescence
uble reactive P (SRP) in the filtrate was an- microscopy at 1,250 x (Hobbie et al. 1977).
alyzed by the ascorbic acid-molybdate At least 400 bacteria were counted per sam-
method of Murphy and Riley (1962). Total ple; no fewer than 10 fields were examined
dissolved P (TDP) was determined by a per filter (expected counting error <5%). It
modification of the oxidation method de- is often difficult to distinguish bacteria and
scribed by Lagler and Hendrix (1982) and photosynthetic picoplankton with acridine
Valderrama (198 1). Dissolved organic P orange. For this reason, unstained samples
(DOP) was calculated as the difference be- were occasionally examined by epifluores-
tween TDP and SRP. NO3 was determined cent microscopy, which is capable of show-
by liquid anion chromatography. ing chlorophyll autofluorescence. Less than
Suspended particulate matter was col- 5% of cells of < l-pm diameter exhibited
lected onto ashed and tared GFK filters, chlorophyll autofluorescence.
dried to constant weight at 6O”C, and re- The mean cell volume of bacterioplank-
weighed. Subsequent to determination of ton was determined for the mixed layer on
total particulate matter, filters were ana- all dates and for the hypolimnion when the
 Bacteria nutrient limitation 1181

lake was stratified. Composite samples were external standard method; blanks consisted
created by abundance-weighted mixing of of samples that were fixed with formalde-
samples from different depths. Cell volumes hyde (2% final concn) before adding the la-
were determined by fluorescence micros- bel.
copy and photomicrography (Fry 1988). Conversion factors for Tdr incorporation
Cells were collected by filtration and stained were determined in triplicate for each sam-
with acridine orange. Fields were photo- pling date from the ratio of Tdr incorpo-
graphed with Kodak T-Max (ASA 2000) or ration to growth rate in unenriched cultures
Kodak Tri-X Pan (ASA 400) high-contrast of bacterioplankton that were collected from
black-and-white film. Negatives were pro- the lake and diluted by the methods given
jected and cell dimensions (100 cells per below. New cells were assumed to have a
sample) were measured with a micrometer. volume equivalent to the mean cell volume
The measurements were routinely calibrat- of bacteria in the stratum of origin on the
ed from photographs of a stage micrometer sampling date.
or standardized microspheres that were pro- Several variables have been converted
jected along with the images of bacterio- from volumetric to depth-integrated (areal)
plankton. The cell volumes were calculated values. For these calculations, each sample
as spheres (cocci) or straight-sided rods with was weighted according to the proportion
hemispheric ends (bacilli). These model of total lake volume found in the stratum
shapes were appropriate because cells were from which it came.
predominantly cocci, bacilli, or shallow vib- Nutrient response experiments-Nutrient
rio. Few deep vibrio or spirillum shapes were limitation was assessed by the growth re-
present in the plankton. Cell volumes were sponse of dilute bacterioplankton commu-
converted to C biomass with the relation- nities to additions of various organic and
ship of Simon and Azam (1989). On the inorganic nutrients. Water collected from
basis of this relationship, the average bac- the epilimnion was split into two fractions
terioplankton in the epilimnion of the lake in’ the laboratory. The first fraction was
(mean cell volume of 0.144 pm”) contained gently filtered (< 130 mm of Hg) through
0.207 pg C pm-3. ashed Whatman GF/D and Gelman A/E
Bacterioplankton production was deter- glass-fiber filters. Although the effective pore
mined from the rate of incorporation of thy- size of the A/E filter is larger than the small-
midine (Tdr) into DNA in situ. Triplicate est bacteria, counts on the filtrate indicated
15-ml samples of lake water were placed that > 98% of the cells were removed by this
into sterile polypropylene test tubes and treatment. Analysis of the filtrate also
spiked with [3H-methyllthymidine (20 nM showed that filtration caused no statistically
Tdr; 1,850-2,590 GBq mmol-‘). The sam- significant increases in dissolved P, N, or
ples were suspended in the lake at the depth DOC:
from which they were collected and allowed The filtrate was used to dilute the natural
to incubate for 1 h (summer) to 4 h (winter). plankton community (1 : 5 to 1 : 20). Dilu-
Uptake of label was stopped by adding tion was greater during summer stratifica-
formaldehyde (2% final concn). Samples tion when bacteria, phytoplankton, and
were then placed on ice and transported to bacteriovores were most numerous at this
the laboratory where they were filtered onto time. The purpose of dilution was threefold:
0.2-pm-pore nylon membrane filters (MS1 to reduce the grazing pressure on bacteria-
Magna) and thoroughly rinsed with 20 ml plankton, to reduce the demand for soluble
of filtered lake water. Filters were frozen nutrients, and to minimize the effect of in-
until the DNA was extracted and purified creased phytoplankton C fixation and exu-
(usually <24 h). DNA was extracted and dation in response to nutrient enrichment.
purified by a modification of the method of During 1987, each enrichment experi-
Findlay et al. (1984), which is based on the ment consisted of three replicates each of a
solubility of DNA in hot and cold TCA. control and seven nutrient addition treat-
Counting efficiency was estimated by the ments. The control was diluted lake water
1182 Morris and Lewis

(summer, > 15°C). The growth response of

bacterioplankton to each experimental
treatment was evaluated by two methods:
determination of bacterioplankton growth
16 Feb 87 0.2 2.0 1.6 124 2.01
rates (p) by direct cell counts (17 dates) and
16 Mar 87 0.9 1.0 1.8 174 1.62 by measurement of Tdr incorporation at the
6 Apr 87 0.7 3.1 1.7 203 1.56 end of the incubation period (all dates). Be-
27 May 87 2.7 2.8 3.9 172 1.51 cause the two methods did not produce ap-
24 Jun 87 1.8 2.1 5.0 62 1.91 preciably different outcomes, Tdr incorpo-
8 Jul 87 1.5 2.0 3.0 75 1.47
22 Jul87 1.7 1.9 2.6 37 I .50 ration data were used to evaluate nutrient
5 Aug 87 1.7 2.4 3.4 22 1.43 limitation on all dates.
19 Aug 87 1.2 1.8 3.1 32 2.37 Estimates of the potential maximum spe-
2 Scp 87 2.7 3.0 2.6 23 1.63 cific growth rate of bacterioplankton (p,,,,)
16 Sep 87 1.8 2.3 3.0 15 2.04 in the lake were obtained for each date from
30 Scp 87 2.5 2.7 3.1 17 1.68
26 Ott 87 2.4 2.6 3.4 52 1.66 the growth rate as determined by direct count
23 Nov 87 1.4 1.9 2.5 132 1.52 of bacterioplankton in treatments contain-
22 Fcb 88 1.9 2.5 2.5 131 1.58 ing either yeast extract (1988) or the com-
14Mar88 0.6 1.3 2.3 128 1.63 bined additions of C, N, and P (1987). Be-
6 Apr 88 1.3 2.1 2.2 164 1.56
23 May 88 1.3 1.5 3.1 207 1.82 cause these treatments usually produced
8 Jun 88 2.9 3.3 10.0 91 1.50 rapidly growing populations, the ratio of in-
22 Jun 88 1.8 2.4 4.9 53 1.91 active cells to actively growing cells was as-
6 Jul 88 2.2 3.0 5.2 33 1.47 sumed to be negligible at the end of the
17Jul88 0.7 1.6 3.9 2 1.50 incubation period. The actual growth rates
3 Aug 88 1.4 1.9 3.5 1 1.43
17 Aug 88 2.4 2.8 2.8 2 2.30 obtained in these treatments were assumed
31 Aug 88 1.9 2.1 3.2 1 1.60 to be reasonable estimates of potential max-
15 Sep 88 3.3 3.3 3.4 <1 2.04 imum specific growth rates, even though
27 Sep 88 2.2 2.2 4.2 2 1.71 small proportions of nongrowing cells may
25 Ott 88 1.6 1.8 3.7 11 1.66
22 Nov 88 2.0 2.3 2.9 61 1.52 have been part of the population.
Results
Nutrient chemistry -Mean concentra-
with no nutrient additions. Nutrient addi- tions of SRP, DOP, PP, N03, and DOC in
tion treatments consisted of diluted lake wa- the mixed layer of the lake are presented in
ter with additions of organic C only (500 pg Table 1. SRP and DOP generally ranged
liter-’ glucose-C and 500 pg liter-’ sodium between 0.2 and 3.0 pg P liter-l but varied
acetate-C), N only (200 pg liter-’ NH&l- considerably between sampling dates. PP
N), and P only (50 pg liter-’ sodium dibasic- varied between 1.6 and 10.0 pg P liter-’ and
P), as well as combinations of these nutri- tended to peak during spring and fall mix-
ents (C+N, C+P, N+P, and C+N+P). In ing. DOC concentrations ranged between
1988, additions of amino acids (50 nM each about 1.5 and 2.5 mg C liter-l and were
of 2 1 amino acids) and yeast extract (2 mg highest and most variable during summer
liter-l) were also made. stratification. NO, concentrations sur-
Depending on the dilution factor for a passed 100 pg N liter-l during winter and
particular date, between 400 and 475 ml of during periods of mixing, but declined
diluent was poured into sterile Whirl Pak steadily after the onset of summer stratifi-
bags. Each bag was then inoculated with cation. NO3 concentrations in the epilim-
sufficient unfiltered lake water to bring the nion were between 15 and 30 pg N liter-’
total sample volume to 500 ml. The samples in 1987 and ~2 pg N liter-l in 1988. NO3
were then incubated in the laboratory under concentrations in the 1988 epilimnion in-
conditions designed to simulate the light and dicate severe N limitation of phytoplankton
temperature regime from which the sample (Morris and Lewis 1988).
was originally collected. Incubations ranged Biomass andproduction ofphytoplankton
in duration from 96 (winter, < 5°C) to 36 h and bacteria -Net primary production was
 Bacteria nutrient limitation 1183

- 40

5w

I
h. - 30

p o!
E
m
-20 o9
f”
b
m 200 H
5 - 10

100

JFMAMJJASONDJFMAMJJASOND
0
JFMAMJJASONDJFMAMJJASOND 1987 Date 1988

1987 Date 1988 Fig. 2. Areal estimates for Chl a, pheophytin a, and
Fig. 1. Areal estimates of primary and bacterial POC for Lake Dillon, 1987-l 988. Hatching-ice cov-
production for Lake Dillon, 1987-l 988. Hatching- er.
ice cover. -i
for the epilimnion, hypolimnion, and the
78 g C mm2 yr-l in 1987 and 101 in 1988. entire water column are shown in Table 2.
In both years, rates peaked in fall as the The volumes were not normally distributed
mixed layer began to thicken but before (Kolmogorov-Smirnov goodness-of-fit test
mixing was complete (Fig. 1). Chl a and for normality, P I 0.001; Sokal and Rohlf
POC peaked about the same time (Fig. 2). 198 1). Cell volumes for the three strata were
Similar primary production trends have contrasted by use of the Kruskal-Wallis test
been noted previously for the lake (Lewis et followed by a nonparametric multiple com-
al. 1984; Morris and Lewis 1988) and can parisons analysis (Devore 1982). The bac-
be explained by the relief of nutrient limi- terioplankton of the hypolimnion were
tation that develops during summer strati- smallest (mean, 0.069 pm3); bacterioplank-
fication. Pheophytin a, which is an index of ton from the whole water column during
decomposition or of zooplankton grazing winter and during spring and fall mixing
(Carpenter and Bergquist 198 5), showed were about the same size (mean, 0.075 pm3),
only slight seasonal variation; maximum although the small difference in size was
values occurred in fall and under the ice statistically significant at a I 0.05. Bacte-
(Fig. 2). rioplankton of the epilimnion were larger
The distribution of bacterioplankton in (mean, 0.144 pm3; cy I 0.05).
the lake showed extensive seasonal and ver- In both years, mean cell volume was low-
tical variation (Fig. 3A). Numbers of bac- est in winter and increased steadily in the
teria were lowest during winter under ice epilimnion during summer stratification.
and during spring and late fall mixing (0. l- Maximum cell volumes occurred near the
0.3 x lo6 cells ml-l); bacteria were distrib- peak of water-column stability and de-
uted uniformly throughout the water col- creased again as the thermocline thickened
umn. The development of larger numbers in late summer and fall. The nonparametric
of bacterioplankton generally corresponded Mann-Whitney U-test (Sokal and Rohlf
to the increasing stability of the epilimnion; 198 1) showed that cells in the epilimnion
in both 1987 and 198 8 bacterioplankton were significantly larger than those in the
abundance peaked in the epilimnion in late hypolimnion (P 5 0.001) for all 16 dates
summer (0.6-1.8 x lo6 cells ml-l). Except when the lake was stratified.
for occasional minor peaks just below the Bacterioplankton biomass showed marked
mixed layer, bacterioplankton density de- seasonal and vertical variation in 1987 and
creased precipitously between the epilim- 1988 (Fig. 3B). A large proportion of the
nion and the hypolimnion. bacterioplankton biomass was restricted to
Pooled bacterioplankton cell volume data the epilimnion. In 1987, the biomass peak
1184 Morris and Lewis

I IMI S IMI I IMI S IM o

20

50

JFMAMJJASONDJFMAMJJASOND JFMAMJJASONDJFMAMJJASOND

IMI S IM-

40

50
i

JFMAMJJASONDJFMAMJJASOND JFMAMJJASONDJFMAMJJASOND
1987 1988 1987 1988
Fig. 3. Temporal and vertical variation for (A) bacterioplankton abundance (lo6 cells ml-l), (B) bacterio-
plankton biomass (mg C m-)), (C) tritiated Tdr incorporation into DNA (pmol liter-l h-l), and (D) bacterio-
plankton production (mg C m-3 d-l) for Lake Dillon, 1987-1988. I-Ice cover; M-mixing; S-stratification.

(33 mg C me3) was lower than in 1988 (75 showed little variation. Biomass per unit
mg C md3). Biomass was often uniformly area (Fig. 4) was lowest under the ice or
distributed within the epilimnion. On some during spring mixing (170-490 mg C mW2)
dates, biomass peaked at or near the ther- and highest during summer stratification
mocline. Bacterioplankton biomass in the (1 ,OOO-2,200 mg C m-‘).
hypolimnion was substantially lower than Incorporation of tritiated Tdr also showed
in the epilimnion (l-8 mg C m-3) and marked seasonal and vertical variation (Fig.

Table 2. Pooled bacterioplankton cell volume data for Lake Dillon, 1987-1988.
Min Max
Mean volume
Sample type N Cm’, SE) Cm’)

Epilimnion 1,632 0.144(0.006) 0.018 0.697

Hypolimnion 1,632 0.069(0.002) 0.016 0.220
Water column 1,152 0.075(0.003) 0.022 0.320
 Bacteria nutrient limitation 1185

Ol’,““““““““““‘I
JFMAMJJASONDJFMAMJJASOND”
JFMAMJJASONDJFMAMJJASOND
1987 Date 1988
1987 Date 1988
Fig. 5. Estimates of population growth rates of bac-
Fig. 4. Areal estimates of bacterioplankton bio- terioplankton in surface and bottom water of Lake Dil-
mass and the ratio ofbacterioplankton biomass to POC lon, 1987-1988. Hatching-ice cover.
in Lake Dillon, 1987-1988. Hatching-ice cover.

3C). The highest incorporation rates were 52 1% of algal production. The highest ratios
observed in the summer epilimnion (1 O-l 2 consistently appeared during the winter un-
pmol liter-’ h-l); the lowest rates occurred der ice when algal and bacterial production
in the hypolimnion or throughout the water were both low.
column in winter or during spring and fall Bacterioplankton production also ex-
mixing (0.2-2.0 pmol liter-’ h-l). Conver- ceeded primary production on several dates
sion factors relating Tdr incorporation to during summer stratification. Annual areal
growth rate were estimated for 13 dates; on bacterioplankton production was 60% of
other dates (mostly winter), growth was too phytoplankton production in 1987 and 66%
slow to allow a reliable estimate. Because in 1988.
conversion factors did not follow any sea- Cell densities and production (as cells
sonal trend, a mean value was used for all produced per unit time) were used as an
calculations (2.03 x lo9 cells nmol-1 Tdr; index of the population growth rate of bac-
SE 0.25 x 109). terioplankton in the lake. The mean pop-
Bacterioplankton production ranged be- ulation growth rate (h) for bacteria was 0.0 12
tween 0.1 and 23.0 mg C m-3 d-’ (Fig. 3D). h-’ (SE 0.0005; n = 409), which is equiv-
Rates of production were lowest in winter alent to a generation time of 2.4 d. Values
under ice and in the hypolimnion (0.1-2.5 of p ranged from 0.001 to 0.08 1 h-l (gen-
mg C rnd3 d-l); the highest rates were found eration time 29.0-0.4 d). In both 1987 and
in the epilimnion during late summer (8.0- 1988, maximum values of p (0.054 and
23.0 mg C m-3 d-l). Within the epilimnion, 0.08 1 h-l) occurred in the epilimnion dur-
production often increased with depth and ing the early stages of stratification (Fig. 5).
peaked at or near the thermocline, but de- Except for these notable peaks, p in the epi-
creased markedly in the hypolimnion. This limnion was generally low (0.005-0.0 10 h-l)
trend resulted from a combination of de- and often even less than p in the hypolim-
creased activity (Tdr incorporation) and a nion.
smaller mean cell volume of bacterioplank-
ton in the hypolimnion. Areal rates of bac- Discussion
terioplankton production (Fig. 1) were low- Biomass, production, and growth rates-
est in winter (lo-50 mg C m-2 d-l) and During 198 7, mean cell volume increased
peaked during summer stratification (270- by a factor of 2.2 from winter minima to
625 mg C m-2 d-l). Annual areal bacterio- summer maxima, while cell abundance in-
plankton production was 47 g C mm2 yr-1 creased by a factor of 2.5. Increases in mean
in 1987 and 67 in 1988. Expressed on an cell volume alone accounted for 47% of the
areal basis, daily estimates of bacterio- seasonal variation of bacteria biomass dur-
plankton production ranged between 22 and ing 1987 and 35% in 1988.
1186 Morris and Lewis

or by a large biomass growing at a low rate.

Population growth rates provide a more di-
rect index of activity and therefore a more
reasonable basis for assessing the impor-
tance of environmental variables in regu-
lating bacterioplankton production.
Mean values of p in the top 10 m of the
lake peaked after the onset of stratification
(Fig. 5), even though bacterioplankton bio-
mass and production were moderate to low
at that time. The high growth rates during
this period resulted in a substantial accu-
6 6 16 mulation of bacterioplankton biomass. Al-
Temperature (“C) though growth rates declined precipitously
Fig. 6. Relationship between F,,,~~and temperature during the remainder of the summer, high
for bacterioplankton of Lake Dillon. production was sustained by the large stand-
ing stock of bacterioplankton.
Measuring activity in terms of growth rate
Bacterioplankton biomass comprised be- is also advantageous because temperature
tween 1 and 14% (mean 4.1 O/o)of total POC places a predictable upper limit on the
when calculated on an areal basis for the growth rates of organisms (Eppley 1972;
entire water column (Fig. 4). Bacterioplank- Goldman and Carpenter 1974). With a tem-
ton biomass accounted for a larger propor- perature conversion, it is possible to com-
tion of POC during summer stratification pare observed growth rates with potential
(5 and 14%) than during winter or spring growth rates that could be expected under
mixing (1 and 3%). ideal conditions.
Although the biomass of bacterioplank- The maximum specific growth rate &,,)
ton comprised only a small proportion of of bacterioplankton in the lake was esti-
the total POC, bacteria accounted for a large mated on 28 dates during the study. As
fraction of the total production (primary + shown in Fig. 6, there is a highly significant
heterotrophic). On an annual areal basis, C relationship between temperature and pmax
equivalent to 60% (1987) and 66% (1988) (P < 0.001):
of net primary production was converted to = 1()-l/0.429 + 0.027(7)1
bacterioplankton biomass. If the growth ef- CL
max

ficiency of bacterioplankton varies between where T is temperature (“C). This equation

40 and 60% (most common values vary be- was used to estimate the potential p,nax for
tween 10 and 90% depending on substrate: bacterioplankton in the top 10 m of the lake
Hobbie and Crawford 1969; Crawford et al. throughout 1987 and 1988 (Fig. 7). The val-
1974; Bell 1984; Jorgensen 1986), then C ues of p,,, varied from 0.0 10 h- 1 (winter)
equivalent to l-2 times net primary pro- to 0.070 (midsummer). Although maxi-
duction was processed by the bacterioplank- mum rates of bacterioplankton production
ton of the lake. were observed during mid- to late stratifi-
Analysis ofpopulation growth rates-Bac- cation, values of p during the same period
terioplankton production has typically been were considerably suppressed relative to pInax
compared to primary production, Chl a, (Fig. 7).
EOC production, and other variables in an The disparity between p and pmax could
attempt to demonstrate how growth is reg- be the result of several factors: decrease in
ulated in situ. This approach implicitly as- the proportion of cells exhibiting Tdr kinase
sumes that production reflects microbial ac- activity, increase in the ratio of dormant
tivity, even though production is affected cells to active cells, or declining specific
by standing stock as well as metabolic rate. growth rates of active cells in the epilimnion
A high level of production could be achieved of the lake. Although the proportion of cells
by a small biomass growing at a high rate incorporating Tdr was not measured, it is
 Bacteria nutrient limitation 1187

_ 0.070
‘;
G 0.060
s!
$ 0.050

fi 0.040

8 0.030

0.020

0.010

O.OW
JFMAMJJASONDJFMAMJJASOND

1967 Date 1666

Fig. 7. Comparison of estimates II,,,,, and actual

population growth rates for bacterioplanktoti in Lake
Dillon, 1987-1988.

unlikely that this mechanism was wholly

responsible for the difference between p and
P,nax. Such a scenario would have required
the proportion of cells incorporating Tdr to
decline substantially during the last 2 weeks
of June (by a factor of 4 in 1987 and by a
factor of 7 in 1988, Fig. 7). After a precip-
itous decline, only a small fraction of the
standing stock of bacterioplankton in the
epilimnion would have exhibited Tdr ki-
nase activity (~25% in 1987 and 14% in
1988). Although seasonal variation occurs
in cells which can incorporate Tdr, these
percentages are unrealistically low (Simek
1986).
Declining growth rates of active cells, or
an increase in the proportion of inactive
cells, is the most probable cause for the dis-
parity between p and pnInx. Although alle-
lopathy could be responsible for declining
growth rate (Lewis et al. 1986; Cannel1 et
al. 1988), it is very unlikely given the low
algal abundance of the lake. A more prob-
able cause for suppression of bacterioplank-
ton growth in the epilimnion is nutrient lim-
itation.
Statistical evidence for nutrient limita-
tion -For Lake Dillon, bacterioplankton
production shows significant bivariate cor-
relation with primary production, Chl a, and
POC (Table 3). Similar results are reported
by Chrzanowski and Hubbard (1988) and
by Simon and Tilzer (1987), who also used
statistical analysis to investigate bacterio-
plankton production in lakes. For Lake Dil-
lon, and possibly for other lakes as well, it
1188 Morris and Lewis

is inappropriate to use the results of this Experimental evidence of nutrient limi-

analysis to make inferences about the reg- tation-In 17 of the 28 nutrient addition
ulation of bacterioplankton growth in situ experiments, the response of bacterioplank-
because bacterial biomass, which affects ton to nutrient additions was evaluated both
bacterial production (r = 0.78 for Lake Dil- on the basis of growth rate (based on cell
lon), reflects cumulative effects of growth counts) and the rate of Tdr incorporation at
and loss over an uncertain period of time, the end of the incubation period. Because
can be low when the growth of individual the outcome was the same for the two meth-
cells is high. Insight into mechanisms that ods of measuring bacterial response, all in-
regulate growth is much more likely to be terpretation of the 28 enrichment experi-
obtained from statistical analysis based on ments is based exclusively on Tdr
growth rate rather than production. incorporation.
A second problem with the use of bivar- After normalization by logarithmic trans-
iate correlation analysis is the lack of in- formation, the Tdr incorporation data were
dependence among a number of variables analyzed by one-way ANOVA to determine
which might affect bacterioplankton growth. whether mean values differed between treat-
This lack of independence is illustrated by ments. When a significant difference (P 5
the highly significant correlation between 0.05) was detected, treatments were com-
production and temperature (P I 0.00 1; r pared a posteriori by use of the Duncan
= 0.77), which is also highly correlated with multiple range test (Sokal and Rohlf 198 l),
every other variable of interest in the data which ordered treatments with respect to
set. The effect of temperature on both me- mean Tdr uptake rate and grouped experi-
tabolism and the physical structure of the mental treatments into statistically homo-
environment (stratification) can be especial- geneous subsets.
ly important in producing statistically sig- Nomenclature and categories of nutrient
nificant correlations between variables that limitation defined for phytoplankton nutri-
have no true cause and effect relationship. ent limitation by Morris and Lewis ( 1988)
Multivariate techniques that control for one are applicable to bacterioplankton com-
or more of the confounding variables are munities and are used here. Four distinct
useful in reducing the importance of this Tdr incorporation responses to nutrient ad-
problem. ditions are possible: no significant difference
Specific growth rates of bacterioplankton in mean Tdr incorporation rates between
in the epilimnion of the lake were analyzed nutrient addition treatments (no limita-
by partial correlation incorporating tem- tion); significantly greater means in treat-
perature as a controlled variable (Table 3). ments that include C (C limitation), N (N
The partial correlation analysis shows that limitation), or P (P limitation); significantly
p is negatively correlated with bacterio- greater means only with simultaneous ad-
plankton biomass (Table 2; P 5 0.001; r = ditions of two or more nutrients (concurrent
-0.42), suggesting competition among bac- limitation, symbolized N+P, N+C, C+P,
teria for nutrients. The concentration of or N + P+ C); and significantly greater means
DOC and of its potential sources (primary with separate applications of two or more
production, POC, Chl a, pheophytin a) is nutrients and with addition of these nutri-
not significantly related to bacterioplankton ents in combination (reciprocal limitation,
growth rate. However, p is significantly re- symbolized N/P, N/C, C/P, or N/P/C).
lated to TDP (P 5 0.001; r = 0.25) and PP Concurrent limitation indicates extreme
(P I 0.001; r = 0.27). Also, the 10 values shortages of more than one nutrient across
of p that comprise the 1987 and 1988 peaks the entire assemblage, whereas reciprocal
occurred at total P concentrations above the limitation indicates limitation of two or
95th percentile for the entire data set. These more components of the bacterioplankton
results suggest that P rather than C is im- assemblage by different nutrients (Morris
portant in regulating bacterioplankton and Lewis 1988).
growth in the lake. Response of bacterioplankton commu-
 Bacteria nutrient limitation 1189

Table 4. Results of the nutrient enrichment experiments for Lake Dillon. Treatments with significantly higher
Tdr production rates that control (P I 0.05): L-low response, < 100% increase relative to control; M-moderate
response, lOO-500% increase relative to control; H-high response, > 500% increase relative to control. Amino
acid (AA) and yeast extract (YE) additions performed only in 1988.
Limitation

Date C N P PC CN NP NPC AA YE Primary Secondary

16 Mar 87 ns ns ns ns ns ns ns None
6 Apr 87 ns ns L L ns L L P
10 Jun 87 ns ns H H ns H H P
24 Jun 87 ns ns H H ns H H P -
7 Jul87 ns L H H L H H P/N -
22 Jul 87 ns ns H H ns H H P -
19Aug87 ns ns M H ns M H P C
2 Sep 87 ns ns ns ns ns ns ns None
16 Sep 87 ns ns H H ns H H P
30 Sep 87 ns ns H H L H H P -
26Oct87 ns ns ns ns ns ns ns None
23 Nov87 ns ns ns ns ns ns ns None
22 Fcb 88 ns ns ns ns ns ns ns ns ns None
14 Mar 88 ns ns ns ns ns ns ns ns ns None
6 Apr 88 ns ns ns ns ns ns ns ns None
23 May 88 ns ns M iti ns M M ns M P -
8 Jun 88 ns ns M M ns M M ns M P -
22 Jun 88 ns ns M M ns M M ns M P -
6 Jul 88 ns ns ns ns ns ns M ns M P+N+C -
17Jul88 ns ns H H ns H H ns H P C
3 Aug 88 ns ns M H ns M H ns H P C
17 Aug 88 ns ns ns ns ns M M ns H P+N -
31 Aug 88 ns ns L L ns L L ns M P N/C
15 Scp 88 ns L L M L M L L H P/N -
27 Sep 88 ns L M M L M H ns H P/N C
25 Oct88 ns ns L L ns L L ns L P -
22Nov88 ns ns ns L ns ns L ns L p+c -

nities to additions of amino acids is ambig- sist of a single nutrient or several nutrients
uous because amino acids provide both C operating concurrently or reciprocally.
and N, but can be interpreted from results The nutrient limitation that was suggest-
of separate and combined additions of C ed by low population growth rates and by
and N. Yeast extract is a multiple nutrient multivariate statistical analysis was con-
source containing a variety of C, N, and P firmed by the nutrient addition experiments
compounds as well as micronutrients. Ad- (Table 4). Bacterioplankton responded to
dition of yeast extract will result in growth nutrient enrichment during most of the
response whenever communities are limit- open-water period; the response was es-
ed by N, P, or C, but could also produce a pecially strong for bacteria in the summer
response in the absence of limitation by N, epilimnion (Fig. 8). P, either alone or in
P, or C if bacterioplankton are limited by combination with N or C, was implicated
trace elements or specific organic sub- in every case of nutrient limitation for the
stances. lake during 1987 and 1988.
Secondary limiting nutrients are those that In 1987, moderate P limitation appeared
are exhausted in cultures that have been en- in April, became more severe in summer,
riched with the primary limiting nutrient. and remained severe until fall turnover.
Their ecological significance is conditional During this period, two of eight experiments
on the availability of the primary limiting showed a strong secondary response to C
nutrient. A secondary limitation may con- or N.
1190 Morris and Lewis

nion (Fig. 8). In summer 1988, samples from

the hypolimnion were used in three nutrient
q Phasphorus (P)
BI
limitation experiments. No significant
I#!
Reciprocal Limitation (P/N)
III
growth response was detected for any of the
nutrient treatments. Nutrient enrichment
Concurrent Llmitatlon (PtN)
also failed to cause a growth response in
Concurrent Umitedion (PtNtC) surface samples during much of the winter
I
under the ice. These results do not neces-
sarily show that nutrient limitation was ab-
JFMAMJJASONDJFMAMJJASOND sent. At low temperatures, differences be-
1987
Date
1988 tween p and pmaX are likely to be small
(~0.0 10 h-l) and difficult to detect without
Fig. 8. Distribution of the forms of bacterioplank-
ton nutrient limitation in Lake Dillon, 1987-1988.
long incubations. Even if nutrient limitation
could be demonstrated in winter or in the
summer hypolimnion, it would have little
The general pattern was similar in 198 8, ecological significance because temperature
but N and C played stronger secondary roles, is the predominant cause for low bacterio-
especially from early August to the onset of plankton growth rates under these condi-
fall mixing. The only growth response to tions.
amino acid additions occurred during this
period. On three dates in August and Sep- Conclusion
tember 1988, the growth response to yeast Bacterioplankton population growth rates
extract was much higher than that to other in Lake Dillon are greatly suppressed rela-
additions. Nevertheless, the primary limi- tive to pmax. Experimental nutrient addi-
tation during this period consisted of P alone tions provide the greatest growth response
or in combination with N. In both years, during periods when p and pnlax differ sub-
the degree of nutrient response declined stantially, indicating that this suppression
during fall mixing as nutrient-rich water en- is the result of nutrient limitation. Both sta-
tered the mixed layer from the hypolim- tistical and experimental studies indicate
nion. that P plays an important role in regulating
Enrichment responses to a large extent bacterioplankton growth. These results,
reflect nutrient concentrations. P was more along with statistical evidence (Bird and
abundant in the lake in 1988 than in 1987 Kalff 1984; Shortreed and Stockner 1986;
because of higher runoff (mean total P concn Currie 1990), P cell content data (Vadstein
in the epilimnion, 7.1 pg P liter-’ in 1988; et al. 1988; Jiirgens and Glide 1990), and
5.3 pg P liter-’ in 1987); response of bac- experimental evidence from other sources
terioplankton to P enrichment was greater (e.g. Jones 1977; Toolan et al. 199 l), suggest
in 1987. The small N response in 1987 (only that P may be important in regulating bac-
one date) is matched by high availability of terioplankton biomass or activity in a va-
inorganic N during that year (NO3 in the riety of aquatic systems.
epilimnion was never < 15 pg liter-l). In There is little doubt that flux of EOC from
contrast, NO3 was severely depleted in the planktonic algae supplies a major fraction
epilimnion in 1988 (< 1 pg liter-‘). The tim- of bacterioplankton substrate (Brock and
ing of the onset and alleviation of bacter- Clyne 1984); in Lake Dillon, the relative
ioplankton N response in 1988 correspond- magnitudes of primary production and bac-
ed closely to the appearance of epilimnetic terioplankton production suggest that bac-
NO3 depletion. Because C limitation oc- teria must depend on algal substrates. How-
curred only once in the lake, it is not pos- ever, it does not follow that algal EOC flux
sible to relate its occurrence to seasonal places a direct constraint on bacterioplank-
variations in C chemistry. ton growth. Inorganic nutrients could reg-
Nutrient limitation of bacterioplankton ulate bacterial growth despite the ultimate
was largely restricted to the summer epilim- reliance of bacteria on organic matter. In
 Bacteria nutrient hnitation 1191

fact, inorganic nutrient deficiency would References

help explain the typical occurrence of mil- BELL, W. H. 1984. Bacterial adaption to low-nutrient
ligram per liter concentrations of DOC in conditions as studied with algal extracellular prod-
surface waters despite the capacity of bac- ucts. Microb. Ecol. 10: 217-230.
BIRD, D. F., AND J. KALFF. 1984. Empirical relation-
teria to take up substrate at low concentra- ships bctwecn bacterial abundance and chloro-
tions (e.g. Jorgensen and Sondergaard 1984). phyll concentration in fresh and marine waters.
A strong role for P in limiting bacterio- Can. J. Fish. Aquat. Sci. 41: 1015-1023.
plankton would seem paradoxical because BROCK, T. D., AND J. CLYNE. 1984. Significance of
algal excretory products for growth of epilimnetic
the principles of uptake kinetics suggest that bacteria. Appl. Environ. Microbial. 47: 73 l-734.
bacterioplankton could obtain sufficient in- CANNELL, R. J. P., A. M. OWSIANKA, AND J. M. WALK-
organic P from the water even when phy- ER. 1988. Results of a large-scale screening pro-
toplankton are severely P limited. The su- gram to detect antibacterial activity from fresh-
perior uptake ability of bacteria has been water algae. Br. Phycol. J. 23: 41-44.
CARPENTER, S. R., AND A. BERGQUIST. 1985. Exper-
demonstrated experimentally in laboratory imental testing ofgrazing indicators based on chlo-
chemostats under conditions designed to in- rophyll degradation products. Arch. Hydrobiol.
duce P competition between bacterioplank- 102: 303-3 17.
ton and phytoplankton (Currie and Kalff CHRZANOWSKI, T. H., AND J. G. HUBBARD. 1988. Pri-
1984a). Currie and Kalff (1984b,c) also re- mary and bacterial secondary production in a
southwestern reservoir. Appl. Environ. Microbial.
ported that small amounts of P added to 54: 66 l-669.
natural, nutrient-deficient assemblages of COLE, J. J., S. FINDLAY, AND M. L. PACE. 1988. Bac-
bacterioplankton and phytoplankton are terial production in fresh and saltwater ccosys-
taken up entirely by bacteria. One physical terns: A cross-system overview. Mar. Ecol. Prog.
Ser. 43: l-10.
basis for the kinetic superiority of bacteria COONEY, J. J., S. A. SILVER, AND E. A. BECK. 1985.
is their high surface-to-volume ratio. Factors influencing hydrocarbon degradation in
Whereas phytoplankton have surface-to- three freshwater lakes. Microb. Ecol. 11: 127-l 37.
volume ratios near 1 and maximum values CRAWFORD, C. C., J. E. HOBBIE, AND K. L. WEBB. 1974.
of 8, bacterioplankton have ratios > 10 The utilization of dissolved amino acids by es-
tuarine microorganisms. Ecology 55: 55 l-563.
(Lewis 1976). In addition, heterotrophic CURRIE, D. J. 1990. Large-scale variability and in-
bacterioplankton have access to organic P teractions among phytoplankton, bacterioplank-
that may not be readily available to phy- ton, and phosphorus. Limnol. Oceanogr. 37: 1437-
toplankton. 1455.
-, AND J. KALFF. 1984a. Can bacteria outcom-
A potential resolution to the paradox of petc phytoplankton for phosphorus? A chemostat
bacterioplankton P deficiency may lie in the test. Microb. Ecol. 10: 205-216.
relative abilities of bacteria and phyto- -, AND -. 1984b. A comparison of the
plankton to store P. The few studies of P abilities of freshwater algae and bacteria to acquire
storage in bacterioplankton suggest that and retain phosphorus. Limnol. Oceanogr. 29: 298-
310.
bacteria do not contain excess P (Jiirgens -, AND -. 1984~. The relative importance
and Glide 1990; Vadstein et al. 1988) and of bacterioplankton and phytoplankton in phos-
thus may store only small amounts of phorus uptake in freshwater. Limnol. Oceanogr.
polyphosphate-P suitable for short-term 29: 311-321.
DEVORE, J. 1982. Probability and statistics for en-
homeostasis rather than long-term reserve. gineering and science. BrooksKolc.
In contrast, phytoplankton can store suffi- EPPLEY, R. W. 1972. Temperature and phytoplank-
cient quantities of polyphosphate-P for up ton growth in the sea. Fish. Bull. 70: 1063-1085.
to 10 cell divisions (Reynolds 1984). If P FINDLAY, S. E. G., J. L. MEYER, AND R. T. EDWARDS.
storage capacity of bacterioplankton is less 1984. Measuring bacterial production via rate of
incorporation of [“HI thymidine into DNA. J. Mi-
than that of phytoplankton, P limitation of crobiol. Meth. 2: 57-72.
bacteria could occur even when phyto- FRY, J. C. 1988. Determination of biomass, p. 27-
plankton are not P deficient. Thus, differ- 72. In B. Austin [ed.], Methods in aquatic bacte-
ential P storage capacity may explain why riology. Wiley.
P limitation of bacterioplankton occurs in GOLDMAN, J. C., AND E. J. CARPENTER. 1974. A ki-
netic approach to the effect of temperature on algal
nature despite the kinetic superiority of bac- growth. Limnol. Oceanogr. 19: 756-766.
teria for P uptake. GLADE, H. 1989. The role of grazing on bacteria in
1192 Morris and Lewis

plankton succession, p. 337-364. Zn U. Sommer REYNOLDS, C. S. 1984. The ecology of freshwater

[ed.], Plankton ecology. Springer. phytoplankton. Cambridge.
HOBBIE, J. E., AND C. D. CRAWFORD. 1969. Respi- RIEMANN, B., AND M. SONDERGAARD. 1984. Bacterial
ration corrections for bacterial uptake of dissolved growth in relation to phytoplankton primary pro-
organic compounds in natural water. Limnol. duction and extracellular release of organic car-
Oceanogr. 14: 528-532. bon, p. 233-247. In J. E. Hobbie and P. J. LcB.
-, J. DALEY, AND S. JASPER. 1977. Use of Nu- Williams [eds.], Heterotrophic activity in the sea.
clepore filters for counting bacteria by fluoresccncc Plenum.
microscopy. Appl. Environ. Microbial. 33: 1225- SCHINDLER, D. W. 1978. Factors regulating phyto-
1228. plankton production and standing crop in the
JONES,J. G. 1977. The effect of environmental factors world’s freshwaters. Limnol. Oceanogr. 23: 478-
on estimated viable and total populations of 486.
planktonic bacteria in lakes and experimental en- SHORTREED, K. S., AND J. G. STOCKNER. 1986. Tro-
closures. Freshwater Biol. 7: 67-9 1. phic status of 19 subarctic lakes in the Yukon
JBRGENSEN,N. 0. G. 1986. Fluxes offrec amino acids Territory. Can. J. Fish. Aquat. Sci. 43: 797-805.
in three Danish lakes. Freshwater Biol. 16: 255- SIMEK, K. 1986. Bacterial activity in a reservoir de-
268. termined by autoradiography and its relationship
-, AND M. SONDERGAARD. 1984. Are dissolved to phyto- and zooplankton. Int. Rev. Gesamten
amino acids free? Microb. Ecol. 10: 30 l-3 16. Hydrobiol. 71: 593-6 12.
J~~RGENS,K., AND H. GLADE. 1990. Incorporation and SIMON, M., AND F. AZAM. 1989. Protein content and
release of phosphorus by planktonic bacteria and protein synthesis rates of planktonic marine bac-
phagotrophic flagellates. Mar. Ecol. Prog. Ser. 59: teria. Mar. Ecol. Prog. Ser. 51: 201-213. .
27 l-284. AND M. M. TILZER. 1987. Bacterial response
KIRCHMAN, D. L., R. G. KEIL, AND P. A. WHEELER. to ’ seasonal changes in primary production and
1990. Carbon limitation of ammonium uptake phytoplankton biomass in Lake Constance. J.
by heterotrophic bacteria in the subarctic Pacific. Plankton Res. 9: 535-552.
Limnol. Oceanogr. 35: 1258-l 266. SOKAL, R. R., AND F. J. ROHLF. 198 1. Biometry, 2nd
LAGLER, C. L., AND P. F. HENDRIX. 1982. Evaluation ed. Freeman.
of persulfate digestion method for particulate ni- SOL~RZANO, L., AND J. SHARP. 1980. Determination
trogen and phosphorus. Water Res. 16: 145l-l 454. of total dissolved phosphorus and particulate
LEWIS, W. M., JR. 1976. Surface/volume ratio: Im- phosphorus in natural waters. Limnol. Oceanogr.
plications for phytoplankton morphology. Science 25: 754-758.
192: 885-887. TOOLAN, T., J. D. WEHR, AND S. FINDLAY. 199 1. In-
-, T. FROST, AND D. P. MORRIS. 1986. Studies organic phosphorus stimulation of bacterioplank-
of planktonic bacteria in Lake Valencia, Vene- ton production in a meso-autrophic lake. Appl.
zuela. Arch. Hydrobiol. 106: 289-305. Environ. Microbial. 57: 2074-2078.
-, J. F. SAUNDERS, D. W. CRUMPACKER, AND C. VADSTEIN, O., A. JENSEN, Y. OLSEN, AND H. REINERT-
BRENDECKE. 1984. Eutrophication and land use: SEN. 1988. Growth and phosphorus status of lim-
Lake Dillon, Colorado. Springer. nctic phytoplankton and bacteria. Limnol. Ocean-
MARKER, A. F., E. A. NUSCH, H. RAI, AND B. RIEMANN. ogr. 33: 489-503.
1980. Measurement of photosynthetic pigments -, AND Y. OLSEN. 1989. Chemical composition
in freshwaters and standardization of methods: and phosphate uptake kinetics of limnetic bactc-
Conclusions and recommendations. Ergeb. Lim- rial communities cultured in chemostats under
nol. 14: 91-106. phosphorus limitation. Limnol. Oceanogr. 34: 939-
MORRIS, D. P., AND W. M. LEWIS, JR. 1988. Phyto- 946.
plankton nutrient limitation in Colorado moun- VALDERRAMA, J. C. 198 1. Simultaneous analysis of
tain lakes. Freshwater Biol. 20: 3 15-327. total nitrogen and total phosphorus in natural wa-
MURPHY, J., AND J. P. RILEY. 1962. A modified single ters. Mar. Chem. 10: 109-122.
solution method for the determination of phos- WETZEL, R. G., AND G. E. LIKENS. 1979. Limnolog-
phate in natural waters. Anal. Chim. Acta 27: 3 l- ical analyses. Saunders.
36. WHEELER, P. A., AND D. L. KIRCHMAN. 1986. Uti-
NUSCH, E. A. 1980. Comparison of different methods lization of inorganic and organic nitrogen by bac-
for chlorophyll and phaeopigment determination. teria in marine systems. Limnol. Oceanogr. 31:
Erbeg. Limnol. 14: 14-36. 998-1009.
PETERS, G. T., J. R. WEBSTER, AND E. F. BENFIELD.
1987. Microbial activity associated with seston Submitted: 17 June 1991
in headwater streams: Effects of nitrogen, phos-
phorus and temperature. Freshwater Biol. 18: 405- Accepted: 5 May 1992
413. Revised: 23 June 1992