JOURNAL OF PLANKTON RESEARCH
j
VOLUME
26
j
NUMBER
6
j
PAGES
681–695
j
2004
Factors affecting the bacteria– heterotrophic nanoflagellate relationship in oligo-mesotrophic lakes RE´MY D. TADONLE´KE´*, B. PINEL-ALLOUL1, N. BOURBONNAIS2 AND F. R. PICK3 ` MONTRE´AL, CP 8888, SUCC. CENTRE VILLE, MONTRE´AL, QC, H3C 3P8, 1GRIL – DE´PARTEMENT DE SCIENCES GEOTOP – UNIVERSITE´ DU QUE´BEC A 2 BIOLOGIQUES, UNIVERSITE´ DE MONTRE´AL, CP 6128, SUCC. CENTRE VILLE, MONTRE´AL, QC, H3C 3J7, MINISTE`RE DE L’ENVIRONNEMENT 3 ´ ˆ ET DE LA FAUNE, GOUVERNEMENT DU QUEBEC, SEPT-ILES, H3C 3P8 AND FACULTY OF SCIENCES, UNIVERSITY OF OTTAWA, OTTAWA, PO BOX
450, STN, ONTARIO, K1N 6N5, CANADA
*CORRESPONDING AUTHOR:
[email protected]
Received on November 4, 2003; accepted on February 24, 2004; published online on March 30, 2004
The coupling between bacteria and heterotrophic nanoflagellates (HNF) was examined in nine lakes of low productivity for evidence of the effects of various metazooplankton (i.e. rotifers, cladocerans and copepods) on this relationship. We considered the size of cladocerans and, in contrast to most previous across-system studies, the three strata of the water column (i.e. epilimnion, metalimnion and hypolimnion). Rotifers were numerically dominant in all lakes and accounted for 45–84% of total metazooplankton abundance, while the abundance of large cladocerans was relatively low, ranging from 0.066 to 15.2 ind. L1. The across-lake relationship between bacteria and HNF was significant in the deeper strata (meta- and hypolimnion) but not in the epilimnion and in the two groups of lakes separated on the basis of their average number of large cladocerans (5 ind. L1, respectively). The results confirmed the negative impacts of large cladocerans on HNF, but also showed that rotifers, probably through grazing on HNF, may be an important factor causing variation in the bacteria–HNF relationship in unproductive waters. Quadratic models best described the relationships between metazooplankton and the ratio of bacteria to HNF. This ratio seemed to be a result of complex interactions between several factors, including the zooplankton composition and abundance and the depth of the lake. Indeed, this ratio significantly decreased across lakes, with increase in depth. In addition, shallower lakes (having 5 large cladocerans L1 and substantial proportions of P. vulgaris). We suggest that the epilimnion, metalimnion and hypolimnion of lakes be taken into account when analysing the bacteria–HNF relationship as well as the cascading effects of zooplankton on microbial communities.
INTRODUCTION Heterotrophic nanoflagellates (HNF) are important predators of bacteria in aquatic systems [e.g. (Haas and Webb, 1979; Fenchel, 1982; Nagata, 1988)]. Acrosssystem and individual lake analyses have shown good correlations between these two compartments of the microbial food web (Berninger et al., 1991). However, other studies, both across systems and within individual lakes, have revealed insignificant or only weak correlations between bacteria and HNF [e.g. (Gasol and Vaque´, 1993; Tzaras and Pick, 1994; Wieltschnig et al., 2001)]. The regulation of HNF abundance by organisms
from both microbial compartments and higher trophic levels (Gasol and Vaque´, 1993; Ju¨rgens et al., 1996) has been considered as one of the most important factors that introduce variability into the relationship between bacteria and HNF. The planktonic cladoceran Daphnia is considered to be a key factor in controlling HNF abundance and negative relationships have been found between Daphnia abundance and HNF within and among lakes [e.g. (Gu¨de, 1988; Gasol and Vaque´, 1993; Pace and Vaque´, 1994; Ju¨rgens and Stolpe, 1995)]. However, some conflicting results have been reported and it is unclear whether all cladocerans
doi: 10.1093/plankt/fbh060, available online at www.plankt.oupjournals.org Journal of Plankton Research Vol. 26 No. 6 Ó Oxford University Press 2004; all rights reserved
JOURNAL OF PLANKTON RESEARCH
j
VOLUME
have a negative impact on HNF. For example, in a comparative study, Vaque´ and Pace (Vaque´ and Pace, 1992) found that the abundance of HNF in lakes dominated by small cladocerans was higher than in lakes dominated by large Daphnia species. On the other hand, an experimental study by Ju¨rgens et al. ( Ju¨rgens et al., 1996) showed that small cladocerans (Bosmina and Ceriodaphnia) also have a strong grazing impact on HNF, their weight-specific clearance rates being higher than that of Daphnia. It is worthwhile to note that the metazooplankton communities of most aquatic systems where negative relationships have been found between HNF and Daphnia, were largely dominated by Daphnia (Gasol, 1994; Ju¨rgens and Stolpe, 1995). Although the interactions between bacteria and HNF in systems with a low number of Daphnia are expected to be different compared to aquatic systems where Daphnia dominates (Gasol and Vaque´, 1993; Ju¨rgens, 1994), comparative studies that examine the bacteria–HNF relationship in lakes with a gradient in Daphnia abundance without a large change in trophic level such as oligo-mesotrophic lakes, are uncommon and have not often taken into account the zooplankton abundance and composition (Tzaras and Pick, 1994). To our knowledge only Gasol et al. (Gasol et al., 1995) have examined the coupling between bacteria and HNF in relation to various metazooplankton groups across freshwater systems with low productivity. Such studies may provide better understanding of trophic interactions in these lakes of low productivity than do larger-scale crossecosystem studies dealing with lakes spanning a wide range of productivity, given that the latter describe (or predict) general patterns of variations (Pace et al., 1998). On the other hand, almost all across-system studies that have analysed the bacteria–HNF relationship have been limited to the euphotic zone (Sanders et al., 1992) or to the epilimnion (Tzaras and Pick, 1994; Gasol et al., 1995) plus the upper part of the metalimnion (Gasol and Vaque´, 1993), although the aphotic zone of fresh waters may contain the bulk of the bacterial community (Tulonen et al., 2000). Furthermore, numerous planktonic crustacean species are known to undergo downward vertical migrations in the water column during the day, which are often more pronounced in unproductive lakes [review in (Wetzel, 2001)]. This could lead to an underestimation of their impact on heterotrophs if deeper waters are not considered. The aims of this study were (i) to examine whether the across-lake relationships between bacteria and HNF differ significantly between the different strata of the water column (i.e. epilimnion, metalimnion and hypolimnion) and (ii) to test the effects of various metazooplankton communities (i.e. rotifers, cladocerans and copepods) on these relationships. All these zooplankton groups were examined
26
j
NUMBER
6
j
PAGES
681–695
j
2004
because they are potentially important sources of predation on HNF. The study was conducted in a set of nine oligo-mesotrophic lakes in Que´bec (Canada). Given that the relative importance of cladocerans increases with the trophic status of a lake (Patalas, 1972), it is expected that cladocerans will be present in low numbers in our lakes. On this basis, we expected significant relationships between bacteria and HNF, if HNF are not limited by other factors. The importance of the abundance and the body-size of the cladocerans was also assessed. For this purpose, we examined whether lakes with >5 large-bodied cladocerans L1 (average number, see the Method section) had higher ratios of bacteria to HNF (B:HNF) than lakes with 5 large cladocerans L1 and GII: 0.9 mm, a criterion commonly used in planktonic food web analysis [e.g. ( Ju¨rgens, 1994; Mazumder, 1994; Pace and Vaque´, 1994)]. In previous works reporting a negative impact of Daphnia on the components of the microbial food web, the effective number of Daphnia is not mentioned (Pace et al., 1998). We chose to base the average number of large cladocerans (five) used to separate lakes in our study on the work of Ju¨rgens ( Ju¨rgens, 1994). Ju¨rgens found that when the number of large cladocerans is 5 cladocerans L1, in which P. vulgaris dominated the rotifer assemblage), and the position of these lakes at this part of the curve corresponded not only to an increase in total metazooplankton abundance but also to a general increasing trend in the mean abundance of large cladocerans from 5.6 ind. L1 in Lake Brompton to
DISCUSSION This study was conducted in a set of lakes of low productivity to examine the effects of various metazooplankton groups on the bacteria–HNF relationship in lakes of this type. In such systems, complex interactions between components of the microbial food web and small zooplankton are expected, but are poorly understood (Sherr and Sherr, 1988). This study, contrary to most crosssystem comparisons, took into account the metalimnion and the hypolimnion, given that metazooplankton can undergo daily vertical migrations. Our total zooplankton densities were generally lower than those reported by Gasol et al. (Gasol et al., 1995) in lakes with trophic status similar to ours. Lake Achigan, which is among our study lakes, was also sampled by Gasol and co-authors during 1990, i.e. the same year as our study. The observed difference, in the metazooplankton abundance in this lake, between our study and that of Gasol et al. (Gasol et al., 1995) might be because these authors conducted their sampling at a monthly frequency from May to September and considered only the epilimnion of lakes. The relationships between bacteria and HNF were not significant in most cases during this study, even in the
Table IV: Among-lake relationships between bacteria (x axis) and heterotrophic nanoflagellates ( y axis) Slope
Intercept
r2
P
n 9
Overall
0.513 (0.158)
0.066 (0.993)
0.599
0.014
Epilimnion
0.287 (0.155)
1.515 (0.972)
0.329
0.106
9
Metalimnion
0.686 (0.227)
1.075 (1.428)
0.565
0.019
9
Hypolimnion
0.598 (0.213)
0.450 (1.323)
0.528
0.026
9
Analyses were carried out with log10-transformed data; the standard errors of estimates are given in brackets; P, probability; n, number of points. Note that each point is an average of 15 samples for overall data and of five samples for each of the three strata.
689
JOURNAL OF PLANKTON RESEARCH
j
VOLUME
26
j
j
6
PAGES
681–695
j
2004
3.5
7.0
A
y = 1.29(0.297)x 2 - 12.48(2.89)x + 32.91(6.97) R2 = 0.764, p = 0.013
A
-1
Log(Bacterial abundance, cells mL )
NUMBER
9.1 Ec 7.5 Ge
Log(B:HNF)
6.5
84.8 Mw 3.0
9 Cw
Cr 10.5
Mg
19.2 Ac 26.3
6.0
Pr 14
Br 42.3
y = -0.26(0.064)x + 7.14(0.225) 2
R = 0.698, p = 0.005
2.5
5.5
4.0
1.5
2.5
3.5
4.5
5.0
5.5 -1
4.5
Log(Rotifer abundance, ind L )
-1
Log(large cladoceran abundance, ind L ) 3.5
B
-1
Log(HNF abundance, cells mL )
B
y = 1.42x 2 - 14.41x + 39.48 R2 = 0.501, p = 0.047
3.6
Ec Log(B:HNF)
Ge
3.4
Cw
3.0
Mw
Cr
Mg Ac Pr
Br
3.2 2.5 y = -0.178 (0.04)x + 3.86(0.139) 2 R = 0.762, p = 0.0046
4.6
4.8
5.0
5.2
5.4
5.6 -1
Log(Total metazooplankton abundance, ind L )
3.0 1.5
2.5
3.5
4.5 -1
Log(large cladoceran abundance, ind L )
Fig. 3. Across-lake relationships of large cladocerans with bacteria (A) and HNF (B). The outlier (open diamond) in (B) was not included in the regression. Standard errors of estimates are given in brackets.
group of lakes with
