Journal of Plankton Research Vol.22 no.1 pp.151–168, 2000
Adult, not juvenile mortality as a major reason for the midsummer
decline of a Daphnia population
Stephan Hülsmann and Winfried Weiler1
Institute of Hydrobiology, Dresden University of Technology, D-01062 Dresden,
Germany and 1Institute of Freshwater Ecology and Inland Fisheries, Department
Limnology of Stratified Lakes, Alte Fischerhütte 2, D-16775 Neuglobsow,
Germany
Abstract. We analysed dynamics and structure of a Daphnia galeata population prior to and during a
midsummer decline of this species in Bautzen reservoir (Saxony, Germany). Patterns of juvenile and
adult mortality were determined by combining field data with laboratory estimates of juvenile growth.
After establishing high densities, fecundity and recruitment of D.galeata markedly decreased, whereas
size at maturity was high. Immediately before the population decline, adult mortality increased and
remained high even after the decline, whereas juvenile mortality was low during the whole investigation period, and virtually absent after the decline. We conclude that the succession of events leading
to a midsummer decline of Daphnia is as follows. (i) A quick increase in Daphnia abundance leads
to the formation of a strong ‘peak cohort’ of about the same age. (ii) During the clear-water phase,
food conditions deteriorate, fecundity declines and hence, recruitment is low. Juvenile mortality
during this period is low, but present. (iii) Adult mortality increases when the ‘peak-cohort’ reaches
its mean life-span, which is reduced due to interactions between age-specific and starvation-induced
mortality. At this point, Daphnia population dynamics can no longer be explained without the onset
of size-selective predation. Hence, the timing between enhanced mortality due to senescence on the
one hand and predation on the other hand, both directed towards adult daphnids, may be decisive for
the initiation of a midsummer decline of Daphnia.
Introduction
A ‘midsummer decline’ of large-bodied cladocerans, especially daphnids, is a
common phenomenon in temperate lakes and reservoirs (Sommer et al., 1986).
After achieving high densities in spring and early summer, Daphnia populations
often decline to very low numbers for extended periods in midsummer. This
decline has been attributed to poor food conditions (Lampert et al., 1986; Sommer
et al., 1986), predation by fish, especially 0+ fish (Mills and Forney, 1983), or a
combination of low food and invertebrate predators (Luecke et al., 1990; De Stasio
et al., 1995). However, only a few studies have quantified fish consumption to
evaluate its impact on Daphnia populations and among those, results were contradictory. Mills and Forney found that consumption by age-0 perch exceeded
production of D.pulex in years with a midsummer decline (Mills and Forney, 1983).
By contrast, other studies found that Daphnia mortality during the decline could
not be explained by age-0 fish consumption alone (Wu and Culver, 1994; Boersma
et al., 1996; Mehner et al., 1998a). There is, however, strong evidence that the indirect effects of size-selective predation may be important. As concluded from
demographic changes in Daphnia populations, a disproportionate loss of adult
daphnids, especially when combined with low fecundity, may enhance direct
predation effects (Gliwicz et al., 1981; Hülsmann and Mehner, 1997; Mehner et al.,
1998a, b). On the other hand, several studies have reported high juvenile death
rates causing a decline (De Bernardi, 1974; Boersma et al., 1996).
© Oxford University Press 2000
151
S.Hülsmann and W.Weiler
Different sources of mortality should have different effects on demography of
a Daphnia population. Planktivorous fish selectively feed on large prey (Gliwicz
and Pijanowska, 1989), though size selection by 0+ fish depends very much on
gape-size (Mehner et al., 1998b). Contrary to fish predation, invertebrate predators are generally considered to prey preferentially on smaller species and instars;
however, this cannot be generalized as it depends on the size of different
zooplanktivorous species (Benndorf, 1995). Moreover, results concerning, for
example, size selection of Leptodora kindtii are contradictory (Herzig and Auer,
1990; Lunte and Luecke, 1990). The effects of non-predatory sources of mortality (starvation, disease, inability to adapt to changing environmental conditions)
might also differ in strength during the life history of an individual Daphnia. In
summary, since, in natural environments, it is virtually impossible to quantify all
these mortality factors simultaneously, size-specific mortality of Daphnia cannot
be determined by looking at its sources.
To overcome this problem, several approaches have been developed for estimating size-specific mortality of cladocerans from population parameters and
growth data (Argentesi et al., 1974; Vijverberg and Richter, 1982; Lynch, 1983).
However, application of existing population dynamics models (De Bernardi,
1974; Hovenkamp, 1989; Boersma et al., 1996) to field data is problematic because
model parameters have to be adjusted to the specific population of interest (see
below). The approach of Lynch requires considerable computational effort and
is difficult to apply (Lynch, 1983). From a Daphnia perspective, however, it is not
important to know the exact instar- or size-specific mortality, but to distinguish
between juvenile (pre-reproductive) and adult (potentially reproductive) mortality.
Although many studies have analysed Daphnia population dynamics during
midsummer declines, the explanations for the observed (total) mortality patterns
have remained somewhat speculative and contradictory (Hall, 1964; Threlkeld,
1979; De Stasio et al., 1995). Consequently, estimation of stage-specific mortality,
and combining these data with indicators of juvenile and adult fitness, may be
major steps towards gaining insight into the mechanisms that lead to a midsummer decline of daphnids. This may elucidate the importance of bottom-up
factors for population dynamics of Daphnia and may help to focus on possible
mortality sources. Studies of this kind are rare (Boersma et al., 1996). Therefore,
we developed a simple method for separately calculating juvenile and adult
mortality of Daphnia, based mainly on data derived from field samples (abundance, size-structure, fecundity) and, additionally, requires estimates of juvenile
growth rates, which is inevitable for the calculation of size-specific mortality. Estimates of juvenile growth can also be used as a measure of fitness (Lampert and
Trubetskova, 1996), whereas fecundity values can be used as an index of starvation in adults (Tessier et al., 1983).
Study site
The study was performed in Bautzen reservoir, situated about 70 km north-east
of Dresden (Saxony, Germany). This highly eutrophic lake has a surface area of
152
Daphnia population decline
533 ha, a mean depth of 7.4 m and a maximum depth of ~12 m. Due to high exposure to wind, Bautzen reservoir stratifies only for short periods in summer.
Usually, thermal stratification develops for a short period in May, but is disturbed
by mixing events. In June and July, further stratification periods may occur but
often, the vertical temperature gradient from the surface to the bottom is only a
few degrees or even absent (Benndorf and Henning, 1989; Köhler, 1992; S.Hülsmann, personal observation). Since 1981, a biomanipulation experiment has been
carried out [(Benndorf, 1995) see also for more details of Bautzen reservoir],
which has resulted in a drastic decrease of zooplanktivorous fish. This, in turn,
led to the dominance of Daphnia galeata in the zooplankton community.
However, a midsummer decline of Daphnia had been observed in many previous
years (Benndorf, 1995; Mehner et al., 1998a).
Method
Zooplankton was sampled at three fixed stations in the pelagic zone at a water
depth of ~10 m twice a week from May to July in 1997. Samples were taken with
a Friedinger-type tube-sampler (Limnos, Finland) of 2 l capacity, integrating
vertically in steps of 1 m from the surface to the bottom of the reservoir by
concentrating single tubes with a plankton net of 30 µm mesh size. Additionally,
samples from the three stations were pooled, resulting in one composite sample
for the pelagic zone of Bautzen reservoir. A subsample of the filtrate was used
for analysis of particulate organic carbon (POC). Zooplankton was immediately
preserved in 3% sucrose–formaldehyde solution (Haney and Hall, 1973). Water
temperature was recorded simultaneously with the zooplankton sampling with a
digital probe (WTW, Germany) in depth intervals of 1 m. For carbon analyses,
the 30 µm filtrate was filtered with pre-combusted glass fibre filters (GMF 5,
Filtrak, Germany), which were dried at 60°C for 4 h, and analysed in a carbon
analyser (C 200, Leco, Germany).
At least 100–150 specimens of D.galeata were counted and measured from the
top of the head to the base of the spine under magnification of 3100. If eggs were
present in the brood pouch, they were counted and classified to developmental
stages (I–IV) according to Threlkeld (Threlkeld, 1979) with the modification that
Threlkeld’s stages 3 and 4 were classified as stage III because their duration is
short and classification difficult. Total egg development time, D, was calculated
from the mean water temperature using the formula given by Bottrell et al.
(Bottrell et al., 1976).
The approach of Johnsen was modified to calculate hatching frequency
(Johnsen, 1983). According to Threlkeld (Threlkeld, 1979), eggs spend different
portions, Ai, of total egg development time, D, in each egg stage i (AI = 0.302,
AII = 0.32, AIII = 0.218, AIV = 0.16). Thus, the duration Di of egg stage i is
Di = D · Ai (d)
(1)
The proportion, Pi, of eggs that are going to hatch during the time interval, ∆t,
can be calculated for each egg stage (I–IV):
153
S.Hülsmann and W.Weiler
if
DIV > ∆t
then
PIV = ∆t/DIV and PI–III = 0
if
DIV + DIII > ∆t
then
PIII =
if
if
else PIV = 1
(2)
else PIII = 1 and
(3)
DIV + DIII + DII > ∆t then
∆t – (DIV + DIII)
PII = –––––––––––––– and PI = 0 else PII = 1 and
DII
(4)
D > ∆t
∆t – (DIV + DIII + DII)
PI = ––––––––––––––––––––
(5)
then
∆t – DIV
––––––––
DIII
and PI–II = 0
DI
else PI = 1
The total number of eggs, E, that are going to hatch during the time interval ∆t
was calculated from Pi and the observed frequency of each egg stage, Fi:
IV
E = ∑ Pi · Fi
(6)
i=I
These calculations will result in an underestimation of E if D < ∆t. During June
and July 1997, this was the case when the sampling interval ∆t was 4 days and,
depending on temperature, exceeded D by 0.5–1 day.
The proportion of adults in the population of D.galeata was estimated after
determining the size at maturity (SAM) according to Stibor and Lampert (Stibor
and Lampert, 1993). In case of extreme low fecundity of Daphnia, this value was
derived from additional samples taken with a 335 µm mesh plankton net and a
minimum of 200 daphnids were measured. Clutch size was calculated as eggs per
adult Daphnia.
Cultures
Neonates of Daphnia were cultured and fed on natural lake seston to measure
their somatic growth during the changing food conditions in the reservoir. Each
week from the end of May until July, egg-carrying females of D.galeata were
isolated from a sample of Bautzen reservoir and neonates born within a 12 h
interval were used for growth experiments in flow-through systems (Stich and
Lampert, 1984). A subsample of the neonates was taken to determine their initial
length (Li). The remaining newborn daphnids were placed in the culture vessels
at a density of less than 50 animals l–1. The number of experimental animals was
always higher than five in each culture. Filtered water from Bautzen reservoir
(30 µm mesh size), which was sampled twice a week, was used as culture medium.
From a stock stored at 4°C in the dark, the water supply in the experimental
room, with a constant temperature (19°C ± 0.5) and low light conditions, was
replaced daily. A peristaltic pump continuously provided the culture vessels with
water (renewal rate > four times per day). The water surface in the culture vessels
was strewn with cetyl alcohol to prevent the daphnids being caught on the water
surface with their hydrophobic carapace (Desmarais, 1997). The experiments
were terminated after 5 days and the final lengths (Lf) of the animals measured.
Growth curves of daphnids generally have a hyperbolic or sigmoid shape but as
a sufficient approximation, growth (in length) of D.galeata can be considered to
154
Daphnia population decline
be linear until maturity [e.g. (Stich and Lampert, 1984; Langeland et al., 1985;
Hovenkamp, 1991)]. Only some of the daphnids carried eggs at the end of the
experiments and therefore, the mean daily growth (gL) was calculated as
gL = (Lf – Li)/∆t (µm day–1)
(7)
Calculation of mortality
The general approach was similar to the discrete event model INSTAR (Vijverberg and Richter, 1982), also used by Hovenkamp (Hovenkamp, 1989, 1990) and
Boersma et al. (Boersma et al., 1996), except that (i) we only discriminated
between juvenile and adult Daphnia, (ii) juvenile growth rates were estimated
weekly in the laboratory under approximately field conditions and (iii) hatching
frequency of D.galeata was estimated according to Johnsen [(Johnsen, 1983) see
above]. The number of individuals that die during one sampling interval is estimated by comparing computed densities of juvenile and adult Daphnia with the
estimated abundance in the field.
For every sampling date (ti) from the end of May (start of the growth experiments), the hypothetical proportion of adults at the next sampling date (ti + 1) was
calculated by applying the following stepwise procedure to all measured animals:
(I)
(II)
if
Lti < SAMti
Lti + 1< SAMti + 1
then
Lti + 1 = Lti + (gL,i 3 ∆t)
juvenile at ti + 1
else
adult at ti + 1
adult at ti + 1
with Lti = measured size at time ti, Lti + 1 = calculated size at time ti + 1, SAMti = size
at maturity at ti, SAMti + 1 = size at maturity at ti + 1, ∆t = time between ti and ti + 1
in days, gL,i = mean daily growth from the culture experiments at ti.
Weekly growth rates were linearly interpolated for the second sampling date
during that week. The influence of changing temperatures on juvenile growth was
neglected in this approach because in some earlier studies, only minor differences
were found within a temperature range of 15 to 20°C (Vijverberg, 1980; Hanazato
and Yasuno, 1985; Langeland et al., 1985; Hovenkamp, 1991). This was similar to
the situation in Bautzen reservoir during the investigation period (mean temperature 14–19°C). We have no indication that D.galeata exhibits daily vertical
migration in Bautzen reservoir [compare (Hülsmann et al., 1999)].
From calculated proportions of juveniles and adults and the estimated number
of newborn daphnids since the last sampling date, a hypothetical (or potential)
abundance of juvenile and adult daphnids at time ti+1 was calculated. The absolute mortality (ind. l–1) during the sampling intervals was estimated as the difference between the hypothetical and the actual abundance of juvenile and adult
specimens of D.galeata at time ti + 1. Juvenile mortality also includes egg mortality. To evaluate the amount of juvenile and adult mortality in the population
decline of D.galeata, the absolute cumulative mortality (ind. l–1), both of juvenile
and adult Daphnia, was calculated and contrasted to cumulative hatching, starting on May 23 when the Daphnia population had stabilized at a high level.
155
S.Hülsmann and W.Weiler
Sampling intervals with negative mortality values were not considered in this estimation (mortality was set to zero). Densities of D.galeata observed in the field at
successive sampling dates, and calculated potential densities, were also used to
compute the rate of population change, r, and the ‘potential’ rate of population
change, r9:
r = (ln Nti + 1, obs. – ln Nti, obs.)/∆t (day–1)
(8)
r9 = (ln Nti + 1, calc. – ln Nti, obs.)/∆t (day–1)
(9)
with Nti + 1, obs., Nti, obs. = observed abundance at ti + 1 and ti, respectively, and
Nti + 1, calc. = calculated abundance at ti + 1. A mortality rate m (day–1) of juvenile,
adult and total mortality was calculated by subtracting r from r9 which results in:
m = (ln Nti + 1, calc. – ln Nti + 1, obs.)/∆t (day–1)
(10)
For comparison, the death rate, d, of D.galeata was also calculated as the difference between birth rate, b, and rate of population change, r, according to the egg
ratio method (Paloheimo, 1974).
Results
In early May, a steep rise in abundance of D.galeata led to the maximum value of
130 ind. l–1 (Figure 1a). Until the middle of June, Daphnia densities fluctuated
between 60 and 80 ind. l–1, followed by a sharp decline, which in a second phase
continued gradually until July. Less than 10 ind. l–1 were found during the rest of
our investigation period. SAM was high in May and early June; highest values
(>1.5 mm) were recorded immediately before the decline. Simultaneously with
the decrease in abundance, SAM declined to 0.8 mm in July (Figure 1b).
Maximum size followed the same pattern with a slight delay. Until the decline,
maximum size of daphnids was 1.8–2 mm although generally, there were only few
‘very large’ individuals. After the decline, daphnids >1.2 mm could only occasionally be found. The proportion of adults (PAD) reflected the dynamics in SAM;
values were low (~0.1) if SAM was high but after the decline, PAD increased to
0.3–0.5 (Figure 1c). Clutch size drastically decreased in early May, increased at
the beginning of June and fluctuated on a low level during the rest of our investigation period (Figure 1d).
In the culture experiments, juvenile growth was low at the end of May but
subsequently increased. Values remained high throughout June–July (Figure 2).
A significant positive correlation was found between POC and juvenile growth
(Figure 3).
Calculated rates of juvenile mortality of D.galeata were low during the whole
investigation period; negative estimates were even more numerous and higher
than positive values (Figure 4a). Adult mortality rates fluctuated strongly in late
May/early June and increased to 0.58 day–1 about 1 week before the Daphnia
population crashed. Values remained high (maximum value 0.61 day–1) during the
156
Daphnia population decline
Fig. 1. Temporal pattern in development of (a) abundance (ind. l–1), (b) size structure, (c) proportion of adults (PAD) and (d) clutch size (eggs per adult female) of D.galeata in Bautzen reservoir. In
1b, the crosses represent the length of the largest daphnids measured (largest 10% of total number
measured) and the line represents size at maturity (SAM) (mm).
decline. Mortality rates of adult daphnids, except for a few sampling intervals,
remained at a comparatively high level throughout summer.
With regard to total mortality of D.galeata, values were generally lower than
adult mortality rates. Highest (maximum 0.31 day–1) rates were recorded during
157
S.Hülsmann and W.Weiler
Fig. 2. Mean daily growth (mm d–1) of D.galeata in culture experiments plotted against the date when
the experiment started.
Fig. 3. Regression of mean daily growth (mm d–1) against POC <30 µm (mg l–1).
the second phase of the decline (Figure 4b). The estimated values were almost
identical to the death rate calculated according to the Paloheimo model.
Despite low mortality rates, absolute values of juvenile mortality before the
population decline were higher than adult mortality due to the high abundance
of juvenile daphnids (Figure 5). Only during the first sampling interval of the
decline (June 12–16) did absolute values of juvenile mortality equal adult
158
Daphnia population decline
Fig. 4. Mortality rate (d–1) of (a) juvenile and adult Daphnia and (b) the whole population of
D.galeata in Bautzen reservoir. In 4b, the death rate according to the Paloheimo-model is also shown.
Note the different scaling of the y-axis in a and b.
mortality but thereafter, juvenile mortality was virtually absent. Cumulative adult
mortality drastically increased during the population decline of D.galeata. In July
(at low Daphnia abundance), this increase diminished.
Following maximum abundance, recruitment of D.galeata (assuming no egg
mortality) was low in late May (Figure 5). About 1 week before the population
decline, recruitment increased but was low again during the first sampling interval when Daphnia declined (June 12–16). After that, recruitment was low but
continuous. In late July, cumulative mortality and recruitment increased with a
similar slope; the Daphnia population had stabilized at a low level.
Discussion
In this study, we have shown that the decline of a population of D.galeata in
Bautzen reservoir was caused mainly by enhanced mortality of adult daphnids,
whereas juvenile mortality was low during the whole investigation period.
Contrary to the general contention that non-predatory mortality of Daphnia is a
direct consequence of low food conditions and primarily concerns juvenile stages,
159
S.Hülsmann and W.Weiler
Fig. 5. Cumulative mortality (ind. l–1) of juvenile and adult Daphnia and cumulative recruitment of
D.galeata in Bautzen reservoir, starting 10 days after the maximum abundance.
we suppose that increasing mortality of adults as an indirect effect of low food
conditions may be related to ageing of Daphnia populations.
The observed patterns of Daphnia mortality gave little indication of direct
effects of food limitation. If present at all, these should be restricted to the period
May–beginning of June. At this time, the POC in the edible size fraction declined
to <0.5 mg C l–1 and lowest values of fatty acids were recorded (Weiler et al., in
preparation). Hence, food limitation is more likely to be a result of low quality
than low quantity of edible particles. As we have no growth data before the
middle of May, we can only speculate about the mortality pattern during the first
sampling intervals of our field investigation. We suppose that during that time,
starvation mortality of adult daphnids (at least, as an interaction with senescence,
see below) might have occurred, as fecundity declined to very low values, thus
indicating severe food limitation. Moreover, it is known that adult daphnids may
adapt their feeding appendages to decreasing food conditions only with considerable time lags (Voigt and Benndorf, 1999). While the enlargement of the filtering area occurred very rapidly in small size classes of Daphnia during the
clear-water stage, adults were at a higher risk of suffering from starvation as their
filter combs were still adapted to high food conditions. When we started our
culture experiments, mortality rates of juvenile daphnids were low, but present
until the beginning of June. We suppose that this did not differ from the situation
in early May. Starvation mortality is more likely for juvenile stages of Daphnia
(Threlkeld, 1976), and low growth rates in late May indeed indicate low fitness
[compare (Lampert and Trubetskova, 1996)]. Although virtually no mortality was
observed during the growth experiments, somatic growth was clearly food-limited
for a short time. However, juvenile growth, and hence fitness, increased well
before the Daphnia population declined. Moreover, juvenile mortality decreased
160
Daphnia population decline
during the decline and therefore contributed only marginally to the midsummer
decline of D.galeata. Negative mortality of juveniles in June may be caused by
hatching from resting eggs. In late July, on the other hand, negative values can be
explained by the difference between sampling interval, ∆t and egg development
time, D. Recruitment is underestimated if ∆t >D and in this case, results in negative mortality of juveniles (see below).
Mortality of adults increased shortly before the population decline of
D.galeata. At that time, SAM was still high and the proportion of adults was low.
Thus, disproportionate loss of adults should have a great impact on the reproductive capacity of the Daphnia population. The sudden rise in Daphnia abundance occurred during one week in May. Hence, a great part of the population
consisted of this ‘peak cohort’ and had about the same age. This was due to high
fecundity and a sharp rise in water temperature, which reduced egg development
time. We could show that fecundity and recruitment of D.galeata was low after
the population had established a high abundance. Hence, the population did not
rejuvenate to a great extent until the beginning of June. This was, however, not
reflected in a rising proportion of adults. We suppose that ageing of the population is masked by high, and even increasing SAM immediately before the
decline. Moreover, the method of determining SAM has its limits under
conditions of severe food limitation, such that it may overestimate the size when
individuals become adult (Stibor and Lampert, 1993). When adult mortality
increased, the peak-cohort was 4–5 weeks old. Under conditions of food limitation, this might correspond to the mean life-span of D.galeata, as in life history
trials with D.galeata under conditions of the clear-water stage, mean duration of
life was 28 days (S.Hülsmann, unpublished results). This is in agreement with
Threlkeld, who proposed an interaction between age-specific and starvationinduced mortality (Threlkeld, 1976). In contrast, Vijverberg (Vijverberg, 1976)
reported that longevity of D.galeata at low food levels was >7 weeks (at 20°C).
However, in this early study of Daphnia life history, as a measure of food quantity only chlorophyll was measured. Moreover, media were replaced only once
per week. Thus, food conditions were probably underestimated because an
unknown amount of bacteria was present (and probably growing) in the culture
medium. In life history experiments with D.pulex, Lynch (Lynch, 1989) found that
age-specific survival drastically decreased once the mean life-span (50 days at
0.5–1 mg C l–1, lower values at food carbon concentration <0.5 mg C l–1) was
reached. There might also be an additional maternal effect, as maximum life span
at low food was reduced in D.pulex when the mothers of the experimental generation were grown in high food conditions (Lynch and Ennis, 1983). This should
have been similar for the mothers of the peak cohort. Thus, if mean life-span of
the peak cohort of D.galeata matched the beginning of the decline, adult mortality might in part be due to senescence. The decline of the maximum size of individuals found in the population of D.galeata demonstrates the vanishing of the
biggest, and thus oldest size classes during the decline. Voigt and Benndorf
proposed yet another mechanism that might enhance adult mortality (Voigt and
Benndorf, 1999). With their feeding appendages being adapted to low food
conditions (filter combs dense and large), adult daphnids may not be able to cope
161
S.Hülsmann and W.Weiler
with rising particle concentrations after the clear-water stage. Clogging of their
filter combs and increased furca movements might lead to enhanced energy loss,
resulting in reduced fitness and possibly, reduced mean life-span.
Our finding that mainly adult mortality was responsible for the midsummer
decline of D.galeata in Bautzen reservoir contrasts with several studies, where
juvenile mortality was found to be decisive for the decline of Daphnia populations e.g. in Lago Maggiore (De Bernardi, 1974), Lake Vechten (Hovenkamp,
1989) and Lake Tjeukemeer (Boersma et al., 1996). In Lago Maggiore and Lake
Vechten, invertebrate predators were considered to be responsible for the
enhanced mortality of juvenile daphnids. Hovenkamp (Hovenkamp, 1990) could
show that predation by Chaoborus and Leptodora accounted for total juvenile
mortality of Daphnia in Lake Vechten for most of the sampling season, confirming results of Dodson (Dodson, 1972), who estimated >90% of Chaoborus predation in the mortality of D.rosea. However, in these studies, Daphnia densities
were distinctly lower [(maximum values of ~10 ind. l–1 in Lago Maggiore (De
Bernardi, 1974), ~30 ind. l–1 in Lake Vechten (Hovenkamp, 1989), ~20 ind. l–1 in
Leechmere pond (Dodson, 1972)] than spring densities of D.galeata in Bautzen
reservoir, where predation impact by invertebrate predators (mainly Leptodora
kindtii) only contributed up to about 40% of Daphnia mortality in 1997
(A.Wagner, Inst. of Hydrobiology, unpublished results). In Lake Tjeukemeer,
Daphnia abundance was comparable with, or even higher than in Bautzen reservoir and—another similarity—0+ fish are the main vertebrate zooplanktivores
(Boersma et al., 1996). However, contrary to the results presented in this study,
Boersma et al. computed the highest mortality for daphnids <1 mm (Boersma et
al., 1996). As this could not be explained by 0+ fish consumption, they concluded
that non-predatory sources of mortality, especially starvation, accounted for this
high juvenile mortality. Boersma and Vijverberg (Boersma and Vijverberg,
1994a) showed that Daphnia in Lake Tjeukemeer was food-limited during most
of the season. However, the significance of increasing mortality of D.galeata at
decreasing food levels found in life-table experiments (Boersma and Vijverberg,
1994b) cannot be judged for the field situation. Fecundity and birth rates of
D.galeata and the hybrid D.galeata 3 cucullata in Tjeukemeer were low prior to
and during the midsummer decline (Boersma, 1995; Boersma and Vijverberg,
1995). This pattern, and also the phytoplankton succession in spring (dominance
of diatoms) and summer (dominance of blue-greens), was similar in Tjeukemeer
and in Bautzen reservoir (Boersma and Vijverberg, 1995; Böing et al., 1998).
Hence, the contradictory results of Boersma et al. (Boersma et al., 1996) and this
study concerning size-selective mortality are unlikely to be caused by bottom-up
factors. They might, however, at least in part, be explained by different computation of Daphnia growth. Obviously, this is crucial for the estimates. Boersma et
al. used laboratory derived data on juvenile growth, established a ratio between
SAM and the maximum size of daphnids, and computed growth of animals in the
field using a von Bertalanffy equation (Boersma et al., 1996). In their model, small
SAM (determined in field samples) reduced somatic growth. In contrast, we
found the highest growth rates of D.galeata after the decline in SAM. No
information concerning the relationship between SAM and juvenile growth is
162
Daphnia population decline
available from other studies. However, as the number of juvenile instars is not
fixed but depends on food availability (Boersma and Vijverberg, 1994b), juvenile growth is retarded at low food conditions (Neill, 1981; Lynch, 1989; Gliwicz
and Lampert, 1990), and high SAM can result from both good and bad food
conditions (McCauley et al., 1990), the proposed relation between juvenile
growth and SAM may not be as clear as assumed. If it is supposed that Boersma
et al. underestimated growth, mortality of small size classes would be overestimated in their study, and vice versa for large size classes (Boersma et al., 1996).
In this study, the highest mortality rates of adult daphnids (>0.5 day–1) mean that
many more adults died during the sampling interval than were present at the
beginning of the interval. This implies that most daphnids that become adult
during this sampling interval die. However, it is also possible that animals that
(due to our growth data) were about to become adult during a sampling interval
died before they actually became adult. In this sense, a part of adult mortality
may actually be mortality of late juvenile instars. However, this possible mechanism may be counteracted by a possible overestimation of SAM and even if this is
not the case, we still have no indication that the smallest juvenile instars died preferentially. This is consistent with the results of McCauley et al. (McCauley et al.,
1990). Recently, the contention that there is a generally higher vulnerability of
cladoceran juveniles, compared with adults, to starvation and crowding [also
implied by (Boersma et al., 1996)] has been questioned by the results of Matveev
and Gabriel (Matveev and Gabriel, 1994), who found that demographic mechanisms vary between species.
Low fecundity of Daphnia populations prior to their decline has frequently
been reported (Threlkeld, 1985; Tessier, 1986; Luecke et al., 1990; Wu and Culver,
1994) and attributed to depletion of food resources (Lampert et al., 1986; Sommer
et al., 1986; Tessier, 1986). However, this causal connection could not be drawn
for the situation in Bautzen reservoir (Mehner et al., 1998a). Due to high phosphorus loading (Benndorf, 1995), the POC <30 µm (see above) is always above
threshold concentrations for Daphnia reproduction known from other systems
(Gliwicz and Lampert, 1990). However, we found high correlations of fatty acid
food components with birth rate of D.galeata in the lower concentration range in
1997 (Weiler et al., in preparation). We therefore conclude that bottom-up limitation (either quantitative or qualitative) of Daphnia fecundity is indeed a prerequisite for a midsummer decline. However, in accordance with Tessier (Tessier,
1986), we infer that the effects of starvation of Daphnia on the initiation of a
midsummer decline only act indirectly and with time lags; ageing of the population leads to increased mortality of adults, but animals do not starve to death
(neither juveniles nor adults). Later in July, low clutch size can no longer be
attributed to food shortage. As clutch size is also a function of body size, it must
be low after SAM drastically decreased.
Interestingly, adult mortality of D.galeata remained high during the second
phase of the decline and even later in July, when the population had stabilized at
a low level. However, it was found in several other studies of midsummer declines
of daphnids, that predation by fish, especially 0+ fish (assumed to be size-selective), although not strong enough to cause the decline, may further decrease
163
S.Hülsmann and W.Weiler
Daphnia abundance and keep it at a low level (Luecke et al., 1990; Wu and Culver,
1994). This was also confirmed by our own results in Bautzen reservoir (Hülsmann and Mehner, 1997; Mehner et al., 1998a). Invertebrate predators in this situation may also contribute to the suppression of daphnids (De Stasio et al., 1995).
This is not necessarily contradictory to our finding that exclusively adult daphnids suffered high mortality. As SAM of D.galeata declined to 0.8 mm, Leptodora
kindtii, the only significant invertebrate predator in Bautzen reservoir at that
time, may well be selective for adult daphnids (Lunte and Luecke, 1990),
although this species is generally considered to prey on small organisms. In
Bautzen reservoir, 0+ percids are considered to be important planktivorous
predators (Mehner et al. 1996, 1998a). Their consumption and size selection in
1997 will be considered in a forthcoming paper.
One major prerequisite for applying our approach is that the sampling interval
has to be shorter than (or equal to) egg development time. Otherwise, recruitment and hence, mortality of daphnids will be underestimated, and juvenile
mortality will be negative (as was the case in the second half of July). Meeting
this requirement also ensures that hatched daphnids will not become adult during
the sampling interval and hence, that their growth does not need to be considered.
However, analysis of population dynamics based on the egg ratio technique will
only give reliable results if the sampling interval is more or less equal to egg
development time (Keen and Nasser, 1981; Gabriel et al., 1987), which implies
that the sampling interval has to be adjusted to prevailing temperatures in the
water under study.
Another requirement to obtain reliable population dynamics estimates is the
accurate determination of population densities. This point has been stressed in
many studies (Lynch, 1982, 1983; Taylor, 1988) and was considered in our investigation by using pooled samples, both horizontally and vertically. However, in so
doing we were unable to consider the influence of vertical differences in temperature and distribution of Daphnia on egg development time and growth, but as
temperature gradients were low according to our own measurements [see also
(Benndorf and Henning, 1989; Köhler, 1992)], only marginal effects can be
expected.
Our method of determining juvenile growth rate of D.galeata did not consider
the influence of temperature. This might be acceptable, as in former studies only
minor effects were found in the temperature range present during our investigation (Vijverberg, 1980; Hanazato and Yasuno, 1985; Hovenkamp 1991), though
uncertainties remain. During our study, mean water temperature of Bautzen
reservoir was lower (14–16°C) than in the growth experiments until June 9 and
hence, application of growth rates to field data up to that date remains somewhat
uncertain. If growth in situ was lower than estimated in the laboratory, our estimated theoretical abundance of adults and hence, mortality would be too high
and vice versa for the juveniles. However, during severe food limitation [which
might have been the case at the end of May, (Weiler et al. in preparation)],
juvenile development of Daphnia may be retarded at increasing temperatures
(Neill, 1981), so growth in situ might also have been better than estimated in the
laboratory (at higher temperatures). In this case, estimated theoretical
164
Daphnia population decline
abundance of adults and hence, mortality would be underestimated and vice versa
for the juveniles. Therefore, no attempt to correct our growth data to actual
temperatures in Bautzen reservoir was made. If our approach should be applied
to Daphnia populations confronted with a much wider temperature range in the
environment or under conditions of vertical migration [compare (Loose and
Dawidowicz, 1994)], experimental expense in Daphnia cultures would have to be
enhanced, at best, similar to the design of Stich and Lampert (Stich and Lampert,
1984).
Another possible source of error is the elimination of inedible particles by
filtering the lake seston with a 30 µm mesh gauze. Filtration of daphnids in the
field might be hindered by the inedible seston fraction, resulting in energy loss
and/or lower energy intake and consequently decreased growth compared with
culture animals. Further, as lake water was only sampled twice per week, the lake
seston was ageing before being offered to the daphnids. Therefore, food
conditions might have been better than in the field. If, via food or temperature
effects, mortality of adults was overestimated, our conclusion that primary adult
mortality was important, would be questionable. However, our growth estimates
can be considered as realistic compared with results from the literature (Stich and
Lampert, 1984; Langeland et al., 1985) and the development of the Daphnia
population structure fits our explanations. ‘Hard’ proof of our conclusion would
be possible using sediment traps. Unfortunately, no data for the critical period
are available for 1997.
In summary, we conclude that the succession of events leading to a midsummer
decline of Daphnia is as follows. (i) A quick increase in Daphnia abundance leads
to the formation of a strong ‘peak cohort’ of about the same age. (ii) During the
clear-water phase, food conditions deteriorate, fecundity declines and hence,
recruitment is low; juvenile mortality during this period is low, but present. (iii)
Adult mortality increases when the ‘peak cohort’ reaches its mean life-span,
which is reduced due to interactions between age-specific and starvation-induced
mortality. At this point, Daphnia population dynamics can no longer be explained
without the onset of size-selective predation. Hence, the timing between
enhanced mortality due to senescence on the one hand and predation on the
other hand, both directed to adult daphnids, may be decisive. This conclusion is
supported by a long-term data analysis of Bautzen reservoir (Benndorf et al.
submitted), which showed that timing of both sources of mortality and hence, the
occurrence of a midsummer decline, is controlled by the water temperature in the
preceding winter and early spring.
Acknowledgements
We would like to thank H.Voigt for providing the POC data and for help during
the sampling and the Daphnia cultures. T.Mehner, T.Petzoldt, A.Wagner,
R.Radke, J.Benndorf and two anonymous referees gave valuable comments on
earlier drafts of the manuscript. The work was supported by project Be 1671/2-2
of the Deutsche Forschungsgemeinschaft (DFG, Germany).
165
S.Hülsmann and W.Weiler
References
Argentesi,F., De Bernardi,R. and Di Cola,G. (1974) Mathematical models for the analysis of population dynamics in species with continuous recruitment. Mem. Ist. Ital. Idrobiol., 31, 245–275.
Benndorf,J. (1995) Possibilities and limits for controlling eutrophication by biomanipulation. Internat.
Revue Ges. Hydrobiol., 80, 519–534.
Benndorf,J. and Henning,M. (1989) Daphnia and toxic blooms of Microcystis aeruginosa in Bautzen
Reservoir. Int. Revue Ges. Hydrobiol., 74, 233–248.
Benndorf,J., Wagner,A., Mehner,T. and Kranich,J. Evidence for temperature control of the
midsummer decline of Daphnia galeata by a long-term data analysis from the biomanipulated
Bautzen reservoir (Germany). Submitted.
Boersma,M. (1995) Competition in natural populations of Daphnia. Oecologia, 103, 309–318.
Boersma,M. and Vijverberg,J. (1994a) Seasonal variation in the condition of two Daphnia species and
their hybrid in a eutrophic lake: evidence for food limitation. J. Plankton Res., 16, 1793–1809.
Boersma,M. and Vijverberg,J. (1994b) Resource depression in Daphnia galeata, Daphnia cucullata
and their interspecific hybrid: life history consequences. J. Plankton Res., 16, 1741–1758.
Boersma,M. and Vijverberg,J. (1995) The significance of nonviable eggs for Daphnia population
dynamics. Limnol. Oceanogr., 40, 1215–1224.
Boersma,M., van Tongeren,O.F.R. and Mooij,W.M. (1996) Seasonal patterns in the mortality of
Daphnia species in a shallow lake. Can. J. Fish. Aquat. Sci., 53, 18–28.
Böing,W., Wagner,A., Voigt,H., Deppe,T. and Benndorf,J. (1998) Phytoplankton responses to grazing
by Daphnia galeata in the biomanipulated Bautzen reservoir. Hydrobiologia, 389, 101–114.
Bottrell,H.H., Duncan,A., Gliwicz,Z.M., Grygierek,E., Herzig,A., Hillbricht-Ilkowska,A.,
Kurasawa,H., Larsson,P. and Weglenska,T. (1976) A review of some problems in zooplankton
studies. Norw. J. Zool., 24, 419–456.
De Bernardi,R. (1974) The dynamics of a population of Daphnia hyalina Leydig in Lago Maggiore,
Northern Italy. Mem. Ist. Ital. Idrobiol., 31, 221–243.
De Stasio,B.T.Jr, Rudstam,L.G., Haning,A., Soranno,P. and Allen,Y.C. (1995) An in situ test of the
effect of food quality on Daphnia population growth. Hydrobiologia, 307, 221–230.
Desmarais,K.H. (1997) Keeping Daphnia out of the surface film with cetyl alcohol. J. Plankton Res.,
19, 149–154.
Dodson,S.I. (1972) Mortality in a population of Daphnia rosea. Ecology, 53, 1011–1023.
Gabriel,W., Taylor,B.E. and Kirsch-Prokosch,S. (1987) Cladoceran birth and death rates estimates:
experimental comparison of egg-ratio methods. Freshwater Biol., 18, 361–372.
Gliwicz,M.Z. and Lampert,W. (1990) Food thresholds in Daphnia species in the absence and presence
of blue-green filaments. Ecology, 71, 691–702.
Gliwicz,M.Z. and Pijanowska,J. (1989) The role of predation in zooplankton succession. In Sommer
(ed.), Plankton Ecology. Succession in Plankton Communities. Springer, Berlin, pp. 253–295.
Gliwicz,Z.M., Ghilarov,A. and Pijanowska,J. (1981) Food and predation as major factors limiting two
natural populations of Daphnia cucullata Sars. Hydrobiologia, 80, 205–218.
Hall,D.J. (1964) An experimental approach to the dynamics of a natural population of Daphnia
galeata mendotae. Ecology, 45, 94–112.
Hanazato,T. and Yasuno,M. (1985) Effect of temperature in the laboratory studies on growth, egg
development and first parturition of five species of Cladocera. Japan J. Limnol., 46, 185–191.
Haney,J.F. and Hall,D.J. (1973) Sugar-coated Daphnia: a preservation technique for Cladocera.
Limnol. Oceanogr., 18, 331–333.
Herzig,A. and Auer,B. (1990) The feeding behaviour of Leptodora kindti and its impact on the
zooplankton community of Neusiedler See (Austria). Hydrobiologia, 198, 107–117.
Hovenkamp,W. (1989) Instar-dependent mortalities of coexisting Daphnia species in Lake Vechten,
The Netherlands. J. Plankton Res., 11, 487–502.
Hovenkamp,W. (1990) Instar-specific mortalities of coexisting Daphnia species in relation to food and
invertebrate predation. J. Plankton Res., 12, 483–494.
Hovenkamp,W. (1991) How useful are laboratory-determined growth curves for the estimation of
size-specific mortalities in Daphnia populations? J. Plankton Res., 13, 353–361.
Hülsmann,S. and Mehner,T. (1997) Predation impact of underyearling perch (Perca fluviatilis) on a
Daphnia galeata population in a short-term enclosure experiment. Freshwater Biol., 38, 209–219.
Hülsmann,S., Mehner,T., Worischka,S. and Plewa,M. (1999) Is the difference in population dynamics
of Daphnia galeata in littoral and pelagic areas of a long-term biomanipulated reservoir affected by
age-0 fish predation? Hydrobiologia, in press.
Johnsen,G. (1983) Egg age distribution, the direct way to cladoceran birth rates. Oecologia, 60,
234–236.
166
Daphnia population decline
Keen,R. and Nassar,R. (1981) Confidence intervals for birth and death rates estimated with the eggratio technique for natural populations of zooplankton. Limnol. Oceanogr., 26, 131–142.
Köhler,J. (1992) Influence of turbulent mixing on growth and primary production of Microcystis
aeruginosa in the hypertrophic Bautzen Reservoir. Arch. Hydrobiol., 123, 413–429.
Lampert,W. and Trubetskova,I. (1996) Juvenile growth rate as a measure of fitness in Daphnia. Funct.
Ecol., 10, 631–635.
Lampert,W., Fleckner,W., Rai,H. and Taylor,B.E. (1986) Phytoplankton control by grazing zooplankton: a study on the spring clear-water phase. Limnol. Oceanogr., 31, 478–490.
Langeland,A., Koksvik,J. and Olsen,Y. (1985) Post-embryonic development and growth rates of
Daphnia pulex De Geer and Daphnia galeata Sars under natural food conditions. Verh. Int. Ver.
Limnol., 22, 3124–3130.
Loose,C. and Dawidowicz,P. (1994) Trade-offs in vertical migration by zooplankton: the costs of
predator avoidance. Ecology, 75, 2255–2263.
Luecke,C., Vanni,M.J., Magnuson,J.J., Kitchell,J.F. and Jacobson,P.T. (1990) Seasonal regulation of
Daphnia populations by planktivorous fish: implications for the spring clear-water phase. Limnol.
Oceanogr., 35, 1718–1733.
Lunte,C.C. and Luecke,C. (1990) Trophic interactions of Leptodora in Lake Mendota. Limnol.
Oceanogr., 35, 1091–1100.
Lynch,M. (1982) How well does the Edmondson-Paloheimo model approximate instantaneous birth
rates? Ecology, 63, 12–18.
Lynch,M. (1983) Estimation of size-specific mortality rates in zooplankton populations by periodic
sampling. Limnol. Oceanogr., 28, 533–545.
Lynch,M. (1989) The life history consequences of resource depression in Daphnia pulex. Ecology, 70,
246–256.
Lynch,M. and Ennis,R. (1983) Resource availability, maternal effects, and longevity. Exp. Geront.,
18, 147–165.
Matveev,V. and Gabriel,W. (1994) Competitive exclusion in Cladocera through elevated mortality of
adults. J. Plankton Res., 16, 1083–1094.
McCauley,E., Murdoch,W.W. and Nisbet,R.M. (1990) Growth, reproduction, and mortality of
Daphnia pulex Leydig: life at low food. Funct. Ecol., 4, 505–514.
Mehner,T., Schultz,H., Werner,M.-G., Wieland,F., Herbst,R. and Benndorf,J. (1996) Do 0+ percids
couple the trophic cascade between fish and zooplankton in the top-down manipulated Bautzen
reservoir (Germany)? Publ. Espec. Inst. Esp. Oceanogr., 21, 243–251.
Mehner,T., Hülsmann,S., Worischka,S., Plewa,M. and Benndorf,J. (1998a) Is the midsummer decline
of Daphnia really induced by age-0 fish predation? Comparison of fish consumption and Daphnia
mortality and life history parameters in a biomanipulated reservoir. J. Plankton Res., 20, 1797–1811.
Mehner,T., Plewa,M., Hülsmann,S. and Worischka,S. (1998b) Gape-size dependent feeding of age-0
perch (Perca fluviatilis) and age-0 zander (Stizostedion lucioperca) on Daphnia galeata. Arch.
Hydrobiol., 142, 191–207.
Mills,E.L. and Forney,J.L. (1983) Impact on Daphnia pulex of predation by young yellow perch in
Oneida Lake, New York. Trans. Am. Fish. Soc., 112, 151–161.
Neill,W.E. (1981) Developmental responses of juvenile Daphnia rosea to experimental alteration of
temperature and natural seston concentration. Can. J. Fish. Aquat. Sci., 38, 1357–1362.
Paloheimo,J.E. (1974) Calculation of instantaneous birth rate. Limnol. Oceanogr., 19, 692–694.
Sommer,U., Gliwicz,Z.M., Lampert,W. and Duncan,A. (1986) The PEG-model of seasonal succession of planktonic events in fresh waters. Arch. Hydrobiol., 106, 433–471.
Stibor,H. and Lampert,W. (1993) Estimating the size at maturity in field populations of Daphnia
(Cladocera). Freshwater Biol., 30, 433–438.
Stich,H.-B. and Lampert,W. (1984) Growth and reproduction of migrating and non-migrating
Daphnia species under simulated food and temperature conditions of diurnal vertical migration.
Oecologia, 61, 192–196.
Taylor,B.E. (1988) Analysing population dynamics of zooplankton. Limnol. Oceanogr., 33,
1266–1273.
Tessier,A.J. (1986) Comparative population regulation of two planktonic cladocera (Holopedium
gibberum and Daphnia catawba). Ecology, 67, 285–302.
Tessier,A.J., Henry,L.L., Goulden,C.E. and Durand,M.W. (1983) Starvation in Daphnia: energy
reserves and reproductive allocation. Limnol. Oceanogr., 28, 667–676.
Threlkeld,S.T. (1976) Starvation and the size structure of zooplankton communities. Freshwater Biol.,
6, 489–496.
Threlkeld,S.T. (1979) The midsummer dynamics of two Daphnia species in Wintergreen Lake,
Michigan. Ecology, 60, 165–179.
167
S.Hülsmann and W.Weiler
Threlkeld,S.T. (1985) Resource variation and the initiation of midsummer declines of cladoceran
populations. Arch. Hydrobiol. Beih. Ergebn. Limnol., 21, 333–340.
Vijverberg,J. (1976) The effect of food quantity and quality on the growth, birth-rate and longevity
of Daphnia hyalina Leydig. Hydrobiologia, 51, 99–108.
Vijverberg,J. (1980) Effect of temperature in laboratory studies on development and growth of
cladocera and copepoda from Tjeukemeer, The Netherlands. Freshwater Biol., 10, 317–340.
Vijverberg,J. and Richter,A.F. (1982) Population dynamics and production of Daphnia hyalina Leydig
and Daphnia cucullata Sars in Tjeukemeer. Hydrobiologia, 95, 235–259.
Voigt,H. and Benndorf,J. (1990) Differences in plasticity of adult and juvenile daphnids in changing
the morphology of their filtercombs. Verh. Internat. Verein. Limnol. (Dublin 1998), in press.
Wu,L. and Culver,D.A. (1994) Daphnia population dynamics in Western Lake Erie: Regulation by
food limitation and yellow perch predation. J. Great Lakes Res., 20, 537–545.
Received on February 18, 1999; accepted on August 2, 1999
168