Daphnia defense strategies in fishless lakes and ponds: one size does not fit all

Abstract

Body size and neck spine development in Daphnia greatly influence this animal’s vulnerability to predation by the size-selective invertebrate planktivore Chaoborus. We develop a stage-classified matrix population model for Daphnia that investigates the interaction and evolution of these two traits in situations (fishless lakes and ponds) where Chaoborus predation constitutes the major source of mortality. This model produces fitness landscapes for these traits in ten distinct Daphnia environments that are characterized by Chaoborus size (medium-sized Chaoborus americanus or large Chaoborustrivittatus), Chaoborus density (0–1.0 L⁻¹) and food level (high or low). Larger Daphnia phenotypes are favored in both high and low food environments that contain C. americanus, and also in a high food situation with C. trivittatus. The environments with C. trivittatus and low food availability, however, select for very small, as well as very large, Daphnia phenotypes (small phenotypes are favored more at high Chaoborus densities), but not those that are intermediate in size. The development of neck spines is advantageous in all situations with Chaoborus, but high food environments that contain C. americanus favor their elimination following juvenile development, while the other model environments favor their retention (to various degrees) after maturity. These model predictions describe alternative antipredator strategies, two of which correspond closely with phenotypic patterns exhibited by two species of Daphnia (Daphnia pulex and Daphnia minnehaha) that commonly coexist with Chaoborus in fishless lakes and ponds.

Received January 18, 2005; accepted in principle March 24, 2005; accepted for publication May 11, 2005; published online May 20, 2005
 Communicating editor: K.J. Flynn

INTRODUCTION

Natural selection has produced a variety of creative responses in zooplankton to the threat of predation. These involve the development of alternative phenotypes that decrease the impacts of various predators through changes in prey morphology, behavior and life history. Nowhere is the diversity and complexity of these antipredator defenses more evident than in species of the cladoceran genus Daphnia (Fig. 1), where evolution has forged an assortment of adaptations to deal with different types of predators. Among these are changes in the size and shape of the head or helmet (O’Brien et al., 1979; Grant and Bayly, 1981; Hebert and Grewe, 1985; Dodson, 1988a; Tollrian, 1990, 1994; Beaton and Hebert, 1997), the development of neck spines (Nachenzähne) on the dorsal margin of the head (Krueger and Dodson, 1981; Havel and Dodson, 1984; Tollrian, 1995a), increases or decreases in body size (Dodson, 1989; Riessen, 1999a,b), alterations in the patterns of vertical migration behavior (Dodson, 1988b; Nesbitt et al., 1996; De Meester et al., 1999), and adjustments in various life history traits (Riessen, 1999a,b; Tollrian and Dodson, 1999). Chemicals (kairomones) released by the predators themselves often directly induce (or at least greatly influence) these defensive responses in Daphnia (Tollrian and Dodson, 1999).

The various traits that function as antipredator defenses do not evolve in isolation from one another, but as integrated suites of characters that constitute an adaptive strategy to a specific set of environmental conditions, such as a particular type and intensity of predation. The form of this adaptive strategy is greatly influenced by both trade-offs that occur as a result of the creation and implementation of these defenses and interactions that occur among different types of defensive traits. Trade-offs result from linkages between traits in which a change in one that produces an increase in fitness is coupled to a change in another that produces a decrease in fitness, thus constraining the simultaneous evolution of optimal conditions in both traits (Stearns, 1992). For example, an increase in body size will decrease the vulnerability of Daphnia to many tactile invertebrate predators; however, it may also result in delayed reproduction and smaller clutch sizes (Riessen, 1999a,b), as well as an increased vulnerability to visually-oriented predators such as fish. Interactions among antipredator defenses occur when the presence of one type of defense influences the effectiveness and usefulness of another. If a Daphnia grows large enough to become invulnerable to a particular invertebrate predator, for instance, then other defenses that are also designed to thwart the predator (e.g. changes in helmet size and shape, development of neck spines and altered vertical migration behavior) become unnecessary and possibly costly. The evolution of an appropriate suite of defensive traits will, therefore, depend on the effectiveness of the individual defenses, their interactions with one another and trade-offs that occur in their development.

In fishless lakes and ponds, Daphnia body size is not constrained by the effects of size-selective fish predation, which otherwise would allow only for the presence of smaller (less visible) phenotypes. Predation pressure on zooplankton in these fishless situations is often exerted primarily (sometimes nearly exclusively) by larvae of the phantom midge Chaoborus. Predation by these tactile, gape-limited planktivores (see below), whose densities can exceed 1 L⁻¹ in many natural systems, can be intense, resulting in high mortality rates in Daphnia and other zooplankton prey (Dodson, 1972; Lynch, 1979; Riessen, 1999a). In general, invertebrate predators in plankton communities tend to select relatively small zooplankton prey because of the difficulty for these tactile predators in handling and ingesting large zooplankton. This would, therefore, favor the presence of only large prey phenotypes in fishless communities where invertebrate predation pressure is intense. The form of size-selective predation exhibited by Chaoborus, however, is different from this typical pattern, which creates potential opportunities for a variety of morphological responses in Daphnia that can decrease predation pressure. The size of the predator’s mouth gape limits its ability to handle and ingest large prey or those with defensive spines (Swift, 1992), as is characteristic for invertebrate planktivores. Thus, relatively large Daphnia (compared to Chaoborus mouth gape diameter) and those that develop a neck spine [extension from dorsal margin of head that projects into one or more pointed ‘teeth’ (see Fig. 1)] may for the most part avoid being eaten. These defenses (neck spines and larger size) are induced in several Daphnia species as a direct response to kairomones produced by the predator (Krueger and Dodson, 1981; Havel, 1985; Beaton and Hebert, 1997; Riessen, 1999b). Chaoborus, however, is also a stationary, ambush predator (unlike most invertebrate planktivores), one consequence of which is that they have low encounter rates with very small Daphnia, which have relatively slow swimming speeds [(Pastorok, 1981); see below]. Thus, small size for these prey may also be advantageous, although it is unclear under what conditions (if any) such a defense strategy would be better than that of large body size.

In this article, we examine the interaction of body size and neck spines as defensive adaptations in Daphnia in response to predation by Chaoborus larvae. This is accomplished by developing a demographic model for Daphnia (stage-classified matrix population model) that simultaneously incorporates the influences of variations in body size and neck spine development on Daphnia fitness in the presence of Chaoborus predation. This includes both (i) the direct effect of these traits on mortality caused by Chaoborus and (ii) trade-offs associated with these traits, such as changes in the number and size of eggs produced and the duration of the various instars. Thus, both the benefits and costs of changes in body size and neck spine development are included in the model (Table I). We use λ(= e^r), the finite population growth rate of Daphnia (which is also the dominant eigenvalue of the projection matrix of the model), as our measure of fitness (Stearns, 1992; McGraw and Caswell, 1996; Riessen, 1999a; Caswell, 2001). The model examines different possible antipredator strategies (various combinations of Daphnia body size and neck spine development) and produces adaptive landscapes that predict the relative fitness of each strategy in different environmental conditions. These conditions are characterized by differences in Chaoborus body size (larger and smaller species with different mouth gape diameters) and density (0–1.0 L⁻¹) and by variation in the food environment for Daphnia (high and low food). We then discuss the implications of these predictions and strategies with respect to different phenotypic profiles found in Daphnia that coexist with Chaoborus.

Daphnia and Chaoborus in fishless lakes and ponds

Daphnia are medium to large-sized (≈1–3 mm adult body length) cladoceran zooplankton that are common and ecologically important in lakes and ponds, both those that contain fish and those that are fishless. During ‘favorable’ periods of the year they reproduce parthenogenetically, producing a variable number of subitaneous (immediately-hatching) eggs in each clutch (one clutch in each adult instar) that is determined by food abundance and the size of the adult female (Green, 1956; Hebert, 1978; Lynch, 1989). Growth is indeterminate and occurs over a short period following each molt. This results in a discrete pattern of size increase in which the animal maintains a constant body length during each instar and does not grow until it molts to the next instar. Most Daphnia have 3–5 juvenile instars (Green, 1956) followed by several adult instars, each of which concludes with the release of a clutch of neonates. Several species of Daphnia live in ponds (temporary and/or permanent) or small lakes where predation by Chaoborus is intense and fish are either absent or of minor importance. These include Daphniapulex, Daphniaminnehaha (Fig. 1), Daphniadentifera and Daphniaarenata, all of which are capable of developing neck spines in response to kairomones released by Chaoborus (Hebert, 1995; Beaton and Hebert, 1997).

Chaoborus larvae are tactile, ambush predators that remain motionless in the water column (buoyed up by two pairs of air sacs) and strike rapidly at approaching zooplankton prey. Mechanoreceptors that are associated with setae located along the length of the predator’s body are used to detect hydrodynamic disturbances made by the swimming and feeding movements of these prey (Pastorok, 1980). The prehensile antennae and mandibles of Chaoborus are then employed (following the strike) to capture the prey organism, which is ingested whole. The maximum size of prey that can be ingested is a function of the diameter of the predator’s mouth gape relative to prey body width (Swift, 1992). Of the four larval instars of Chaoborus, instars III and IV are sufficiently large to consume at least some Daphnia, although relatively large individuals of this prey may not be able to be ingested. Chaoborus are common in the plankton communities of both lakes and ponds and are especially abundant in bodies of water that lack fish. Species commonly found in these fishless habitats include the very large C. trivittatus and the medium to large-sized C. americanus (von Ende, 1975).

THE MODEL

General features

The fitness model is a stage-classified matrix population model designed for Daphnia (stages = instars), which follows the methods described by Caswell (Caswell, 2001) and is similar to (except as noted below) a recent model developed by Riessen (Riessen, 1999a). It is comprised of 15 instars in the Daphnia life cycle graph, with the fifth being the first adult instar. Table II describes the various parameters in the model, their values and sources of information. The population projection matrix includes three basic elements: F_i (fertility of instar i, which is a function of the clutch sizes of the various adult instars), P_i (probability that Daphnia survives and remains in instar i during a 1-day projection interval) and G_i (probability that Daphnia survives and grows to instar i + 1 during this interval). The transition probabilities, P_i and G_i, are functions of instar duration (T_i) and instar-specific survival probability (σ_i) (Table II). Survival probability, σ_i, is in turn the product of a non-predation (background) survival probability (b_i) and the probability of survival in the presence of Chaoborus predation (c_i). We calculate λ (dominant eigenvalue of the projection matrix) using a program designed with the software package Mathematica 4.0 (Wolfram, 1999; see Young, 2002 for program details) and report the relative fitness (range = 0–1) of each phenotype in a particular environment as λ λ_max⁻¹, where λ_max = highest value of λ in that environment. Each of the ten distinct environments examined in this study is defined by predator size (larger or smaller Chaoborus), predator density (0, 0.5 or 1.0 L⁻¹), and food level (high or low food). In these environments, Chaoborus size alters the vulnerability of different sizes of Daphnia to the predator, Chaoborus density alters encounter rate and thus predation intensity, and food availability influences both the growth rate (and thus body size) of Daphnia and its fecundity.

Daphnia body size and neck spine patterns

Daphnia body size has an effect on both vulnerability to Chaoborus predation and adult clutch size, while the presence of a neck spine influences both Chaoborus predation and instar duration. These effects produce trade-offs (summarized in Table I) that are incorporated in the model described below. Body size differences in Daphnia are examined in the model by creating 21 phenotypic patterns that vary in size at birth (body length of first instars ranges between 0.40 and 0.80 mm at 0.02-mm intervals) but then grow at the same rate (percent increase in body length from instar i to i+1) between given instars. This produces a series of growth trajectories in which differences in the size of the first instar are carried forward into all subsequent instars by applying the same growth rate to instars with different body lengths. In a given food environment, the same growth rate is applied to a particular instar in each of the 21 body size patterns, but this rate of growth is greater in the ‘high food’ environments than in those with a ‘low food’ level (Table III). This generates a range of patterns in Daphnia body length that encompasses the variation typically observed in lakes and ponds (Table IV).

The extent of the development of neck spines by Daphnia is represented in the model by 16 different patterns. These include one that has no instars with neck spines, one that has neck spines only in instar 1, one that has neck spines only in instars 1–2, one that has neck spines only in instars 1–3 and so on, concluding with a pattern where Daphnia exhibits neck spines in all 15 instars. The model, therefore, assumes that neck spines are always first developed in the first instar and then retained for variable durations in the life cycle. The combination of the 21 body size and 16 neck spine patterns produces 336 distinct Daphnia phenotypes in each of the ten environments, forming the adaptive landscape of the model.

Chaoborus predation

The effect of predation by Chaoborus on the relative fitness of the different Daphnia phenotypes is incorporated into the model by calculating the survival probability (over the 1-day projection interval) of each Daphnia instar in the presence of this predator (c_i) and integrating this into the overall instar-specific survival probability, σ_i (Table II). This calculation is accomplished by analyzing the predation cycle of the Chaoborus–Daphnia interaction, which can be broken down into three conditional probabilities: (i) the probability of encounter between the stationary predator and swimming prey (P_enc), which is calculated from the encounter rate, E_i, using the Poisson distribution, (ii) the probability that an encounter results in a strike (attack) by Chaoborus (P_strike) and (iii) the probability that a strike results in successful ingestion of the Daphnia (strike efficiency, S_i). Daphnia body length influences both encounter probability and strike efficiency, but with opposite effects (Table I). Larger Daphnia have faster swimming speeds, resulting in higher encounter rates with Chaoborus, but also are more difficult to handle and ingest by the predator, resulting in lower strike efficiencies (Swift and Fedorenko, 1975; Pastorok, 1981; Riessen et al., 1988; Swift, 1992; Riessen, 1999a). The net result of these counteracting effects of Daphnia body length is Chaoborus selection for intermediate-sized Daphnia (Pastorok, 1981; Riessen, 1992, 1999a). Neck spine development in Daphnia has a more straightforward effect—the presence of these structures decreases the strike efficiency of Chaoborus by making the prey more difficult to handle and ingest (Havel and Dodson, 1984).

We derive a function describing the relationship between encounter rate with Chaoborus and Daphnia body length (Table II; Fig. 2) from the data of Pastorok (Fig. 1 in Pastorok, 1981). We apply a power function regression to these data (Fig. 2), however, since Pastorok’s original linear regression model (Eq. 11 in Pastorok, 1981) results in encounter rates ≤0 for Daphnia with body lengths ≤0.58 mm (which includes many of the first instars and some of the second instars in the model). Encounter rate is also dependent on both Chaoborus density and spatial overlap between predator and prey in the water column (model assumes a uniform distribution of Daphnia and/or Chaoborus in the water). We ran simulations with Chaoborus densities of 0.5 L⁻¹ (moderate level that is frequently observed in lakes and ponds) and 1 L⁻¹ (high level that is nevertheless not uncommon) and also included in the model densities of 0 L⁻¹ in order to compare the shape of fitness landscapes produced in the absence of predation. In addition, turbulence within the water column may possibly influence encounter rates between Daphnia and Chaoborus; however, its exact effect on animals of these sizes and swimming abilities is unknown. In any event, any such effects will be of minimal importance in the fishless lakes and ponds that are the focus of this study, since these bodies of water are typically small and/or sheltered from the wind, characteristics which greatly decrease the creation of turbulence.

Following the results of Riessen (Riessen, 1999a), we use a constant value of 0.727 as the probability of a strike by Chaoborus given an encounter with Daphnia (P_strike). Unlike encounter probability and strike efficiency, this probability appears to be independent of Daphnia body length (Riessen, 1999a).

Strike efficiency by Chaoborus (S_i) is influenced by Chaoborus size, Daphnia body length and the presence of neck spines on the prey. We model predation effects by fourth-instar larvae of two species of Chaoborus that are common in fishless lakes and ponds, the very large (head capsule length ≈1.8 mm) C. trivittatus and the somewhat smaller (head capsule length ≈1.3 mm) C. americanus (Table II; Fig. 3). Neck spines have been shown to decrease the strike efficiency of Chaoborus on Daphnia to values between 25 and 50% of those on undefended individuals of the same body length (Krueger and Dodson, 1981; Havel and Dodson, 1984; Tollrian, 1995a). In this model, we decrease the Chaoborus strike efficiency on Daphnia instars with neck spines to 50% of that calculated for individuals of the same size without neck spines (Riessen, 1999a).

Clutch size

The number of parthenogenetic eggs carried by Daphnia in their brood chambers in each adult instar (ε_i) determines the values of F_i (fertility of instar i) in the matrix model (Table II). This number (clutch size) is a function of (i) the total amount of egg mass or volume produced in the instar and (ii) the size of the individual eggs. Total egg mass or volume in a clutch is dependent on the amount of nutrition collected during feeding, which is determined by both Daphnia body size and the quantity and quality of algae available in the lake or pond. The size of the individual eggs in a clutch is determined by the size of the neonates they need to produce at the end of the instar; larger-sized eggs are required to produce the larger neonates associated with larger Daphnia phenotypes (Fig. 4). The result of these relationships is a classic trade-off between egg size and egg number (Stearns, 1992).

We use relationships (for high and low food concentrations) linking adult body length to clutch size in D. pulex (neonates = 0.66-mm body length) as standards to determine the clutch sizes of the various instars of the different Daphnia phenotypes in the model (Fig. 5). By multiplying clutch size in these standard functions by 6.894 × 10⁶ μm³ (egg volume for a 0.66-mm neonate; Fig. 4), we obtain relationships of total clutch volume to adult Daphnia body length that we use for all phenotypes in the model. To finally obtain clutch sizes for the instars in each Daphnia phenotype at high and low food levels, we then divide the value of total clutch volume for a given adult body length by the egg volume related to the neonate size of that phenotype (Fig. 4). In this way, we are able to calculate clutch size for an adult instar taking into account the size of the adult Daphnia, food concentration and individual egg size.

Instar duration

The duration of a Daphnia instar (T_i) is a function of water temperature, stage of development (juvenile or adult) and presence or absence of a neck spine. Instar durations used in this model for Daphnia without neck spines (listed in Table II) are those described by Riessen (Riessen, 1999a; Table I) for D. pulex raised at 20°C. These values are similar to those measured by Lynch (Lynch, 1989). Daphnia instars that develop neck spines in response to the presence of Chaoborus kairomones typically exhibit an extended duration that results in delayed reproduction (Riessen, 1999b). In this model, the duration of an instar with a neck spine is 15% longer than for the same instar without this defensive structure, a delay in development which is typical for D. pulex (Riessen and Sprules, 1990).

RESULTS

In the model environments that contain the smaller Chaoborus predator, C. americanus, the largest Daphnia phenotypes (neonate body length = 0.80 mm) have the highest relative fitness (Fig. 6). This result holds for Daphnia that exhibit all degrees of neck spine development, for both high and low food conditions and for Chaoborus densities at 0.5 and 1.0 L⁻¹. While there is a general increase in fitness with increasing body size, this pattern is not monotonic. Relatively small Daphnia (but not the smallest) have the lowest fitness; minimum values occur in Daphnia phenotypes with neonate body lengths of 0.42–0.44 mm at high food and 0.50–0.56 mm at low food. Increasing body length beyond these points produces rapid increases in relative fitness. In high food environments, this increase tends to level off somewhat for the largest phenotypes (neonate body lengths between 0.70 and 0.80 mm), but under low food concentration fitness continues to steadily increase with body length.

The extent of neck spine development in Daphnia that yields the highest fitness under predation by C. americanus depends upon Daphnia body size, food conditions and, to a lesser extent, Chaoborus density (Fig. 6). Smaller Daphnia phenotypes have a fitness advantage when neck spines are retained through most or all of the adult instars. This is especially true under low food concentration, where growth rates are lower (Table III) and Daphnia therefore smaller at a given instar (Table IV), and at high Chaoborus densities (1.0 L⁻¹). Under either of these circumstances, maximal fitness for small Daphnia phenotypes (neonate body length ≤ 0.50 mm) is always achieved when neck spines are maintained throughout the individual’s life. Large Daphnia phenotypes (neonate body length ≥ 0.70 mm) living in environments with a high food concentration have highest fitness values when neck spines are retained only through juvenile development (instar 4). In environments with a low food concentration, however, there is a fitness advantage for these larger Daphnia to retain their neck spines after maturity. This extension of neck spine development is influenced by the intensity of predation; high Chaoborus densities (1.0 L⁻¹) select for longer retention of these defensive structures. For all Daphnia in the various C. americanus environments, there is strong selection pressure (indicated by the steep slope of the fitness surface) to retain neck spines through at least the four juvenile instars. The fitness curves for phenotypes associated with each body size pattern, however, tend to level off, indicating only weak selection pressure on the larger adult instars to either retain or eliminate their neck spines.

Overall, the fitness landscape for Daphnia subject to predation by C. americanus (Fig. 6) is one where natural selection favors very large phenotypes that develop neck spines which are retained at least through juvenile development. High food concentrations (and therefore rapid growth) favor the loss of neck spines in these large Daphnia at maturity (instar 5), but in low food environments, selection favors Daphnia that retain these defensive structures somewhat longer, at least through the first few adult instars. Lowest fitness values occur in phenotypes that are relatively small and lack any neck spine development.

Predation on Daphnia by the larger Chaoborus species, C. trivittatus, under high food concentrations produces fitness landscapes that are very similar to those generated in the C. americanus predation environments with low food (Fig. 7). Highest fitness occurs for the largest Daphnia phenotypes and for those that retain neck spines through most or all of the adult instars. Lowest fitness is associated with relatively small Daphnia (minimum values occur for phenotypes with neonate body lengths between 0.48 and 0.54 mm) and those with no neck spines. For medium and large-sized phenotypes there is strong selection pressure for larger body size, and for phenotypes of all sizes there is strong selection pressure to develop neck spines and retain them through juvenile development and into at least the early adult instars.

The Daphnia fitness landscapes that are associated with C. trivittatus predation and low food conditions, on the other hand, are fundamentally different from those that represent the other simulated environments (Fig. 7). The primary difference is in the effect of body size on fitness, which results in changes in selection pressures on this trait. Fitness values here are very high for both extremely large and extremely small Daphnia phenotypes and relatively low for those at intermediate sizes. Highest values always occur in either the largest (neonate body length = 0.80 mm) or smallest (neonate body length = 0.40 mm) phenotypes, while lowest values are associated with those phenotypes that have neonate body lengths between 0.62 and 0.68 mm. In these fitness landscapes, larger Daphnia phenotypes would experience strong selection pressure to further increase body size, while smaller phenotypes would be under strong selection pressure for a further decrease in size. This will ultimately drive Daphnia to either very large or very small body sizes. Interestingly, increasing Chaoborus predation pressure (as a result of increased predator densities) shifts the fitness landscape in a way that tends to favor the very small Daphnia phenotypes. At C. trivittatus densities of 0.5 L⁻¹ the largest sized Daphnia tend to have fitness values that are slightly higher than the smallest phenotypes; however, this pattern is reversed when predator densities are increased to 1.0 L⁻¹. The overall pattern of selection, which favors both small and large (but not intermediate-sized) Daphnia, is very different than in the other model environments (C. americanus predation—high food and low food; C. trivittatus predation—high food), where high fitness values are limited to only large Daphnia phenotypes and natural selection favors only increases in body size.

The C. trivittatus—low food environments also favor a long retention of neck spines in Daphnia (Fig. 7). Highest fitness values nearly always occur in those Daphnia phenotypes in which neck spines are present in all 15 instars. Very strong selection pressure exists to retain neck spines through both the juvenile and earlier adult instars, and an especially large increase in fitness results from extending neck spine duration from instar 4 (end of juvenile development) to instar 5 (first adult instar). This selection pressure is considerably decreased in the later adult instars, although continued retention of neck spines nearly always increases fitness. Overall, the model environments with C. trivittatus predation and low food conditions select for either very large or very small Daphnia phenotypes that retain neck spines well beyond maturity.

Selection pressures in the absence of predation (Chaoborus density = 0 L⁻¹) are weaker than in any of the Chaoborus environments, as indicated by the generally more gradual slopes on the fitness landscapes (Fig. 8). There is, nevertheless, under both high and low food concentrations, a fitness advantage for those Daphnia phenotypes that are larger and completely lack neck spines. Since there is a cost for developing neck spines (delayed reproduction), but no benefit in the absence of Chaoborus, highest fitness values in the absence of Chaoborus predation always occur in Daphnia phenotypes that never form these defensive structures, and fitness decreases progressively with their development and increased retention. On the other hand, the effect of body size on fitness is not quite as straightforward. There is a general increase in fitness with increasing body size, but the smallest Daphnia do not have the lowest fitness (minimum values occur for phenotypes with neonate body lengths between 0.44 and 0.46 mm) and under high food concentrations fitness values peak prior to the maximum body size in the model (highest values occur for phenotypes with neonate body length = 0.76 mm). Overall, the model predicts that freshwater planktonic environments that lack predators would select for large Daphnia that do not develop neck spines.

DISCUSSION

Analysis of fitness landscapes

The precise shapes of the fitness landscapes produced by the model (Figs 6–8) result from the interplay between the different Daphnia body size patterns that are present (each of which produces a unique trajectory of instar-specific body sizes) and the relative sizes of the two Chaoborus species, along with the interaction that occurs between the defensive effects of neck spines and prey body size. Daphnia body size patterns are determined by both body length at birth (which varies from 0.40 to 0.80 mm among the different model phenotypes) and growth rate (which is controlled by the food environment) (Table IV). The effect of Chaoborus predation on Daphnia fitness depends on these body size patterns, but also on the size of the predator. Very large Chaoborus, such as C. trivittatus, have very large mouth gapes and are, therefore, capable of ingesting much larger Daphnia than are somewhat smaller species, such as C. americanus (Fig. 3). A Daphnia may be able to grow large enough at some point in its life to become completely invulnerable to Chaoborus predation. The earlier this occurs, the greater the fitness advantage for that animal. Strong selection pressure will exist in favor of those phenotypes that are able to achieve such a size refuge, especially those that do so relatively early in life. Since larger Daphnia phenotypes are most likely to accomplish this, natural selection will generally favor the evolution of larger body size in response to Chaoborus predation. This tendency, however, will be modified by the size of Chaoborus that is present and the nature of the food environment. A size refuge from this predator earlier in life is more likely to occur for Daphnia that are subject to predation by smaller Chaoborus (smaller mouth gape and thus Daphnia can become invulnerable at a relatively small size) and for those that live in high food environments (faster growth rates). In the C. americanus environments in the model, natural selection favors very large Daphnia phenotypes, which can produce individuals that are invulnerable to this predator (due to large size) by instar 5 in high food conditions and by instar 7 in low food situations. Similar selection pressures exist for Daphnia under C. trivittatus predation in a high food environment since a fast growth rate allows very large Daphnia phenotypes a complete size refuge from this large predator by instar 6. However, even the largest Daphnia phenotype in a low food environment (with a subsequent slow growth rate) does not reach a size invulnerable to C. trivittatus predation until instar 11, and some relatively large Daphnia phenotypes never grow large enough to be invulnerable to this predator. Environments with C. trivittatus and a low food concentration, therefore, subject many of the adult instars of very large Daphnia phenotypes to intense Chaoborus predation. This negates much of the benefit of large body size and allows the predation advantages of smaller phenotypes (decreased encounter rate with this ambush predator) to more or less balance those of larger phenotypes (decreased strike efficiency). The ultimate effect on Daphnia under these conditions is selection for either very large or very small phenotypes.

Daphnia instars that are large enough to be invulnerable to Chaoborus predation no longer receive any benefit from retaining a neck spine. As a consequence, natural selection will favor the extended retention of neck spines (well beyond maturity) for smaller Daphnia phenotypes (size refuge from Chaoborus predation occurs later, if at all), for Daphnia that experience predation by large C. trivittatus (relatively difficult to outgrow vulnerability and neck spines therefore remain beneficial), and for Daphnia living in a low food environment (slower growth produces smaller individuals). In contrast, there is relatively strong selection pressure on larger Daphnia phenotypes in the C. americanus—high food environments to eliminate neck spines after instar 4 (final juvenile instar) since the Daphnia are then too large to be consumed.

In the absence of predation, natural selection will favor large Daphnia phenotypes that lack any neck spine development (Fig. 8). In all Chaoborus environments, however, the cost of neck spines (delayed reproduction) is strongly counteracted by their effectiveness in inhibiting the ability of Chaoborus to handle and ingest Daphnia. This produces fitness landscapes that exhibit strong selection pressures for the development of these defensive structures in Daphnia and their retention at least through juvenile development (Figs 6 and 7). On the other hand, the tendency in the absence of predators for natural selection to favor large Daphnia (Fig. 8) is strongly reinforced in most Chaoborus environments—ones that contain C. americanus (Fig. 6) and those that combine C. trivittatus predation with high food concentrations (Fig. 7). However, in low food environments (slow growth) with C. trivittatus (capable of ingesting relatively large prey) (Fig. 7), the advantage of small size for Daphnia (decreased encounter rate with the predator: Fig. 2) can counteract (and even exceed) the combined advantages of large body size, which include decreased strike efficiency (Fig. 3) and increased reproduction (as evidenced in the fitness landscapes in the absence of predation: Fig. 8).

Daphnia in fishless habitats

Among the Daphnia species that coexist with Chaoborus in fishless lakes and ponds, two in particular, D. pulex and D. minnehaha, present an interesting contrast in body size and neck spine development. Daphnia pulex is a very large species that is common in both temporary and permanent ponds of temperate (especially cool temperate) regions, including most of temperate North America (Hebert, 1995). Neonates range in body length from 0.55 to 0.80 mm, size at first reproduction under high food conditions varies from 1.6 to 2.2 mm and maximum body length can reach 3.5 mm (Dodson and Havel, 1988; Riessen and Sprules, 1990; Spitze, 1992; Hebert, 1995; H. P. Riessen, unpublished results). In response to kairomones released by Chaoborus larvae, D. pulex develops neck spines that are maintained only during juvenile development, typically through either instar 3 or instar 4 (Havel, 1985; H. P. Riessen, unpublished results). Daphnia minnehaha, on the other hand, is a much smaller species than D. pulex—neonate body length is <0.60 mm and can be as low as 0.40 mm (values are usually between 0.45 and 0.55 mm), while size at first reproduction typically varies from 1.0 to 1.4 mm with values occasionally as low as 0.8 mm (H. P. Riessen, unpublished results). Daphniaminnehaha also develops neck spines in response to Chaoborus kairomones (Fig. 1), but unlike D. pulex, retains them at least through its early adult instars (H. P. Riessen, unpublished results). This species, in fact, appears to be unique in having adult females that possess neck spines (Hebert, 1995). Daphniaminnehaha (a species closely related to D. catawba) is commonly found in ponds and small lakes (especially those with an acidic pH) from northern New England and the maritime provinces of Canada westward through much of the Great Lakes watershed, a range that is entirely encompassed within that of the more widely distributed D. pulex (Hebert, 1995).

The phenotypes for these two distinct Daphnia species match alternative antipredator strategies that are predicted by our model for different fishless environments. Daphniapulex fits the phenotypic profile (large body size and neck spine development that is limited to juvenile instars) expected to be optimal in high food environments where predation is dominated by moderately large Chaoborus (C. americanus). This species is, in fact, very common (and often the dominant zooplankter) in shallow fishless ponds in temperate North America that are inhabited by C. americanus and typically have relatively high algal productivity. In contrast, the phenotype of D. minnehaha (small body size with neck spines retained into adult instars) appears designed to take advantage of low food (oligotrophic) environments with high densities of extremely large Chaoborus (C. trivittatus). This prediction of small size to counteract large invertebrate planktivores is counterintuitive since there is a general expectation that plankton communities dominated by invertebrate predators will be composed of mainly large-sized zooplankton that are difficult to consume (Dodson, 1974). This principle, however, may not be as broadly applicable as believed since several very large invertebrate predators (e.g. large Chaoborus, notonectids, mysids and Bythotrephes) are known to be capable of ingesting large Daphnia and other large zooplankton prey, making large size an ineffective defense in some cases (Swift and Fedorenko, 1975; Murtaugh, 1981; Cooper, 1983; Scott and Murdoch, 1983; Schulz and Yurista, 1998; Young, 2002). This could lead to circumstances that favor small-sized zooplankton communities in fishless lakes and ponds. A recent study by Malkin (Malkin, 2003) on fishless lakes in Algonquin Provincial Park, Ontario, Canada, in fact, found that these systems were mostly dominated not by the expected large body-sized species, but by relatively small zooplankton. Unfortunately, fishless lakes and ponds in which C. trivittatus larvae are the main predators have not been well studied, and thus the influence of these especially large Chaoborus on Daphnia and other zooplankton is not clearly understood.

While there is not as yet sufficient evidence to constitute a proper test of our model, various studies on responses of D. pulex to C. americanus predation do provide some experimental support for its predictions. In a selection experiment, Spitze (Spitze, 1991) subjected populations of D. pulex (cultured under high food conditions) to predation by fourth-instar C. americanus (0.25 L⁻¹) over a 10-week period and determined the evolutionary responses of several life history traits. He observed a pronounced increase in Daphnia body size in all instars as a direct evolutionary consequence of Chaoborus predation. Other studies have investigated phenotypic plasticity in D. pulex, specifically the induced responses of individuals to kairomones released by Chaoborus. In addition to the development of neck spines in juvenile instars, D. pulex often exhibits an increase in body size and growth that results in a larger size at maturity (Tollrian, 1995b; Riessen, 1999b). Both this response and the evolutionary changes observed by Spitze (Spitze, 1991) are consistent with the predictions of our fitness model with respect to adaptive responses of D. pulex to predation by C. americanus.

Certainly factors other than Chaoborus predation are important in shaping Daphnia body size, life history and distribution in fishless lakes and ponds. Competitive interactions with other herbivorous zooplankton, which are greatly influenced by body size, may result in severe food limitation and thus greatly limit growth, reproduction and survivorship (Gliwicz, 1990). Other features of lakes and ponds may have similar effects on Daphnia. These would include poor food quality, resulting from nutrient limitation in phytoplankton (Sterner et al., 1993; Gulati and DeMott, 1997), low pH and low calcium concentration (Hessen et al., 2000). Each of these must surely be taken into account in any complete analysis of Daphnia life history evolution and distribution. Nevertheless, high densities of Chaoborus larvae are common in fishless lakes and ponds and constitute, through their size-selective predation effects, a major ecological and evolutionary force in these systems. The model developed in this article allows us to better understand this force and thus to better comprehend important features of Daphnia ecology and evolution.

ACKNOWLEDGEMENTS

We are grateful to the Ontario Ministry of Natural Resources and the staff at the Harkness Fisheries Laboratory (Mark Ridgway, Trevor Middel and Gary Rideout in particular) for the logistical support they provided for our sampling of lakes in Algonquin Provincial Park. We are also indebted to several individuals who assisted us in sampling these lakes and their Daphnia populations—Sairah Malkin, Wiebke Böing, Björn Wissel and Gary Pettibone. Finally, we thank Charles Ramcharan for assistance in developing parts of the model and Randy Snyder for his comments on the manuscript and for technical help in the preparation of the figures.

REFERENCES

Author notes

¹Department of Biology, Suny College at Buffalo, 1300 Elmwood Avenue, Buffalo, NY 14222, USA and ²Department of Biology, York University, 4700 Keele Street, Toronto, Ontario, Canada M3J 1P3