Marine Biology

, Volume 146, Issue 4, pp 771–780

Palatability of autotrophic dinoflagellates to newly hatched larval crabs

Authors

  • M. F. Perez
    • Shannon Point Marine Center
    • Shannon Point Marine Center
Research Article

DOI: 10.1007/s00227-004-1482-8

Cite this article as:
Perez, M.F. & Sulkin, S.D. Marine Biology (2005) 146: 771. doi:10.1007/s00227-004-1482-8

Abstract

To determine the general palatability of autotrophic dinoflagellates to newly hatched crab larvae and whether there are taxonomic, predator/prey size relationships, or toxicity components to their ability to discriminate among dinoflagellates, larvae of six species of crabs from two families were fed 16 species/strains of dinoflagellates from three orders. Dinoflagellate cell length ranged from 18 to 50 µm, and toxic and non-toxic species/strains were included. Experiments measuring incidence of prey ingestion, grazing rates on individual constituents of selected prey combinations, and development on one toxic species shown to be readily ingested were conducted between 2000 and 2002. Thirteen of sixteen dinoflagellates were palatable to larvae, with no consistent pattern of prey discrimination based on taxonomic affinity, toxicity, larval hatching season, or predator/prey size relationships. Although the three dinoflagellates not ingested were toxic, three other toxic species/strains were ingested, with accelerated mortality occurring in the one case. Ingestion of non-favored prey occurred only at very low rates when mixed with readily ingested prey, indicating selectivity. Larvae hatching in winter generally ingested dinoflagellates as readily as did zoeae hatched in spring and summer. Newly hatched larvae ingested a wide variety of dinoflagellates, while discriminating among related species. Such discrimination will not always prevent larval ingestion of prey that will result in mortality.

Introduction

Brachyuran crabs have a complex life cycle that includes a free-living planktonic larva, the zoea. Zoeae of most crab species are planktotrophic, requiring nutrition obtained from feeding on particulate carbon in the water column (e.g. Costlow and Bookhout 1959; Sulkin 1978). Obtaining nutrition from the plankton may be particularly important for newly hatched larvae, since laboratory studies have shown that residual nutrition in the egg supports unfed larvae for only a few days. If larvae do not ingest prey soon after hatching, development will be delayed and many will not survive to the second zoeal stage, even if prey is subsequently provided (Paul and Paul 1980; Anger and Dawirs 1981; Staton and Sulkin 1991; Hartman 1994). Furthermore, larvae of most crab species hatch on the bottom and possess behavioral traits that promote active swimming up into the water column (Sulkin 1984), an activity that consumes energy. Brachyuran zoeae appear to satisfy their nutritional demands by being flexible in the types of prey they will ingest, including large micro- and meso-zooplankton (e.g. Sulkin 1975; Bigford 1978; McConaugha 1985; Harms and Seeger 1989), algae (e.g. Sulkin 1975; Hartman and Letterman 1978), heterotrophic dinoflagellates, and detrital particles (Lehto et al. 1998). Indeed, Levine and Sulkin (1984) and Factor and Dexter (1993) demonstrated that zoeae will ingest artificial microparticles.

Thus individual brachyuran zoeae appear to ingest a wide variety of prey types in a wide range of prey sizes (with lengths of 10–400 µm). Somewhat paradoxically, however, larvae also discriminate among closely related prey types (Hinz et al. 2001). Autotrophic dinoflagellates demonstrate this phenomenon well. Ingestion by brachyuran zoeae has been reported on the dinoflagellates Prorocentrum reticulatum (Incze and Paul 1983), Alexandrium tamarense (Yazdandoust 1987), and Prorocentrum micans (Lehto et al. 1998; Sulkin et al. 1998b). However, Hinz et al. (2001) reported that the same larvae that ingested P. micans rejected three strains/species of Alexandrium spp., two of which were toxic and one non-toxic. Moreover, if larvae were exposed to mixtures of the favored P. micans and one of the three Alexandrium spp. not typically ingested, some of the latter cells appeared in the larval gut, indicating at least a brief period of ingestion.

While there is evidence that larvae discriminate among autotrophic dinoflagellates, the numbers of both predator and prey species tested to date are too limited to determine whether dinoflagellates as a group are generally palatable to brachyuran zoeae and to assess whether feeding cues producing discrimination among cell types are related to prey taxonomic affinity, cell toxicity, or possible relationships between prey cell size and larval size. To address these issues, we tested 16 species/strains of autotrophic dinoflagellates belonging to three orders (Gonyaulacales, Gymnodinales, and Prorocentrales), ranging in cell length from 18 to 50 µm, and including toxic and non-toxic strains. Dinoflagellates were fed to newly hatched larvae of six species of crabs belonging to two families (Grapsidae and Cancridae), whose larvae were of different sizes and included winter-, spring-, and summer-hatching species.

Materials and methods

Experimental approach

The percent of larvae ingesting specific dinoflagellates was determined by examining gut contents of larvae exposed for 24 h to the prey. Ingestion rates were used to determine if larvae exposed to combinations of favored and non-favored prey (based on incidence of ingestion) could distinguish between the prey types and feed selectively. Finally, larvae were raised on one readily ingested toxic dinoflagellate to determine the effects of such feeding on development.

Dinoflagellate species/strains were selected based on taxonomic affinity, toxicity, and predator/prey size relationships to test specific hypotheses based on previous studies. For example, Prorocentrum micans Ehrenberg was readily ingested by zoeae of three species of crabs (Lehto et al. 1998; Sulkin et al. 1998b). To determine whether species/strains belonging to this genus were generally palatable, we tested three different strains of P. micans and three other Prorocentrum spp., two of which were toxic, with cell lengths ranging from 18 to 45 µm (Table 1). We tested these prey on zoeae of the same three crab species previously used (Cancer magister Dana, C. oregonensis Dana, and Hemigrapsus oregonensis Dana) and three additional species (C. productus Randall, C. gracilis Dana, and H. nudus Dana). Zoeae of the former three crab species were shown by Hinz et al. (2001) to avoid ingesting three species/strains of Alexandrium (fundyense 1719 Balech, tamarense 118 LeBour, and tamarense 115), the first two being toxic. To determine whether Alexandrium spp. were generally unpalatable, we repeated these experiments with the same dinoflagellate species/strains on zoeae of all six crab species and added two additional Alexandrium species, one toxic (A. andersoni Balech) and the other non-toxic (A. catenella Balech). To expand the scope of taxonomic groups, toxicities, and prey sizes tested, two additional non-toxic species of the order Gonyaulacales (Gonyaulax spinifera Diesing and Protoceratium reticulatum Butschii) and two species of the order Gymnodinales (Gymnodinium catenatum Graham, one toxic and one non-toxic strain, and the non-toxic Gyrodinium instriatum Freudenthal et al.) were included.
Table 1

Characteristics of autotrophic dinoflagellates used in experiments (DSP diarrhetic shellfish poisoning; #, previously tested on crab larvae; CCMP Provasoli-Guilliard National Center for the Culture of Marine Phytoplankton; SPMC Shannon Point Marine Center, laboratory of S. Strom). Species/strain abbreviations in parentheses

Species and strain

Toxin

Length (µm)

Width (µm)

Order Gonyaulacales

  Gonyaulax spinifera CCMP 409 (Gs)

No

50

50

  Protoceratium reticulatum CCMP 1889 (Pr)

No

38

34

Order Gymnodinales

  Alexandrium andersoni CCMP 1718 (Aa)

Saxitoxins

27

23

  Alexandrium catenella CCMP 1911 (Ac)

No

34

33

  Alexandrium fundyense CCMP 1719# (Af)

Saxitoxins

35

35

  Alexandrium tamarense CCMP 115# (A5)

No

35

35

  Alexandrium tamarense CCMP 118# (A8)

Saxitoxins

40

35

  Gymnodinium catenatum CCMP 1937 (G37)

Saxitoxins

38

33

  Gymnodinium catenatum CCMP 1940 (G40)

No

41

36

  Gyrodinium instriatum CCMP 431 (Gi)

No

50

38

Order Prorocentrales

  Prorocentrum dentatum CCMP 1517 (Pd)

No

19

11

  Prorocentrum hoffmanianum CCMP 683 (Ph)

Okadaic acid

45

35

  Prorocentrum micans CCMP 689 (P9)

No

38

17

  Prorocentrum micans CCMP 693 (P3)

No

40

25

  Prorocentrum micans SPMC# (Pm)

No

35

15

  Prorocentrum minimum CCMP 699 (Pn)

Venerupin, DSP

18

16

Experimental organisms

Adults of the six species of crabs used in the experiments are distributed along a habitat gradient from upper intertidal rocky beach to sub-tidal eelgrass beds in the vicinity of Anacortes, Washington, USA. One or another of the six crab species are ovigerous from February through August. This group of species thus provided zoeae that hatched in winter, spring and summer, permitting analysis of whether ingestion of dinoflagellates is related to the likelihood of encountering them in the water column. These species also provided a range of sizes of newly hatched larvae as follows (tip of dorsal spine to tip of rostral spine): Hemigrapsus oregonensis (Ho), 1.10 mm (Hart 1935); Cancer gracilis (Cg), 1.12 mm (Ally 1975); H. nudus (Hn), 1.20 mm (Hart 1935); C. oregonensis (Co), 1.81 mm (H. Ko, personal communication); C. productus (Cp), 2.06 mm (Trask 1970); and C. magister (Cm), 2.12 (H. Ko, personal communication). Prior research on larval nutrition has been conducted on all species except H. nudus (Lehto et al. 1998; Sulkin et al. 1998a, 1998b; Sulkin and McKeen 1999).

Ovigerous C. magister and C. productus were collected by SCUBA in late winter 2000 and 2001 from sub-tidal eel-grass beds in Ship Harbor, Anacortes, Washington, USA. C. oregonensis was collected in late winter by rock dredge from Burrows Bay near Anacortes. Ovigerous H. nudus, H. oregonensis, and C. gracilis were collected by hand in spring and summer of 2002 (H. oregonensis was also collected in spring and summer 2001) from the intertidal at Shannon Point Marine Center. Ovigerous C. magister, C. productus, and C. gracilis were held in the laboratory, initially in running seawater tables at ambient temperatures (8–11°C as the year progressed). Smaller (<4 cm carapace width) H. nudus, H. oregonensis, and C. oregonensis were held in 80-mm-diameter glass bowls containing 0.2-μm-filtered seawater (FSW) in light- and temperature-controlled incubators (14 h light:10 h dark cycle; 12°C). When an oviger showed signs of hatching (darkening and loosening of egg sacks), the female was isolated in a separate container. Several hundred newly hatched larvae were collected with a 200-μm-mesh sieve and placed in 170-mm-diameter glass bowls or 500-ml beakers filled with FSW and used in experiments within 24 h. When possible, larvae from at least two broods that hatched on the same day were pooled and distributed haphazardly among treatments in each experiment.

Dinoflagellates

Dinoflagellate cultures were obtained from the Provosoli-Guillard National Center for Culture of Marine Phytoplankton (CCMP) (Boothbay Harbor, Maine, USA) and the laboratory of S. Strom at the Shannon Point Marine Center (SPMC). Characteristics of the 16 species/strains used in the ingestion experiments are shown in Table 1. Toxicity status of each strain is based on that published on the CCMP website at the time of purchase (http://ccmp.bigelow.org). Dinoflagellates used in the experiments ranged in length from approximately 18 µm (Prorocentrum minimum Schiller) to 50 µm (Gyrodinium instriatum, Gonyaulax spinifera).

All dinoflagellate strains were maintained in exponential growth phase (ranging from 103 to 104 cells ml−1 depending upon species) in 14 h light:10 h dark cycles under cool white fluorescent bulbs at approximately 75 µmol photons m−2 s−1. Prorocentrum hoffmanianum Faust and two strains of Gymnodinium catenatum were maintained at 22°C; all others were maintained at 12°C. Cultures were maintained in their specific media (F/2, F/10, or L1) as specified on the CCMP website (http://ccmp.bigelow.org/Cl/Cl_01.html). Stock cultures of each strain were maintained in 10-ml glass culture tubes, from which larger experimental cultures were initiated periodically. Experimental cultures were grown and maintained in 250-ml or 1-l polycarbonate Nalgene bottles.

Experimental cell densities of 103 cells ml−1 were used for most experiments. Using cells from stock cultures in exponential growth phase reduced one source of variation due to differences in surface properties and toxicities of cells, both of which are known to change with growth phase (Taroncher-Oldenburg et al. 1997; Aguilera and Gonzales-Gil 2001). A common experimental cell density of 103 cells ml−1 was used, because it falls within the range of natural dinoflagellate blooms (Smayda 1997; Vila et al. 2001) and has been used previously in tests with larval crabs (Hinz et al. 2001).

Incidence of ingestion

Incidence of ingestion of dinoflagellates was determined by the presence of chl a in the larval gut after 24 h of exposure to the prey. Larvae were examined for the presence of chl a in the gut under epifluoresence at 354 nm, using a Leica DMR microscope. Although epifluorescence color varied among prey species, the typical aspect was red to reddish-orange. Gut coloration often indicated a gut packed with cells, as for C. oregonensis zoeae feeding on P. micans (Fig. 1). However, there were frequently only a few cells or faint reddish-orange coloration present. Ambiguous results were counted as negative; the results are therefore conservative with respect to incidence of ingestion.
Fig. 1

Cancer oregonensis. Zoeae shown after feeding on Prorocentrum micans: top panel under epifluorescence at 354 nm; bottom panel light microscopy

For tests with each prey species, an unfed control and a P. micans treatment were included (for prey other than P. micans). The unfed treatment confirmed that ingestion of dinoflagellates was required to produce gut coloration under epifluoresence. The P. micans treatment confirmed that larvae being tested would ingest any dinoflagellate prey, based on previous reports that this prey was readily ingested by zoeae of at least three species of crabs (Lehto et al. 1998; Hinz et al. 2001). In all experiments, no unfed larvae showed evidence of fluorescence in the gut, and, in all cases, P. micans was ingested by the larvae being tested.

Experiments were set up as follows for all target prey. Several hundred newly hatched larvae from two or more broods were placed in a glass bowl containing FSW. From the pooled assemblage of larvae, groups of 50 larvae were placed in each of three 80-mm-diameter glass bowls each containing 80 ml of the target prey that had been diluted with FSW to the test density of 103 cells ml−1. Larvae were exposed to the target prey for 24 h at 12°C under dimmed light (approximately 17 µmol photons m−2 s−1) on a 14 h light:10 h dark cycle. Although it was obvious that some larvae had ingested some prey soon after initial exposure, measurements were made for all tests after 24 h (Hinz et al. 2001), to facilitate comparisons among prey treatments and crab species. Fifteen larvae were then haphazardly selected from each of the three bowls to determine gut fluorescence. The percent of larvae showing such fluorescence from each bowl was determined. The entire experiment was typically run twice with two different batches of larvae, providing six independent measurements for each diet treatment. Each of the other two treatments (unfed and P. micans) were run in an identical manner. Only results for the target species are presented.

In addition to the percent incidence of ingestion, a number of ingestion events observed while examining the larvae under epifluorescence microscopy are reported.

Mixed prey grazing

Ingestion rates for each member of specific two-prey combinations were determined. The prey mixtures consisted of equal densities (125 cells ml−1) of P. micans, a prey shown to be readily ingested, and either Alexandrium fundyense (for Cancer magister, C. oregonensis, C. productus, Hemigrapsus oregonensis) or Protoceratium reticulatum (for H. oregonensis). The latter two prey were ingested only rarely by larvae when presented alone.

Rate of ingestion was determined by measuring the number of prey cells removed from a known volume of culture water by newly hatched larvae in a 24-h period. For each ingestion-rate experiment, the density of the combined prey culture was adjusted by diluting the stock culture of each constituent with F/10 medium to achieve a nominal 125 cells ml−1 and by combining them at the beginning of the experiment. To determine the exact initial cell concentration, three 1- to 2-ml samples were collected from the combined treatment, preserved with Lugol’s solution, and counted by light microscopy using a Palmer–Maloney chamber. A total of 100 ml of the combined prey treatment was added to each of six 250-ml bottles. Fifty larvae were then added to each of three of the bottles; the remaining three were used as predator-free controls. Bottles were maintained under low light conditions (approximately 17 µmol photons m−2 s−1) for 24 h at 12°C. At the conclusion of the experiment, a 1- to 2-ml sample of culture water was collected, preserved, and counted from each of the six bottles.

The equations of Frost (1972) were used to calculate ingestion rates for each dinoflagellate in the mixture (number of cells ingested per zoea per day), and these were compared with one-way ANOVA or Kruskal–Wallis ANOVA as appropriate. Post hoc power analyses were conducted on non-significant results using G-power.

Development of zoeae fed A. andersoni

Newly hatched larvae of C. magister, H. oregonensis and H. nudus were raised in the laboratory from hatching to determine the effects on development of feeding on the toxic strain of A. andersoni. For larvae of each species, treatments included unfed, non-toxic P. micans, toxic A. andersoni, and freshly hatched nauplii of the brine shrimp Artemia sp. (Argentenium, Argent Chemical Laboratory, Redmond, Washington, USA). Artemia sp. nauplii have been shown to support high survival in the laboratory for first stage larvae of both C. magister and H. oregonensis (Poole 1966; Sulkin et al. 1998a). Prey densities were 8 nauplii ml−1 for Artemia sp. and 103 cells ml−1 for each dinoflagellate.

For each experiment, larvae from two broods that hatched on the same day were pooled. Individuals were removed haphazardly from the larval culture and placed individually in wells of a 12-well plastic tray. Each well contained 3 ml of FSW containing the appropriate dietary treatment; 24–36 larvae were used for each experiment. All larvae were examined daily for evidence of mortality and molting, and transferred by pipette to fresh culture medium (with appropriate prey). Each experiment was concluded when all larvae had either died or molted to zoeal stage 2.

Results

Genus Prorocentrum (order Prorocentrales)

Virtually all larvae examined from each of the six crab species ingested each of the three Prorocentrum micans strains tested (Fig. 2). Larval midguts for all six species typically were filled with cells as shown for Cancer oregonensis (Fig. 1). No epifluorescence was noted in the hindgut. When Hemigrapsus oregonensis larvae were fed P. micans, prey cell thecae present in fecal matter were empty, with no fluorescence detected, indicating the complete breakdown of chl a, while cells were in the midgut.
Fig. 2

Prorocentrum spp. Mean percent ingestion (n=6; ±SE) by larvae of indicated crab species (abbreviations as in “Materials and methods—Experimental organisms”) on the indicated dinoflagellates (asterisk indicates toxic strain; NT indicates not tested; no bar indicates 0% response; cross-hatched bars indicate significant differences between feeding on P. dentatum and P. minimum for Cancer oregonensis (Co) and Hemigrapsus oregonensis (Ho) as described in “Results”)

Larvae of the five crab species fed the smaller, non-toxic P. dentatum ingested the cells, with only larvae of C. gracilis showing <50% incidence of ingestion (Fig. 2). Thus, while this smaller Prorocentrum sp. is palatable, incidence of ingestion was lower than for P. micans in all five crab species (one-way ANOVA on arcsine-transformed data for H. oregonensis, P<0.05; no statistical tests run on other crab species due to zero variance in P. micans fed larvae in those experiments).

Larvae of five crab species ingested the small toxic dinoflagellate P. minimum (Fig. 2). To compare directly a non-toxic and a toxic species of Prorocentrum of equal sizes, one-way ANOVA (arcsine-transformed data) compared incidences of ingestion between P. dentatum and P. minimum for each crab species (Fig. 2). There were no significant differences in C. magister, C. gracilis, or H. nudus, while in C. oregonensis, ingestion of P. minimum exceeded that of P. dentatum (P<0.02) and in H. oregonensis, more feeding occurred on P. dentatum than on P. minimum (P<0.01; Fig. 2).

In both P. dentatum and P. minimum, gut fluorescence appeared to be less intense than was the case for P. micans strains. Typically, small reddish-orange fluorescent particles were present in the midgut, with individual cells apparent in the hindgut. This probably reflects the ingestion of fewer cells and, possibly, incomplete digestion of those cells.

In contrast to these five strains of Prorocentrum spp., the toxic P. hoffmanianum was almost never ingested (Fig. 2). However, the cultures of this dinoflagellate included dense strands of cells imbedded in mucous material that often covered the larvae and likely physically impeded ingestion of cells.

Genus Alexandrium (order Gonyaulacales)

Incidence of ingestion of A. fundyense exceeded 10% only in C. gracilis and C. oregonensis (Fig. 3), with only a few larvae of any crab species ingesting non-toxic A. tamarense 115 (Fig. 3) and none ingesting toxic A. tamarense 118 (data not shown). Gut fluorescence in larvae that did show some ingestion of the non-toxic strain suggests only a few intact cells present in the midgut.
Fig. 3

Mean percent ingestion (n=6; ±SE) by larvae of indicated crab species (abbreviations as in “Materials and methods—Experimental organisms”) on the indicated dinoflagellates (asterisk indicates toxic strain; NT indicates not tested; no bar indicates 0% response)

In contrast, larvae of three of the six crab species exposed to a non-toxic strain of A. catenella, similar in cell size to those described above, ingested substantial numbers of cells (Fig. 3). Incidence of ingestion equaled or exceeded 60% for larvae of C. gracilis, C. oregonensis, and H. oregonensis, with only larvae of C. productus showing no ingestion of this prey. Both whole cells and fragments were visible in the midguts of larvae ingesting the cells.

Larvae of all six crab species readily ingested a toxic strain of A. andersoni (Fig. 3), with mean incidence of ingestion always exceeding 50% and approaching 100% for larvae of C. gracilis. Under epifluorescence, larval guts were a dark reddish-orange, with bits of particulate matter often visible. There was evidence of fluorescence in the hindguts of a few H. nudus and C. gracilis larvae. In one case, an H. nudus larva was observed holding an A. andersoni cell in its maxillipeds, then releasing it.

The comparatively high palatability of A. andersoni begged the question of its impact on the development of newly hatched larvae, from both a nutritional and toxicity perspective. When mean incidence of ingestion was compared among the six crab species, a significantly higher percentage of C. gracilis and the two Hemigrapsus spp. larvae ingested prey than was the case with C. productus, C. oregonensis, or C. magister larvae (Table 2; one-way ANOVA on arcsine-transformed data; P<0.05). Based on this analysis, H. oregonensis (Ho), H. nudus (Hn), and C. magister (Cm) were selected for the development experiment, to include species showing both relatively high and low incidence of ingestion.
Table 2

Alexandrium andersoni. Mean percent of larvae ingesting prey and results of Tukey’s multiple-comparison test (n=6; P=0.05). Shared letters indicate no significant difference. Species’ abbreviations as in “Materials and methods—Experimental organisms”

Crab sp.

Percent ingestion

Tukey’s test

Cg

100

a

Hn

92

a

Ho

96

a

Cp

73

b

Cm

66

b

Co

57

b

Larvae of all three crab species survived well to zoeal stage 2 on the brine shrimp diet (mean days of molt: Cm, 11.0; Ho, 8.0; Hn, 10.0). Unfed larvae had all died by days 6 (Cm), 9 (Ho), or 11 (Hn). In C. magister, there was no delay in mean day of death for larvae fed the readily ingested P. micans as compared to unfed controls, although in both Hemigrapsus spp., significant delay occurred (ANOVA, P<0.05; Tukey’s HSD-test P=0.05; Table 3). Using the same tests, it was determined that there was no significant difference in mean days of death in C. magister between unfed larvae and those fed toxic A. andersoni. However, for both Hemigrapsus spp., mean day of death occurred significantly earlier for larvae fed A. andersoni than for larvae that were not fed at all (Table 3).
Table 3

Mean (±SE) days of death for first stage larvae of the three indicated crab species fed the indicated diets. Shared letters indicate no significant differences (Tukey’s HSD, P=0.05) within each crab species

Crab species and diet

Mean day of death

n

Tukey’s HSD

Cancer magister

  Prorocentrum micans

7.3±0.35

36

a

  Alexandrium andersoni

6.9±0.24

36

a

  Unfed

6.9±0.23

36

a

Hemigrapsus nudus

  P. micans

12.0±0.36

36

a

  Unfed

8.6±0.27

36

b

  A. andersoni

4.1±0.17

36

c

H. oregonenesis

  P. micans

7.8±0.66

22

a

  Unfed

4.6±0.20

24

b

  A. andersoni

3.2±0.40

24

c

Order Gymnodinales

While no larvae from any of the six crab species ingested a toxic strain of Gymnodinium catenatum (CCMP 1937), a few H. oregonensis larvae did ingest a non-toxic strain (CCMP 1940) (Fig. 4).
Fig. 4

Mean percent ingestion (n=6; ±SE) by larvae of indicated crab species (abbreviations as in “Materials and methods—Experimental organisms”) on the indicated dinoflagellates (asterisk indicates toxic strain; no bars indicate 0% response)

Incidence of ingestion of non-toxic Gyrodinium instriatum varied among crab species (Fig. 4). While no larvae of C. magister or C. productus ingested cells, 40% or more larvae of the other four crab species ingested this species. Gut coloration under epifluorescence was pale orange, with reddish bits of particulate matter present in the mid- and hindguts.

Order Gonyaulacales

Results similar to those described above for Gyrodinium instriatum were seen when larvae were exposed to non-toxic strains of both Gonyaulax spinifera and Protoceratium reticulatum (Fig. 4), although few C. oregonensis larvae fed on either prey, and few H. nudus larvae fed on P. reticulatum.

Mixed prey ingestion rates

Ingestion rates for H. oregonensis larvae exposed to a 50:50 mixture of Protoceratium reticulatum (Pr) and Prorocentrum micans (Pm) cells showed ingestion of both cell types (mean number of cells ingested per zoea per day, ±1 SD: Pr, 215±7; Pm, 231±3; n=3). Mean ingestion rate of P. micans was significantly higher than that of P. reticulatum (Kruskal–Wallis ANOVA; P<0.05).

Additional mixed prey experiments were carried out on larvae of H. oregonensis, C. magister, C. productus, and C. oregonensis, using a 50:50 mixture of P. micans and toxic A. fundyense. Ingestion rates differed significantly between the cell types in three of the four crab species tested (Fig. 5). Only for C. oregonensis were there no significant differences in ingestion rates between cell types (Fig. 5; Kruskal–Wallis ANOVA, P=0.88; n=3; G-power: 0.07). In the other three cases, ingestion rates were significantly higher for P. micans (n=3; P<0.05 in each case). In all cases, there was little or no evidence of ingestion of A. fundyense.
Fig. 5

Ingestion rates (±SD) for each constituent of a 50:50 prey mixture (Pm: Prorocentrum micans; Af: Alexandrium fundyense) for larvae of indicated crab species

Discussion

A broad range of autotrophic dinoflagellates are palatable to larval crabs (Table 4). Thirteen of the sixteen dinoflagellates were ingested by larvae of at least two crab species, although two of them (non-toxic Alexandrium tamarense 115 and non-toxic Gymnodinium catenatum) showed very low incidence of ingestion in only one and two of the crab species, respectively. Although all three dinoflagellates that were not ingested by any crab species were toxic, there is no consistent pattern of prey discrimination based on taxonomic affinity, toxicity, season of hatching, or predator/prey size relationships.
Table 4

Mean percent ingestion of dinoflagellates by larvae (NT not tested; asterisk indicates toxin-producing strain; diet abbreviations as in Table 1)

Crab species

Diet species

Aa*

Ac

Af*

A5

A8*

Gs

Gi

G37*

G40

Pm

P9

P3

Pd

Ph*

Pn*

Pr

Cancer gracilis

100

60

53

NT

0

49

44

0

0

100

100

100

42

0

51

78

C. magister

65

9

11

0

0

0

0

0

3

100

100

100

69

0

42

0

C. oregonensis

57

75

37

12

0

14

52

0

0

100

100

100

52

0

72

1

C. productus

73

0

0

0

0

0

0

0

0

100

100

100

NT

0

NT

0

Hemigrapsus nudus

92

16

4

0

0

87

60

0

0

100

100

100

68

0

67

12

H. oregonensis

92

84

0

0

0

84

30

0

7

87

100

98

81

0

59

34

Consistent with previous reports (Lehto et al. 1998; Hinz et al. 2001), Prorocentrum micans was readily ingested by a large proportion of larvae of all six crab species. Both non-toxic P. dentatum and toxic P. minimum were also palatable, although incidence of ingestion of these smaller cells was lower than that of P. micans. However, there was no apparent relationship between prey size and that of predators, with larvae of all crab species showing virtually the same results. Only toxic P. hoffmanianum was not palatable, the results confounded by its production of mucous strands that appeared to impede ingestion physically. The generally high incidence of ingestion of Prorocentrum spp. may be due to their flattened shape, with cells repeatedly observed being ingested whole. In contrast, cells of most other dinoflagellates appeared to be broken open before being ingested.

Our results do not support the suggestion of Hinz et al. (2001) that dinoflagellates of the genus Alexandrium are generally unpalatable to larval crabs. Although we confirm and expand upon their report that very few larvae ingest toxic and non-toxic strains of A. tamarense and that toxic A. fundyense is only rarely ingested (Hinz et al. 2001), both non-toxic A. catenatum and toxic A. andersoni were readily ingested by five and all six crab species, respectively.

The unpalatability of toxic A. tamarense contrasts with the results of Yazdandoust (1987), who reported that larvae of Cancer anthonyi thrived on this dinoflagellate. However, Yazdandoust (1987) used a different species of Cancer than used here, and his strain of A. tamarense appears to be quite different. He actually identified his strain as Gonyaulax catenella (a species identified as A. tamarense by Balech 1985) and described them as chain-forming. The toxic A. tamarense strain used in the present study showed no aggregation characteristics, although the non-toxic strain tended to collect at the surface.

There is no evidence of a taxonomic basis to dinoflagellate palatability with respect either to the predator or prey. Ingested dinoflagellates belong to all three orders (Prorocentrales, Gonyaulacales, and Gymnodinales), and the three dinoflagellates that were never ingested by larvae of any crab species belong to two of the three orders (Table 4). The six crab species tested belong to two families, Cancridae and Grapsidae. Larvae of all species except C. productus ingested more than half of the prey types presented, with C. productus ingesting only 4 of 14. C. productus thus appears to be the most selective of the crab species tested, with three of the four dinoflagellate strains ingested belonging to one species, the apparently universally palatable P. micans. However, other species of the same crab genus are more flexible with respect to ingestion of dinoflagellates, and C. productus is neither the largest larva nor the only winter-hatching species tested. Thus, the reasons for higher selectivity by C. productus are not apparent in the present study.

There is no apparent relationship between larval size and that of dinoflagellates ingested, with the largest larva (C. magister) ingesting the smallest dinoflagellate (P. dentatum) and the smallest larvae (Hemigrapsus oregonensis and C. gracilis) ingesting the largest cells (e.g. G. spinifera). This is not surprising, given that larval crabs ingest microparticles as small as 1.0 µm (Factor and Dexter 1993), meso-zooplankton as large as 400 µm, and a heterotrophic dinoflagellate of 350 µm in diameter (Lehto et al. 1998).

There also seems to be little relationship between the propensity for larvae to ingest dinoflagellates and whether they are likely to encounter high densities of them in their environments. Primary productivity in the Puget Sound region is very low in winter (Copping 1982; Brainard 1996), so that species with larvae hatching in winter are unlikely to encounter dense patches of dinoflagellates. Although one of the three species hatching in winter, C. productus, ingested fewer dinoflagellate types than did the other five crab species; the other two, C. magister and C. oregonensis, ingested about the same numbers of dinoflagellate types as did the three species that hatch in spring and summer. However, because larvae of all three winter-hatching species are in the water column until well after spring bloom, later larval stages are likely to encounter dense patches of phytoplankton, including dinoflagellates. Mechanisms involved in prey selection may thus be present in the hatching stage, even when they may not be truly significant until later in development.

The capacity of larvae to discriminate among dinoflagellate species/strains is perhaps most vividly illustrated when they are presented with toxic cells. It is clear that cell toxicity is not the sole basis of discrimination. Of the six toxic strains presented to larvae, two were frequently ingested by five crab species; one, frequently ingested by two crab species; and three, never ingested. Of the two frequently ingested dinoflagellates, one contains saxitoxins; the other, venerupin. Avoidance of A. tamarense was not related to strain toxicity. It is possible that ingestion of toxic cells may occur for only a brief initial period. Hinz et al. (2001) indicated that C. oregonensis ingested toxic A. fundyense during the first 6 h of exposure, with ingestion declining over the following 42 h. Moore (2003) found that gut fluorescence associated with ingestion of P. micans persisted for up to 24 h after larvae were removed from the prey, although ingestion of such cells occurred continually over that time period when larvae were continuously exposed to them. Since observations were made only after 24 h, it was not possible to determine, based on gut fluorescence, whether ingestion of toxic cells had been continuous.

Hinz et al. (2001) reported that, based on gut fluorescence, larvae would ingest toxic cells when they were presented in combination with P. micans, even though such toxic cells were not ingested when presented alone. However, ingestion rate results reported here for such prey combinations (Fig. 5) indicate that if ingestion of toxic cells occurs under such circumstances, the number of such cells ingested is low. Furthermore, although Perez (2003) observed that exposure to even a few toxic cells impaired locomotory activity in some larvae, Sulkin et al. (2003) suggested that reduced swimming, resulting from encounters with blooms of toxic algae, might lead to larvae sinking out of the bloom, thus reducing their exposure. Nevertheless, when the capacity to discriminate among prey does not prevent significant ingestion of a toxic dinoflagellate, the result can lead to accelerated mortality. Thus, the capacity to discriminate among dinoflagellate types will not always prevent ingestion of cells that will prove harmful.

Hinz et al. (2001) and the present study have demonstrated that larval crabs discriminate among dinoflagellate prey species. Hinz et al. (2001) further reported that such discrimination occurs at the time of actual ingestion; that is, virtually all dinoflagellate prey are captured, but only some are ingested. This often occurs after manipulation of the prey cell by the mouth parts, implying the presence of a cue that is located on the surface of the cell. The basis for such discrimination does not appear to be related to prey taxonomic affinity, size, or toxicity. Dinoflagellates excrete complex macromolecules onto their cell surface, aiding in regulation of cellular uptake of dissolved compounds and in recognition of predators or prey (Wolfe 2000; Sakaguchi et al. 2001). Such organic moieties or other surface properties may serve as ingestion cues for larval crabs.

The present study adds to what is already a complicated and somewhat paradoxical model of feeding in newly hatched crab larvae. Such larvae must satisfy nutritional needs by feeding in the plankton soon after hatching, and typically can survive only if they feed on large micro- or meso-zooplankton. Given their uncertain prey environment, however, larvae appear to be opportunistic encounter feeders that will ingest protists and even detritus, presumably using this nutrition either to sustain them until or between encounters with zooplankton prey or to supplement such prey. Although a wide variety of prey are ingested, larvae nevertheless can and do discriminate among protists (dinoflagellates) after capturing them and will even ingest prey that are destined to harm them, in spite of the capacity to discriminate among cells.

The role played by these selective predators in regulating protist communities has not been examined. However, given the vast numbers of larvae produced in local habitats over a relatively short time (e.g. Coyle and Paul 1990; Schwamborn et al. 1999), “top down” control of species composition and relative abundance within phytoplankton communities could be substantial. Furthermore, since most larvae are themselves prey for jellyfish, finfish, and larger plankton, most of the protist carbon consumed by them will enter the metazoan food web.

Acknowledgements

This paper includes data submitted in partial fulfillment of the Master’s of Science degree to Western Washington University by the first author. The first author was supported by graduate assistantships provided by the Shannon Point Marine Center and a U.S. Environmental Protection Agency STAR Graduate Fellowship. We thank N. Moore and H. Ko for technical assistance. Experiments conducted as part of this research comply with the laws of the United States of America.

Copyright information

© Springer-Verlag 2004