Amer. Zool., 32:503-511 (1992)
Constraint on Reproductive Output in Brachyuran Crabs: Pinnotherids Test the Rule1
Anson H. Hines Smithsonian Environmental Research Center, P.O. Box 28, Edgewater, Maryland 21037
Synopsis. Brood weight and number of eggs per brood are primarily determined by female body size in brachyuran crabs. Brood weight exhib? its an isometric relationship spanning more than four orders of magnitude in body weight among 33 non-pinnotherid species, with brood size being constrained to about 10% of female body weight by the space available for yolk accumulation within the cephalothorax. In contrast, two pin? notherids (Pinnotheres ostreum, Fabia subquadrata) have relative brood sizes which are 66% and 97%, respectively, of body weight. The allometric constraint is circumvented in these extraordinarily large brood sizes by two pinnotherid features which allow more space for yolk accumulation. Unlike other brachyurans, their ovaries extend out of the cephalothorax into the abdomen, and calcification of their exoskeleton is greatly reduced, potentially allowing distension of the body. As in crabs, pinnotherid egg size is highly variable among species, which tends to increase the variance among species in fecundity per brood, making pinnotherid fecundities not extraordinary. The number of broods produced per year in these two species is limited by short reproductive seasons, which tends to bring their annual reproductive output in line with other brachyurans. The trade-offs among reproductive variables in pinnotherids is adaptive for the small body size and parasitic niche of these species.
Introduction strained to about 10% of female body weight available for accumulation Physical limits and phylogenetic history by space yolk within the calcified sometimes impose constraints upon evo? cephalothorax (Hines, size is a trait lutionary solutions to ecological problems 1982). Egg major reproductive which is variable (Stearns, 1977). Surface-to-volume ratios highly among brachyuran As and effects of fluid viscosity are examples species (Hines, 1982,1986, 1988,1991). a of the constraint in brood of physical properties which can constrain consequence size relative to the scaling patterns in such physiological prob? body size, high interspe? cific in size results in much lems as metabolism, movement in water, variability egg variation in and circulation. As a phylogenetic example, greater interspecific fecundity small versus few than in evolution of an external skeleton imposes (many large eggs) brood mass for sized crabs limits on growth patterns in arthropods. equivalently These kinds of constraints set limits for (Hines, 1982). Recent analysis of covaria- tion in these traits within two variation in certain traits, and suites of traits reproductive families of crabs and Cancri- may evolve in patterns of trade-offs as a (Geryonidae that the in consequence of these constraints. dae) suggests, however, range relative brood size of some be Brood size and fecundity are primarily species may as much as about twice the 10% determined by female body size in brachy? average a wide of uran crabs (Hines, 1982, 1988, 1991). Brood among array species (Hines, 1988, weight in brachyuran crabs is generally con- 1991). In this paper, I provide a further test of 1 the hypothesis that reproductive output per From the Symposium on The Compleat Crab pre? brood in is constrained sented at the Annual Meeting ofthe American Society brachyurans by space of Zoologists, 27-30 December 1990, at San Antonio, available in the cephalothorax, resulting in Texas. limited variation in brood weight relative 503 504 Anson H. Hines
Table 1. Collecting information for crab species.
to female body weight. Because parasitic tive purposes, samples of 7-15 ovigerous species often have higher reproductive out- females of four non-pinnotherid species of puts than free-living species, I measured the crabs (Neopanope sayi, Portunus spinicar- brood size, egg size, and fecundity relative pus, Menippe nodifrons, Loxorhynchus to female body size in two species of pin? grandis) were also collected from locations notherid crabs which live obligatorily within listed in Table 1. The samples for analysis the mantle cavities of bivalve molluscs. Pin- were selected to span the size range of ovig? notheres ostreum occurs within the oyster erous females collected at the sites. Although Crassostrea virginica (also some species of the size range of sampled crabs did not nec? scallops and mussels) along the east coast essarily span the range reported in the lit? of North America, throughout the Gulf of erature for reproductive females, the range Mexico, Caribbean, and to Brazil (Chris- of sampled crabs was in all cases more than tensen and McDermott, 1958; Williams, adequate to determine the size-dependent 1984). Fabia subquadrata inhabits the mus- regressions for reproductive variables (see sel Mytilus californianus (also other mussel Results). The collectors selected crabs with species and some clams) along the west coast brooding embryos in developmental stages of North America from Alaska to southern from late blastula to early gastrula to avoid California (Pearce, 1966; Garth and Abbott, confounding effects of egg diameters swell- 1980). Both crab species consume food par? ing during late stages of development or of ticles and mucus on the gills of their host, egg loss from the brood during incubation. sometimes inflicting significant damage to The samples were frozen and then fixed in one ctenidium. The life histories of both 10% formalin-seawater and stored in 70% species are similar. After post-larval mega- ethanol until they were processed. The fol? lopae infect the bivalve host from the plank- lowing variables were measured for each ton, the crabs begin juvenile growth. Mating crab: maximum (spine-to-spine) carapace occurs outside the host, however, when the width (mm) ofthe female; female dry body female leaves the host and re-enters the water weight (g); dry weight (g) ofthe brooded egg column temporarily. Inseminated females mass; average diameter (\im), including the return to a bivalve host and continue to chorionic membrane adhering tightly to the grow and produce eggs, which are brooded embryonic surface, ofa subsample of seven in the usual brachyuran way. Males are dwarf eggs before drying; and the number of eggs and are also found within the bivalve host per brood, extrapolated from the dry weights part of the time, though usually not with a ofthe total brood and a counted subsample female. of about 2,000 eggs (Hines, 1982, 1988, 1991). All broods were examined under a Materials and Methods dissecting microscope for the presence of Samples of 12-16 ovigerous females of nemertean worms or other potential infec? Pinnotheres ostreum and Fabia subquadrata tions by egg predators. Data on numbers of were collected from Crassostrea virginica broods produced per year in the pinnothe- and Mytilus californianus, respectively, at rids were extracted from Christiansen and locations shown in Table 1. For compara- McDermott (1958) and Pearce (1966). Constraint on Reproductive Output in Crabs 505
r^ o\ o\ o\ vo Tt d d d d o c5
o\ o co o r^ r^ 00 OnOO O ?? vo *1 ^ P ?* vq tj- U^U^ Tf + + + + + + XXX XXX
yielded significant regression coefficients in 102-, ? Loxorhynchusgrandis all species, which is an important compo? ? Menippenodifrons f Neopanopesayl nent of the subsequent regression analyses * Portunusspinicarpus and allometric considerations (Table 2). S 10? They are also a good indicator ofthe range in body size among species (see Hines 1982, ?10-1 1988, 1991, and below). o oe Nemertean worms (Carcinonemertes spp.) m W2 or other potential egg predators were not observed in the egg masses of any of the 10-3 samples. 106! Mean sizes and egg (diameter volume, 105 varied the respectively) considerably among 10< four non-pinnotherid species: Neopanope UJ 103 314 0.0164 Portunus sayi /*m, mm3; spini- 5 102. carpus 336 /im, 0.0201 mm3; Menippe nodi- Z 101 frons 365 /un, 0.0258 mm3; and Loxorhyn- o g 10? chus grandis 666 nm, 0.159 mm3. Egg sizes 10-1 (diameter and volume, respectively) varied 10-2. * widely between the two pinnotherid species: 10-3- Pinnotheres ostreum 260 iim, 0.0092 mm3; .1 1 10 Fabia subquadrata 414 jum, 0.0372 mm3. BODYWEIGHT (g) Egg size did not vary significantly within Fig. 2. Brood weight (top) and number of eggs per species (ANOVA, P > 0.05). Egg size did brood (bottom) versus body size in samples of four non- vary significantly among the six pinnotherid pinnotherid species of crabs spanning four orders of magnitude in dry body weight. Note log-log scales. and non-pinnotherid species (ANOVA on = Regression equations are given in Table 2. Log Egg Diameter: F(5>64) 221.95, P > 0.001). Although the largest species, Lox- orhynchus grandis, had the largest egg, and female body weight both within and among Pinnotheres ostreum (the smallest species) species (Table 2, Fig. 2, ANOVA of Log had the smallest egg, mean egg size was not Brood Weight on Log Body Weight, P < correlated with female body size among the 0.05); and mean dry brood weight ranged six species (ANCOVA of Log Egg Volume from 0.0259 g in Neopanope sayi to over = on Log Body Weight, F(5>64) 0.36, P > 18.2 g in Loxorhynchus grandis. Similarly, 0.55). However, egg size was significantly brood weight for the pinnotherid species correlated with female body size among the increased significantly with female body 35 species in the larger data set (Fig. 1; weight within species (Fig. 3; ANOVA on ANOVA of Log Egg Volume on Log Body Log Brood Weight on Log Body Weight, P = Weight, F(U4) 8.328, P < 0.01), but vari? < 0.001); and mean brood weight ofthe ation in body size only explained a small pinnotherid species overlapped that of the = portion of the variation in egg size (R2 smallest non-pinnotherid species: Pinno- 0.201) (Hines, 1982, 1988, 1991; see also theres ostreum 0.0237 g; Fabia subquadrata Hines, 1986). Along with Callinectes sapi? 0.0697 g. Because regression slopes differed dus, P. ostreum had the smallest egg of the significantly among the 6 species (ANCOVA 35 species, and had an egg which was smaller of Log Brood Weight on Log Body Weight than the 95% confidence limits ofthe mean with significant interaction of Log Brood for the ? 95% x = group (mean confidence lim? Weight Log Body Weight, F(5?64) 3.06, its: Egg Diameter 406 ? 20 ixm; Egg Volume P < 0.02), further contrasts of reproductive 0.045 ? 0.008 mm3). Fabia subquadrata, effort using ANCOVA were not possible. To on the other hand, had an egg size which compare brood size ofthe pinnotherids with was near the mean for the group. an array of other non-pinnotherid species, Brood weight for the four non-pinnothe? mean log brood weight was regressed on rid species increased significantly with mean log body weight for each ofthe 6 spe- Constraint on Reproductive Output in Crabs 507
10? 10J ? Pinnotheresostreum O Fa6/asubquadrata 102 3 3 Y= 0.962X - 0.940 R*2 = 0.930 + " _ "? .V UJ O o ? 10o Q ?' ? ? ? o Q o >** cc 810-1 m z < UJ < i 10-3 10-4 105. 10?
OC UJ m
O 104
103 .01 .1 1 10 100 1000 MEANBODY WEIGHT (g) MEANBODY WEIGHT (g)
Fig. 3. Brood weight (top) and number of eggs per Fig. 4. Mean brood weight (top) and mean number brood (bottom) versus dry body weight in samples of of eggs per brood (bottom) versus mean body weight two species of pinnotherid crabs. Note log-log scales. in 35 species of crabs from this paper and Hines (1982, Regression equations are given in Table 2. 1988, 1991). Regression equations and R2 values are indicated. Dashed lines indicate the 95% confidence belts ofthe data in the regreassions. Note that only the mean brood for the two cies along with comparable data for 29 other weights pinnotherid species fall outside the confidence limits. Note log-log scales. species extracted from earlier publications (Hines, 1982, 1988, 1991). Whereas mean brood weights for all 33 non-pinnotherid among non-pinnotherid species, ranging species lie within the 95% confidence limits from 1,790 eggs per brood in the smallest of the regression, brood weights of the two Neopanope sayi to 634,000 eggs per brood pinnotherid species were significantly larger in the largest Loxorhynchus grandis (Fig. 2; and fell outside the confidence belt (Fig. 4). Table 2). However, variance in size-specific Based on aposteriori chance alone for a 95% fecundity among species was greater than confidence limit for a sample size of 35, 1.9 for brood weight, and it is not initially clear species might be expected to lie outside the for this limited number of species whether confidence belt of the regression; but the a the increase in fecundity with body size priori chance of both pinnotherids being among species is best described by a linear those species is only 1 in 400 (0.25%). or curvilinear regression of log egg number Mean brood weight as a fraction of mean versus log body weight (Fig. 2, but see below). body weight for the 33 non-pinnotherid spe? For the two pinnotherid species, log egg cies averaged 10.7% and ranged from as low number per brood was also a linear function as 3.2% in Portunus spinicarpus to as high of log body weight, with mean fecundities as 23.0% in Chaceon quinquedens (Fig. 5). overlapping the small non-pinnotherid spe? In contrast, mean brood weight for the two cies: Pinnotheres ostreum 5,680 eggs per pinnotherids was very much larger at 66.2% brood; Fabia subquadrata 7,560 eggs per for Pinnotheres ostreum and 96.7% for Fabia brood (Fig. 3). Again, because regression subquadrata (Fig. 5). slopes differed significantly among the 6 Fecundity per brood also increased sig? species (ANCOVA of Log Brood Weight on nificantly with body weight within and Log Body Weight with significant interac- 508 Anson H. Hines
z o
tn o < w
n Q U. OO 20 o f? ffi *???? 4*
MEANBODY WEIGHT (g)
Fig. 5. Brood weight in proportion to female body size for 35 species of crabs from this paper and Hines (1982, 1988, 1991) spanning over four orders of magnitude in dry body weight. Note that brood weight in non- pinnotherid species averaged about 10% (minimum = 3%, maximum = 22%) of female body weight, but brood sizes in the two pinnotherid species were 66% (Pinnotheres ostreum) and 97% {Fabia subquadrata). Note log scale on the X-axis. tionofLog Egg Number x Log Body Weight, they lie within the 95% confidence limits of F(5,64)= 3.85, P < 0.01), further comparison the mean. of fecundities using ANCOVA was not The mean brood weight per year and the possible. Regression of mean log egg num? mean fecundity per year produced by each ber per brood on mean log body weight for species was estimated by multiplying the each ofthe 35 species reported here and in mean brood weight or the mean number of earlier papers (Hines, 1982, 1988, 1991) eggs per brood, respectively, by the mean showed that the two pinnotherid species had number of broods per year. Mean annual relatively high fecundities for their body size reproductive output (brood weight) is a lin? (Fig. 4). However, while the fecundity of P. ear function of body weight (Fig. 6; ANOVA, ostreum fell on the 95% confidence = P < The upper F(1,34} 426.1, 0.001). pinnotherid limit, that for F. subquadrata was within the species had large reproductive outputs, but belt at a deviation from the mean similar outputs of none ofthe 35 species fell outside to that for Callinectes sapidus (Fig. 4). the 95% confidence belt of the regression. Mean numbers of broods per year pro? Mean annual fecundity is also a linear func? duced by the four non-pinnotherid species tion of body weight (Fig. 6; ANOVA, F(134) = are: 1.5 in Neopanope sayi (Swartz 1972); 190.3, P < 0.001). Annual fecundities of 2.0 in Portunus spinicarpus (Camp and the pinnotherids were high, but only Cal? Whiting, 1991), 2.0 in Menippe nodifrons linectes sapidus had a reproductive output (Wilber, 1989; Hines, personal observa? outside the 95% confidence belt for the tion); and 3.5 in Loxorhynchus grandis regression. Again, by definition, there is an (Hines, unpublished). Similarly, the two a posteriori probability of 1.9 species falling pinnotherid species each produce a mean of outside the 95% confidence belt. 1.5 broods (range 1-2 broods) per year over a distinct reproductive season from June Discussion through August in Pinnotheres ostreum Comparison of reproductive variables (Christensen and McDermott, 1958) or between pinnotherid species and other fam- December through March in Fabia ilies provides insight into the evolutionary subquadrata (Pearce, 1966). Considering the patterns of life history strategies in the total of 35 species reported in this and ear? Brachyura. Body size is the principal deter- lier publications (Hines, 1982,1988, 1991), minant of reproductive output and fecun? the pinnotherids produce somewhat fewer dity in brachyuran species ranging over 4 broods per year than the overall average of orders of magnitude in body weight (Hines, 2.4 broods (range 1.0 to 8.0 broods), but 1982, 1988, 1991). Although pinnotherids CONSTRAINT ON REPRODUCTIVE OUTPUT IN CRABS 509
available in any ofthe non-pinnotherid spe? > cies, which have ovaries confined to the m ' * ' UJ Y= 0.879X - 0.547 R*2 = 0.928 , ^fc , cephalothorax. Second, unlike well-calci- a. 10ij fied, non-pinnotherid species, females (but not males) of these pinnotherids have very 1 io?-l little calcium carbonate deposited in their Q exoskeleton (Christiansen and McDermott, O O 10-1-, 1958; Pearce, 1966; Hines, personal obser? s ffi vation), which makes their exoskeleton flex? ible, and which provides the possibility that 10-24 the body is distensible during yolk accu- mulation. < Just as egg size is a major variable of UJ ? Y= 0.711X +4.498 RA2 =0.852 ? - reproductive strategies in non-pinnotherid 106 brachyurans (Hines, 1982, 1986, 1988, UJ 1991; Fig. 1), egg size is highly variable S io5 between the two pinnotherids from very 3 - ? ? ? Z small to in size. Two con? (3 nearly average ui 104 sequences of this variation in egg size are z evident in fecundities of the pinnotherids Fig. 7. (Left) Camera lucida drawing of ventral view of female Pinnotheres ostreum with ripe ovaries extending into abdomen prior to egg extrusion. (Right) Camera lucida drawing of frontal view of brooding female Fabia subquadrata. In specimens newly preserved in formalin, the ovaries and newly extruded eggs appear light orange and the crabs are light yellow. Note very large abdomenal flap, which aids in incubation of extraordinarily large brood mass relative to body size in both species. nal parasites. Just as many invertebrates face crabs only by evolving two mechanisms "evolutionary difficulties" imposed by small (extension of ovaries into the abdomen and body size upon reproductive tactics (Menge, distensible exoskeleton) that circumvent this 1975; Strathmann and Strathmann, 1982), allometric constraint. a crab the size of a pinnotherid is markedly limited in its absolute reproductive output. ACKNOWLEDGMENTS In species which brood, there are only three I thank the for options for increasing fecundity: reducing following people collecting assistance: S. A. Reed for Pinnotheres egg size (more but smaller eggs); increasing ostreum and F. A. the number of broods per year; and/or Menippe nodifrons, for Fabia P. Had- increasing brood mass. Because the cum- Chapman subquadrata, don for and J. D. Shields mulative data indicates that 250 ^m appears Neopanope sayi, for I am to to be approaching the lower limit of egg size Loxorhynchus grandis. grateful E. L. T. and D. for brachyurans (Fig. l;Hines, 1982, 1988, Wenner, Mantel, Wolcott, Wolcott for their efforts to the 1991; Hines and Morgan, unpublished), and organize sym? in honor of Bliss called since eggs of Pinnotheres ostreum are near posium Dorothy "The of which this that size limit, reducing egg size alone would Compleat Crab," paper is a E. Wenner and two reviewers not appear to increase fecundity "enough." part. pro? vided comments on the Both of these species also appear to be helpful manuscript. This work was in restricted in the number of broods produced supported part by funding from the Smithsonian Marine Station at by relatively short reproductive seasons. Link this is contribution no. 282 of Since this seasonal limit applies to both Port; the Smithsonian Marine Station at Link winter and summer cycling species, it is not Port. clear whether ecological (e.g., timing of lar? val settlement) or energetic (plankton pro? duction for the host's food) considerations References pertain. Clearly, the parasitic habitat ofthe Camp, D. K. and N. H. 1991. adult females of these species minimizes the Whiting. Population biology of four species of crabs, of reduced calcification in the swimming genus disadvantages Portunus (Decapoda, Portunidae), off Cape exoskeleton, because the host bivalve both Canaveral, Florida. Florida Marine Research Pub? restricts movement and provides protection lications, St. Petersburg, Florida. (In press) M. E. and J. J. McDermott. 1958. Life- from predators, thus negating two of the Christensen, history and biology ofthe oyster crab, Pinnotheres major functions of a rigid, non-distensible ostreum Say. Biol. Bull. 114:146-179. exoskeleton. these Thus, pinnotherids pro? Garth, J. S. and D. P. Abbott. 1980. Brachyura: The vide a test of the rule that brood size is true crabs. In R. H. Morris, D. P. Abbott, and E. constrained to about 10% of female body C. Haderlie (eds.), Intertidal invertebrates of Cal? ifornia, pp. 595-630. Stanford Univ. Palo weight by space available for yolk accu? Press, Alto, California. mulation in the Pinnothe? cephalothorax. Hines, A. H. 1982. Allometric constraints and vari? rids are able to produce broods which are ables of reproductive effort in brachyuran crabs. proportionately much larger than in other Mar. Biol. 69:309-320. CONSTRAINT ON REPRODUCTIVE OUTPUT IN CRABS 5 11 Hines, A. H. 1986. Larval patterns in the life histories research. 2nd ed. Freeman Co., New York, New of brachyuran crabs (Crustacea, Decapoda, York. Brachyura). Bull. Mar. Sci. 39: 444-466. Stearns, S. C. 1977. The evolution of life history Hines, A. H. 1988. Fecundity and reproductive out? traits: A critique ofthe theory and a review ofthe put in two species of deep-sea crabs, Geryon fen- data. Ann. Rev. Ecol. Syst. 8:145-171. neri and G. quinquedens (Decapoda: Brachyura). Strathmann, R. R. and M. F. Strathmann. 1982. The J. Crust. Biol. 8:557-562. relationship between adult size and brooding in Hines, A. H. 1991. Fecundity and reproductive out? marine invertebrates. Am. Nat. 119:91-101. put in nine species of Cancer crabs (Crustacea, Swartz, R. C. 1972. Postlarval growth and reproduc? Brachyura, Cancridae). Can. J. Fish. Aq. Sci. 48: tive biology of the xanthid crab, Neopanope tex- 267-275. ana sayi. Ph.D. diss., College of William and Mary, Menge, B. A. 1975. Brood or broadcast? The adaptive Williamsburg, VA. Dissertation Abstracts 33: significance of different reproductive strategies in 7150-7151B. the two intertidal sea stars Leptasterias hexactis Wilber, D. H. 1989. The influence of sexual selection and Pisaster ochraceus. Mar. Biol. 31:87-100. and predation on the mating and postcopulatory Pearce, J. B. 1966. The biology ofthe mussel crab, behavior of stone crabs (Xanthidae, Menippe). Fabia subquadrata, from the waters of the San Behav. Ecol. Sociobiol. 24:445^151. Juan Archipelago, Washington. Pacific Sci. 20:3- Williams, A. B. 1984. Shrimps, lobsters, and crabs of 35. the Atlantic coast of the Eastern United States, Sokal, R. R. and F. J. Rohlf. 1981. Biometry. The Maine to Florida. Smithsonian Institution Press, principles and practice of statistics in biological Washington, D.C.