Egg Size Determines Offspring Size in Neotropical Fishes (Teleostei: Cichlidae) Author(s): Ronald M. Coleman and Alison P. Galvani Source: Copeia, Vol. 1998, No. 1 (Feb. 3, 1998), pp. 209-213 Published by: American Society of Ichthyologists and Herpetologists Stable URL: http://www.jstor.org/stable/1447721 Accessed: 01/05/2009 00:45

Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at http://www.jstor.org/page/info/about/policies/terms.jsp. JSTOR's Terms and Conditions of Use provides, in part, that unless you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you may use content in the JSTOR archive only for your personal, non-commercial use.

Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at http://www.jstor.org/action/showPublisher?publisherCode=asih.

Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed page of such transmission.

JSTOR is a not-for-profit organization founded in 1995 to build trusted digital archives for scholarship. We work with the scholarly community to preserve their work and the materials they rely upon, and to build a common research platform that promotes the discovery and use of these resources. For more information about JSTOR, please contact [email protected].

American Society of Ichthyologists and Herpetologists is collaborating with JSTOR to digitize, preserve and extend access to Copeia.

http://www.jstor.org Copeia,1998(1), pp. 209-213

Egg Size Determines Offspring Size in Neotropical Cichlid Fishes (Teleostei: Cichlidae)

RONALD M. COLEMAN AND ALISON P. GALVANI

Within the family Cichlidae, the size of eggs We investigated the relationship between egg varies tremendously across species (Breder and size and hatchling size in 26 species of Neotrop- Rosen, 1966; Fryer and Iles, 1972; Coleman, ical . We also examined eggs and hatch- 1991). Sizes range from the small eggs (0.9 mm lings produced by 17 females within one species in diameter) of rams (Microgeophagusramirezi) to [convict cichlids, "Cichlasoma()" ni- the massive eggs (4.5 mm in diameter) of some grofasciatum], and eggs and hatchlings from of the African mouthbrooding cichlids such as within single clutches of two species (convict Cyphotilapiafrontosa (unpubl. data). The hatch- cichlids and rainbow cichlids, Herotilapia multis- lings of the large-egged mouthbrooders are pinosa) to look at the effects of intraspecific and clearly much larger than the hatchlings of the intraclutch variation in egg size. rams and other small-egged species (pers. obs.). the between size and However, relationship egg MATERIALSAND METHODS hatchling size for smaller differences in egg size has not been For within the explored. example, Cichlid eggs are not spherical but generally the of sizes is Neotropical cichlids, range egg are better described as prolate spheroids (elon- much smaller: such as Cich- small-egged species gated spheres). To compare nonspherical eggs, lasoma dimerushave an diameter of average egg Coleman (1991) proposed using the "effective 1.3 whereas the mm, largest-egged species, diameter," which is the diameter the egg would "Cichlasoma tuba has an (Theraps)" average egg have if its contents were reshaped into a perfect diameter of 2.4 mm Does (RMC, unpubl. data). sphere. In the case of a prolate spheroid, the this small difference in size actu- relatively egg effective diameter (de), as used throughout this translate into a consistent difference in the ally paper, is the cube root of the major axis multi- size of the does the hatchlings? Furthermore, plied by the square of the minor axis (i.e., cube variation in size within a ac- egg single laying root of length X width X width). count for differences in size? Under- hatchling The systematics of Neotropical cichlids are in the of differential allo- standing consequences a state of flux, creating confusion in the generic cation to size between in a egg eggs single status of many of the species used in this study and will clutch and within among species help (Kullander, 1983; Stiassny, 1991). We follow the evolution us understand the factors shaping Stiassny (1991) in referring to certain of these of egg size. fishes by the genus " Cichlasoma,"including the not translate into A larger egg may necessarily section names provided by Regan (1906, 1906- a larger hatchling for several reasons. For ex- 1908) as though they designate subgenera; how- ample, a given measure of egg size may not be ever, the exact phylogenetic meaning of these a true representation of the energetic and ma- sections is uncertain. The 26 species examined terial resources of an egg. This could be the were Aequidens pulcher, Cichlasoma bimaculatum, case if larger eggs are simply filled with more C. dimerus, C. festae, C. pusillum, "Cichlasoma water than smaller eggs, and size differences ()" alfari, "C. (Am.)" citrinellum,"C. may solely be accounted for by water content. (Am.)" longimanus, "C. (Am.)" rostratum, "C. In such a case, the difference in egg size may (Nandopsis)" managuense, "C. (Archocentrus)" be insignificant to the resulting hatchling. This centrarchus,"C. (Ar)" nigrofasciatum,"C. (Ar.)" is particularly plausible because fish eggs are hy- septemfasciatum,"C. (Ar)" spilurum, "C. (Ther- drated and expand to some extent when they aps)" bifasciatum, "C. (Th.)" intermedium, "C. leave the female (Kamler, 1992). An ideal mea- (Th.)" nicaraguense, "C. (Th.)" sieboldii, "C. sure of egg size may involve a complete bio- (Th.)" tuba, Cleithracaramaronii, Crenicichlalepi- chemical analysis of the contents of an egg, re- dota, Herotilapiamultispinosa, Microgeophagus altis- porting protein, lipid, and carbohydrate con- pinosa, Nannacara anomala, Neetroplusnematopus, tent (e.g., Kamler, 1992); however, this is not and Pterophyllumscalare. practical for many studies. Indeed, for situations The fishes were maintained in the laboratory when the eggs are particularly small and yet using a typical freshwater aquarium setup. Each must be kept alive, even weighing individual setup consisted of a glass aquarium (38-280 li- eggs or measuring their volume can be difficult. ters), gravel, sponge filter, heater, and either

? 1998 by the American Society of Ichthyologists and Herpetologists 210 COPEIA, 1998, NO. 1 pieces of slate, flowerpots, or broken flowerpots. pect significant positive relationships in either All species examined were biparental substra- analysis. tum spawners (Loiselle, 1994), though differing To examine intraspecific variation, we sam- in the precise choice of spawning site (e.g., flow- pled eggs and hatchlings from single spawnings erpot, piece of slate, exposed glass tank bot- of 17 convict cichlid females. To assess the effect tom). Water temperatures were maintained be- of variation within a clutch in size of egg on size tween 24 and 28 C, and the light cycle was 12D/ of offspring, we obtained individual hatchlings 12L. Fishes were fed twice daily on Purina trout from eggs of known size using single clutches chow and Tetramin cichlid flakes with weekly of convict cichlids and rainbow cichlids. In both supplements of live brine shrimp. cases, eggs were scraped off the laying site as Aquaria were observed daily for the presence described above and their size measured in a of eggs or hatchlings. When a clutch of eggs was petri dish. For the within-clutch analyses, we se- sizes as as of more in- found, a sample of 30-50 eggs was scraped from lected extreme well eggs of the substratum with a blunt metallic spatula. termediate size to maximize the range egg sizes used. Each was measured under the The rest of the eggs were left to hatch. Eggs egg and then in a numbered were then either measured immediately or microscope placed in the As each stored in 70% isopropyl alcohol. Storage in iso- hatching cup hatching aquarium. hatched several the was propyl alcohol does not significantly alter egg egg days later, hatchling removed and in a numbered vial of 4% diameter (unpubl. data). placed To calculate effective diameter, we measured formalin. Eggs that did not hatch were dropped from The were the length and width of 20 eggs from each sam- subsequent analysis. hatchlings measured as above and similar using a Wild dissecting scope fitted with an regressions per- ple formed. ocular micrometer, calibrated to 0.01 mm. The (2.0 cm diameter, 1.5 cm 20 values were averaged to produce a mean egg Hatching cups size for the On the of high) were made from the bottom portion of a species. day hatching, vial secured with silicon to a hatchlings were collected with a turkey baster. plastic specimen small square of glass to make the cup sink. The Because live hatchlings are difficult to measure cups were placed in a half-filled 38-L aquarium, accurately, they were immediately preserved in with a heater and airstone. This hatching-cup 4% formalin and measured several days later. design proved successful in other experiments Three parameters were measured with each because it allows water to circulate over the eggs hatchling lying on its side: the total length of but keeps the eggs from blowing out of the the hatchling (TL), and the length and width cups. of the yolk sac. Typically, 20 hatchlings were measured to obtain a mean hatchling value for RESULTS the species, though in a few cases, particularly when were rare, fewer than 20 were hatchlings Mean hatchling length was positively corre- used [C. 13; "C. (Am.)" alfari 13; "C. festae lated with mean egg diameter in the 26 species (Am.)" rostratum 18; "C. (Ar.)" septemfasciatum examined (Fig. 1; r = 0.83; df = 24; P < 0.01). 16; "C. 19; "C. inter- (Th.)" bifasciatum (Th.)" Diameters of most of the eggs were 1.3-1.9 mm. medium13; "C. tuba 3; "C. sieboldii (Th.)" (Th.)" The extreme point at 2.4 mm was "C. (Th.)" Cl. maronii M. N. anomala 12; 16; altispinosa 11; tuba, the only known cichlasomine with such a P scalare We used a 10; 11]. only single spawning large egg (RMC, unpubl. data). The means for from a female for each to avoid single species hatchling yolk diameter (ye) also correlated pos- in the pseudo-replication interspecific analysis. itively with mean egg diameter in these species To determine the of importance collecting (ye = 1.00 X de - 0.11 mm; r = 0.97; df = 24; the hatchlings on the first day of hatch, we also P< 0.01). took of to see how daily samples hatchlings Hatchlings change size rapidly after hatching much their size changed from day to day. We (Fig. 2). Hatchling length increased 15% on the sampled a single spawn for each of two species, first day for midas cichlids and 13% for "C. "C. (Am.)" citrinellum and "C. (Th.)" bifascia- (Th.)" bifasciatum. Also, as expected, yolk di- tum, from the day of hatch until the stage of ameter declined as hatchling length increased. free-swimming. It is therefore important that hatchlings be ob- To test whether egg size explains offspring tained on the day of hatching (as was done in size, we regressed mean hatchling length on this study) for hatchling size comparisons; oth- mean egg size, and we correlated mean yolk di- erwise changes in hatchling size may cloud the ameter with mean egg size. If larger eggs are relationship between egg size and hatchling simply filled with more water, we would not ex- size. COLEMAN AND GALVANI-CICHLID OFFSPRING SIZE 211

7.0 5.U

15 6.0 4.8 - a 0 I 5.0 4.6 - ci 0 & 4.0 I| 4.4-

3.0 4.2- ? / ?~~~~~~~~~~

1 2.0 4 n a 1.2 1.4 1.6 1.8 2.0 2.2 2.4 1.4 1.5 1.6 Egg Diameter (mm) Egg Diameter (mm) Fig. 1. Mean hatchling length (TL) as a function Fig. 3. Mean hatchling length (TL) as a function of mean effective egg diameter (de) for 26 species of of mean effective egg diameter (de) for spawnings Neotropical cichlids (TL = 2.76 X de + 0.02 mm). from 17 female convict cichlids (TL = 2.62 X de + Each number representsone or more species, as fol- 0.49 mm). lows: 1 = N. anomala; 2 = P scalare, 3 = C. dimerus,4 = M. altispinosa and Cl. maronii; 5 = "C. (Am.)" lon- gimanus and H. multispinosa;6 = C. bimaculatum;7 = ter and between yolk diameter and egg diame- A. pulcher, 8 = C. pusillum; 9 = "C. (Ar)" nigrofascia- ter. For a rainbow cichlid clutch, hatchling tum; 10 = Cr lepidota;11 = "C. (Ar)" centrarchus,12 length increased with egg diameter (Fig. 4; r = = "C. (Ar)" spilurum; 13 = "C. (Ar)" septemfasciatum; 0.62; df = 14; P < 0.05), and yolk diameter in- 14 = "C. 15 = "C. mana- (Th.)" intermedium; (N.)" creased with egg diameter (r = 0.93; df = 14; P 16 = "C. 17 = "C. guense; (Am.)" citrinellum; (Th.)" < 0.01). the outlier at 18 "C. 19 = "C. Removing possible egg sieboldii; = (Th.)" bifasciatum; (Am.)" size of 1.27 did not 20 = "C. 21 = C. 22 = (Fig. 4) significantly change alfari; (Am.)" rostratum; festae, either result. The coefficients of variation "C. (Th.)" 23 = N. and 24 = (So- nicaraguense, nematopus; kal and for di- "C. (Th.)" tuba. Rohlf, 1981) egg diameter, yolk ameter, and hatchling length were 8.6%, 9.4%, and 5.8%, respectively. For a convict cichlid Intraspecific comparison of single spawnings clutch, the results were similar for hatchling from 17 convict cichlid females also revealed a length (r = 0.63; df = 12; P < 0.05) and yolk positive correlation between mean hatchling diameter (r = 0.90; df = 12; P < 0.01). The length and mean egg diameter (Fig. 3; r = 0.54; coefficients of variation for egg diameter, yolk df = 15; P < 0.05). The mean for hatchling yolk diameter and hatchling length were 5.5%, diameter was also significantly correlated with 6.1%, and 2.9%, respectively. egg diameter (r = 0.64; df = 15; P < 0.05). Within an individual clutch, we found positive DISCUSSION correlations between hatchling and egg diame- Our results indicate that a considerable source of hatchling-size variation derives from 7.0 - 1.9

4.5 -

! 4.2 6.0- I . a 1 3.9 ? ?0 ? ? l3 s fb u 3.6 - H .i 5.0 .m 3.3-

1.J 1 3 4 5 j.0 Day After Hatch 1.2 1.3 1.4 1.5 1.6 1.7 Egg Diameter (mm) Fig. 2. The change in hatchling length (solid lines) and yolk diameter (dashed lines) on the days Fig. 4. Hatchling length (TL) of individualhatch- after hatching. Circles represent "C. (Th.)" bifascia- lings as a function of effective egg diameter (de) with- tum, squares represent "C. (Am.)" citrinellum. Each in a single clutch of rainbow cichlids (TL = 1.09 X symbol representsthe averagefor 20 individualssam- de + 2.23 mm). A plot of yolk diameter (Ye) on effec- pled on that day, except for "C. (Th.)" bifasciatumon tive egg diameter (de) looks similar (Ye = 0.97 X de day 1 (n = 19), day 4 (n = 17), and day 5 (n = 10). - 0.05 mm). 212 COPEIA, 1998, NO. 1 differences in egg sizes. This relationship holds stated, that they measured live fry, not pre- across species of Neotropical cichlids, within served ones. Second, they measured the hatch- one of the species, and even within single lings at the free-swimming stage, when the yolk clutches of two species. This result suggests that sac was completely absorbed, several days later small differences in egg size can have significant in development than in our study. Third, they consequences for the life history of the hatch- used random samples of eggs from each of the lings. It also suggests that offspring size at hatch broods they examined, whereas we specifically is under the control of the female, to the extent selected a range of eggs of different sizes. Cal- that she can control egg size allocation, more culating the coefficient of variation in egg size than it is determined by historical constraints of for each of their broods with the data provided phylogeny, i.e., which particular species of cich- in their table 1, the CVs are in almost all cases lid the egg comes from. comparable to the CVs we obtained (range = Across species, differences in egg size ac- 6.9-19.3, mean = 11.0). As they pointed out, counted for 69% of the variation in hatchling the CVs they found for fry length were much total length and 94% of the variation in the size smaller (range = 1.3-6.1, mean = 2.7). We do of the yolk of hatchlings. Few interspecific stud- not know whether their lesser variation in size ies are available to compare with our findings. of free-swimming hatchlings was the result of Blaxter (1969) presented data on egg size and measuring live hatchlings or whether the varia- length of hatchlings for 14 species of oviparous tion they found in size of eggs truly disappeared fishes drawn from a wide range of fish taxa. The after the eggs hatched. smallest egg was that of a surgeonfish (diameter The positive association between the yolk size of 0.7 mm), whereas the largest was a dogfish of hatchlings and egg size indicates a metabolic (egg length of 65 mm). The data were heavily cost to producing larger eggs. Larger eggs are clustered toward small eggs, but even so, small not larger because they are filled with more wa- eggs clearly produced small hatchlings, and ter but because they contain more of the ma- large eggs produced large hatchlings. terials necessary to produce a larger hatchling. A number of researchers have exploited dif- Consequently, from a parental-investment per- ferences between the egg size of different pop- spective, making a large egg costs a female more ulations, for example, Atlantic herring Clupea per egg than making a small egg. Larger eggs harengus (Blaxter and Hempel, 1963), and dace entail a reduction in the number of eggs that Leuciscusleuciscus (Mann and Mills, 1985), or be- can be produced (Smith and Fretwell, 1974; RMC, tween different females, for example, orange- unpubl. data). Therefore, larger size must con- throat darter Etheostomaspectabile (Marsh, 1986), fer a compensatory advantage, such as the abil- striped bass Morone saxatilis (Eldridge et al., ity to escape predators, to exploit an inhospita- 1982), Atlantic cod Gadus morhua (Knutsen and ble environment, or to consume a broader array Tilseth, 1985), chum and coho salmon Onco- of foods. No studies to date have examined this rhynchus keta and 0. kisutch (Beacham et al., question at such a fine level of variation in the 1985), rainbow trout 0. mykiss (Springate and Cichlidae. Bromage, 1985), and Atlantic salmon Salmo sa- Our study shows that relatively tiny differ- lar (Thorpe et al., 1984), to examine hatchling ences in initial egg size may lead to differences size. All of these studies of intraspecific variation in size of hatchlings. Within a few weeks after found that larger eggs produced larger hatch- hatching, dramatic differences in the size of sib- lings, as we found with convict cichlids. lings are obvious (unpubl. data). Further, the Within a single clutch of rainbow cichlids, egg phenomenon of "growth depensation," namely size accounted for 38% of the variation in that through time the largest fish in a cohort hatchling total length and 87% of the variation get larger and the smaller fish lag further be- in yolk size of hatchlings. Similarly, for a clutch hind, widespread in fishes (Brown, 1957; Brett, of convict cichlids, we found 40% and 81%, re- 1979), is commonplace in cichlids (pers. obs.). spectively. Our intraclutch results differ from a The intriguing question remains: is it the initial previous report by Lagomarsino et al. (1988) differences in egg size that ultimately determine who detected only a weak correlation between the size of a fry in the hierarchy, or do these egg size and hatchling length in the midas cich- initial differences disappear and other forces lid, "C. (Am.)" citrinellum. They argued that generate subsequent variation? their correlation was so weak as to be biologi- cally insignificant. ACKNOWLEDGMENTS We suspect three possible reasons why Lago- marsino et al. (1988) found only a weak corre- We thank G.W. Barlow for use of his facilities, lation. First, it appears, though it is not explicitly financial support, commenting on the manu- COLEMAN AND GALVANI-CICHLID OFFSPRING SIZE 213 script, and discussions about this research. We KNUTSEN,G. M., AND S. TILSETH.1985. Growth, de- and success of Atlanticcod thank B. Lavery for supplying us with fish, eggs, velopment, feeding lar- vae Gadusmorhua related to size. Trans. Am. and knowledge about breeding cichlids; the egg members of the Pacific Coast Cichlid Associa- Fish. Soc. 114:507-511. S. 0. 1983. A revision of the South tion for their of RMC's re- KULLANDER, ongoing support American cichlid Cichlasoma(Teleostei: Cich- search on and the genus egg-size evolution; GuyJor- lidae). Swedish Museum of Natural History,Stock- dan Fund of the American Cichlid Association holm. for financial support. We particularly thank B. LAGOMARSINO,I., R. C. FRANCIS,AND G. W. BARLOW. Blair, R. and L. Bireley, and S. Sung for helping 1988. The lack of correlation between size of egg us to obtain the P scalare, "C. (Th.)" sieboldii, and size of hatchling in the midas cichlid, Cichla- and "C. (Th.)" tuba, respectively. The work was soma citrinellum.Copeia 1988:1086-1089. P. V. 1994. The cichlid Rev. ed. conducted under use protocol RO50- LOISELLE, aquarium. 0896 at the of California, Tetra-Press,Melle, Germany. University Berkeley. R. H. AND This research was in MANN, K., C. A. MILLS. 1985. Variation in supported part by grant the size of and larvae of the from the National Science gonads, eggs dace, BNS 9109852 Foun- Leuciscusleuciscus. Environ. Biol. Fishes 13:277-287. dation to G. W. Barlow and is num- publication MARSH,E. 1986. Effects of egg size on offspring fit- ber 1 from the Cichlid Egg Project. ness and maternal fecundity in the orangethroat darter, Etheostomaspectabile (Pisces: Percidae). Co- LITERATURE CITED peia 1986:18-30. REGAN,C. T. 1906. A revision of the fishes of the BEACHAM, T. D., F. C. WITHLER, AND R. B. MORLEY. South American cichlid genera Cichla,Chaetobran- 1985. Effect of egg size on incubation time and chus and Chaetobranchopsis,with notes on the genera alevin and fry size in chum salmon (Oncorhynchus of the AmericanCichlidae. Ann. Mag. Nat. Hist. 7: keta)and coho salmon (0. kisutch).Can. J. Zool. 63: 230-239. 847-850. . 1906-1908. Pisces. Biologia Central-America BLAXTER,J. H. S. 1969. Development:eggs and lar- 8:1-203. vae, p. 177-252. In: Fish physiology.Vol. 3. Repro- SMITH, C. C., AND S. D. FRETWELL.1974. The optimal duction and growth. W. S. Hoar and D. J. Randall balance between size and number of offspring.Am. (eds.). Academic Press, New York. Nat. 108:499-506. , ANDG. HEMPEL.1963. The influence of egg SOKAL,R. R., ANDF.J. ROHLF.1981. Biometry.2d ed. size on herring larvae (Clupea harengusL.) .J. Cons. W. H. Freeman,San Francisco,CA. Perm. Int. Explor.Mer. 28:211-240. SPRINGATE,J. R. C., ANDN. R. BROMAGE.1985. Effects BREDER,C. M., JR., ANDD. E. ROSEN.1966. Modes of of egg size on earlygrowth and survivalin rainbow reproduction in fishes. T.F.H. Publications, Nep- trout (Salmogairdneri Richardson). Aquaculture 47: tune City,NJ. 163-172. BRETT,J. R. 1979. Environmental factors and growth, STIASSNY, M. L. J. 1991. Phylogenetic intrarelation- p. 599-675. In: Fish physiology.Vol. 8. Bioenerget- ships of the family Cichlidae:an overview,p. 1-35. ics and growth.W. S. Hoar, D. J. Randall,and J. R. In: Cichlidfishes: behaviour, ecology and evolution. Brett (eds.). Academic Press, New York. M. H. A. Keenleyside (ed.). Chapman and Hall, BROWN,M. E. 1957. Experimentalstudies of growth, New York. p. 361-400. In: Physiologyof fishes. Vol. 1. M. E. THORPE,J. E., M. S. MILES, AND D. S. KEAY.1984. De- Brown (ed.). Academic Press, New York. velopmentalrate, fecundityand egg size in Atlantic COLEMAN,R. M. 1991. Measuring parental invest- salmon, Salmosalar L. Aquaculture43:289-305. ment in nonspherical eggs. Copeia 1991:1092- 1098. (RMC) DEPARTMENT OF INTEGRATIVE BIOLOGY, M. A. AND M. BOWERS. ELDRIDGE, B., J. WHIPPLE, J. UNIVERSITY OF CALIFORNIA, BERKELEY, CALI- 1982. and of bass, Mo- Bioenergetics growth striped FORNIA 94720-3140; AND (APG) NEW COL- rone saxatilis,embryos and larvae. Fish. Bull. 80: 461-474. LEGE, OXFORD UNIVERSITY, OXFORD, EN- OX1 3BN. E-mail: cole- FRYER,G., AND T. D. ILES. 1972. The cichlid fishes of GLAND, (RMC) Send the GreatLakes of Africa.T.F.H. Publications, Nep- [email protected]. reprint tune City,NJ. requests to RMC. Submitted: 9 Oct. 1996. Ac- KAMLER,E. 1992. Earlylife history of fish: an ener- cepted: 2 June 1997. Section editor: S. T. getics approach. Chapman and Hall, New York. Ross.