Prevalence of Male Brood Care in Teleosts

98 Nature Vol. 276 2 November 1978

well understood as is the basis for the families with at least one parental species site sequestration would be necessary. differential. Most temperate-zone agri for which the spawning behaviour is Given a commitment to spawn defence,

culture crops are of the C3 type and known, 23 (60.5%) are characterised by selection could then operate to produce 13 3 5 possess a 6 C near -27%. Many tropical simultaneous expulsion of gametes - • As complex brood care, following the schema 25 grasses (including maize, sorghum and neither sex enjoys a temporal advantage outlined by Barlow • millets) and relatively few dicots are C4 in these cases, the desertion argument Within the framework of this model, and possess a 613C near -13%. The mean should predict a random distribution of biparental brood care is a derived state, its human tissue values of around -23%'. paternal, maternal and biparental care. evolution favoured by environmental reported by Lyon and Baxter may simply On the contrary, paternal care charac factors such as extreme scarcity of spawn 6 7 8 18 26 reflect a dietary carbon source which terises 16 of these families • , a ing sites • • or severe spawn preda 2 consists mainly of C3 plants. Thus, the significantly nonrandom proportion (x = tion 10·18·27, which make switching avail human samples in the above study were 19.00, d.f. = 2, P < 0.005). able energy from egg production to probably of Northern European deriva The desertion argument predicts behavioural activities advantageous for 25 tion. Also, I would expect subject vari paternal care in species in which oviposi females • Maternal care could be derived ability to be high, as the ratio will vary in a tion and ejaculation are well separated in from a biparental precursor, with male manner related to the dietary carbon. time. In the cave-spawning cichlids, the desertion explicable in terms of the Presumably, this would range from near best known group of fish of this type, modified parental investment model pro 1 -13% for diets derived from C4 plants to brood care ranges from biparental to fully posed by Dawkins and Carlisle • 4 8 11 near -27% for food from C3 plants. maternal, however • - • Maternal To conclude, the limiting nature of The real question is whether there is mouthbrooding cichlids have the most spawning sites and the differential bio secondary fractionation following the clearly sequential spawning pattern energetics of maturing eggs and sperm, 2 initial carboxylation of CO2 • There is known 1 • Yet the female commits herself rather than the mechanics of spawning some direct evidence for this in plant irrevocably to a parental role before the with internal fertilisation, seem to account 3 4 8 12 tissues and biochemical fractions • and eggs are even fertilised • • for the prevalence of male brood care 3 indirect evidence in animals • This addi Finally, maternal care is predicted among teleosts. tional discrimination, however, is small where the male spawns first. This pattern PA UL V. LOISELLE and seems most important in lipid occurs in the maternal mouthbrooding Department of Zoology and synthesis where further discrimination cichlid Sarotherodon macrochir13 and the Museum of Vertebrate Zoology, 13 5 against C occurs • Recent experimental paternal custodial goby Bathygobius University of California, 14 data suggest that derived organic matter soporator • In neither case has this Berkeley, California 94720 13 reflects dietary 8 C but tissues and bio behaviour led to a departure from the I. Dawkins, R. &Carlisle, T. R. Nature 262, 131-133 (1976). 13 2. Trivers, R. L. in Sexual Selection and the Descent of Man chemical fractions differ in 8 C values 7. genus-typical pattern of brood care. (Aldine, Chicago, 1972). As the organic matter of secondary The data suggest that factors other than 3. Frank, S. The Pictorial Encyclopedia of Fishes (Hamlyn, production will reflect the dietary values, a the sequence of gamete deposition may London, 1971 ). 4. van den Nieuwenhuizen, A. Tropical Aquarium Fishes. further complication is introduced. Each account for the prevalence of paternal Their Habits and Breeding Behavior (Constable, tissue and biochemical fraction may also care in fishes. I propose an alternative London, 1964 ). 5. Loiselle, P. V. Env. Biol. Fish. (in the press). have an isotopic 'memory'. This will be a model, based on (1) the limited availabil 6. Breder, C. M. & Rosen, D. E. Modes of Reproduction in 13 15 19 function of the 6 C of the carbon at the ity of suitable spawning sites - , which Fishes (Natural History Press, New York, 1966). 13 7. Loiselle, P. V. & Barlow, G. W. in Contrasts in Behavior time of synthesis, the 6 C values of makes sequestration of a breeding ter (Wiley Interscience, New York, 1978). subsequent foodstuffs, and the biochemi ritory selectively advantageous for many 8. Brichard, P. Rev. Zoo/. afr, 89, 871-888 (1976). 5 7 9. Sterba, G. Freshwater Fishes of the World (Studio Vista, cal turnover rate of the tissue or fraction. teleosts with demersal eggs - , and (2) the London, 1966). The persistence of this memory as well as differential energetic unit-cost of produc 10. Mera!, G. H. thesis, Univ. Calif., Berkeley (1973). 2 20 21 the magnitude of secondary dis ing ova and spermatozoa • • , which 11. Wickler, W. Aquar.- u. Terrar.-Z 19, 9-13 (1966). 12. Fryer, G. & Iles, T. D. The Cich/id Fishes of the Great Lakes crimination for various biochemical frac makes sequestration less profitable for of Africa (Oliver and Boyd, Edinburgh, 1972). tions must be established before results of females than males as a means of maxi 13. Wickler, W. Nature 208, 595-596 (1965). 6 14. Tavolga, W. N. Bull. Amer. Mus. nat. Hist. 104, 431-459 certain ecological and perhaps metabolic mising reproductive output. Extending (1954). 22 studies can be considered reliable. the reasoning of Maynard-Smith and of 15. Breder, C. M. Anal. Rec. 65,118 (1935). LARRY L. TIESZEN Dawkins and Carlisle 1 to include access to 16. Chapman, W. Mel. Copeia (3), 168-170 (1943). 17. Kodric-Brown, A. Evolution 31, 750-766 (1977). Department of Biology, any resource necessary for reproductive 18. McKaye, K. R. Ecology 58, 291-302 (1977). Augustana College, success, male desertion after spawning 19. Winn, H. E. Brit. J. Anim. Behav. 5, 25-28 (1965). 20. Bateman, A. J. Heredity 2, 349-368 (1948). Sioux Falls, South Dakota 57102 would be selectively advantageous only 21. Rohwer, S. Amer. Nat. 112, 429-440 (1978). 1. Lyon, T. D. B. & Baxter, M. S. Nature 273, 750 (1978). where there is high probability of securing 22. Maynard-Smith, J. Anim. Behav. 25, 1-9 (1977). 2. Grimwade, S. Nature 269,201 (1977). 23. Echelle, A. A. Copeia (I), 68-76 (1973). 3. Lerman. J.C. in Proc. 8th Int. Con{. Radio Carbon Dating, another breeding site. Where these are 24. Itzkowitz, M. Behaviour 68, 1-22 (1974). Hl 7 (1972); in Environmenlal and Biological Control of limiting, this eventuality is unlikely. 25. Barlow, G. W. Science 139, 851-852 (1963); Z. Tier Photosynthesis (ed. Marcelle, R.) 323 (Junk, Graven Persistence is thus favoured because an psycho/. 21, 99-123 (1964); Amer. ZooL 14, 9-34 hage, 1975). (1972). 4. Smith, B. N. & Jacobsen, B. S. Pl. Cell Physio/, Tokyo 17, immediate payoff to a territorial male is 26. Allen. G.D. The Anemone Fishes. Their Classification and 1089 (1976). the continued opportunity to spawn. Biology (T.F.H. Publications, Neptune City, 1972). 5. DeNiro, M. J. & Epstein, S. Science 197,261 (1977) 27. Loiselle, P. V, Biol. Behav. 2, 129-142 (1977). 6. Tieszen, L. L. et al. Oecologia {in the press). Defending an area against intrusion 7. DeNiro, M. J. & Epstein. S. Geochim. Cosmochim. Acta incidentally affords some protection to DAWKINS REPLIES-I agree with 42,495 (1978). 23 24 eggs within its boundaries ' • Thus in any Loiselle that the comparative data do not Prevalence of system characterised by nest disturbance support the small part of our paper which male brood care in teleosts or egg predation, a difference in fry was concerned with male brood care DAWKINS AND CARLISLE 1 have production between guarded and among teleosts. A similar conclusion is suggested that male brood care among unguarded sites would strongly favour reached in a recent exhaustive review of 1 teleosts arises from the female's ability to post-spawning territorial behaviour even the literature on paternal care • The pri spawn first and then desert her consort, if other factors reduce the likelihood of mary purpose of our paper, to expose the 2 placing him in Trivers' 'cruel bind' • The sequential polygyny. Its evolution would 'Concorde fallacy', remains valid. vulnerability of sperm to diffusion osten be facilitated because it would not require RICHARD DAWKINS sibly bars this option to males. A incorporation of new motor patterns or Animal Behaviour Research Group, comparative analysis of teleost reproduc endocrine control systems into the Department of Zoology, tive biology lends little support to this reproductive repertoire. Only prolonga University of Oxford, Oxford, UK argument, however. Of the 38 teleost tion of elements already involved in prior I. Ridley, M. Anim. Behav. 26, 904--932 (1978).