Supplementary methods – details on literature searches and data preparation

Literature search

Egg size data were predominantly obtained from a systematic search in Web of Science (WoS) using the keyword combinations: marine fish and “egg size” (title) and “egg size” (topic); reef fish and “egg size”, marine fish and “egg size” and “life history character*”; marine fish and “egg size” and fecundity; fish and egg size, and papers citing Thresher (1984). If the study included a reef fish family and had data on egg diameter, volume, or wet weight it was retained. Reef fish families were determined by a list compiled from Thresher (1984), Leis (1991), and Cowen andSponaugle (1997). All species included were crosschecked on FishBase (Froese andPauly 2003) to ensure that species included were associated withcoralreef or rockysubstrate. References were traced for articles with compiled datasets, particularly Thresher (1984), Ware (1975), and Einum and Fleming (2002). Post-hoc searches were done for specific families using the family name and egg size as keywords. Additional egg size data were obtained from three larval guides (Okiyama 1988; Moser 1996; Richards 2006). When egg size searches were exhausted, clutch size data were added post hoc for species with available egg size data. If the egg size paper did not give clutch size, FishBasewas searched using the search term “fecundity parameter” followed by a search on WoS.

Data preparation

Due to the lack of consistency in whether error values were reported and the form in which they were reported, intraspecies variation for egg and clutch size was not considered. If a study listed a range of values – in the form of several individuals of the same species or multiple measurements of the same species throughout a breeding season – the midpoint was retained. If a study represented egg size graphically, the values were approximated using the axis units and the mean value was calculated. Several families had species with ellipsoidal eggs (Apogonidae, Chaetondontidae, Gobiesocidae, Gobiidae, Labridae, Malacanthidae, Ostraciidae, Pomacentridae, Scaridae, Synodontidae), given in papers as a length and width and these dimensions were retained.

If clutch size was reported as a range or across seasons, the midpoint was retained. For two demersal species (Parablenniusruber and Rhinecanthusaculeatus), the diameter of a single egg mass was provided and we extrapolated the number of eggs based on clutch and egg volume. Papers listing the population mean egg production or fecundity as a volume of eggs were excluded from the database because no simple conversion among units was possible. The majority of clutch sizes obtained for demersal spawners were collected experimentally in the original studies to ensure that the clutch size was per female per spawning and not an aggregate of multiple females laying eggs.

The maximum adult standard length for each species was obtained from FishBase. If total length (i.e. the maximum length from the mouth and including the full caudal fin) values were given, they were converted to standard lengths (mouth to caudal peduncle) using the length-length parameters, which provide species-specific variables for the general equation where total length is known and standard length unknown. Two species – Apogonhungi and Archamialineolata – did not have any length value listed in FishBase and therefore the standard lengths were obtained from the original source (see ESM Table S6). Several species in FishBase werelisted with only total lengths without listing any length-length parameters. In these cases, the mean a andb values for the genus were calculated and used as the approximate parameters. Standard length was chosen over mass for adult size because there are more length data available and because length, not mass, is thecommon variable in reproductive theory.

Phylogenetics

To create a tree including the majority of our focal species, we used elements of acanthomorph phylogenies (Li et al. 2009; Near et al. 2013) as backbones and augmented with addition sequence data (as in Riginos et al. 2014). Specifically, we parsedrhodopsin (Rhod) sequences (from Li et al. 2009) to retain focal fish families with species represented in our data and to preserve selected fossil dated calibration points (from Santini et al. 2009, nodes 24, 33, 34, 36, 37, 38, 42, and 44 from their Table 2). We added additional Rhod sequences representing focal families as available from NCBI. NCBI and the Barcode of Life Database (BOLD) were also queried for cytochrome oxidase I (COI) sequences for all focal species and for each species with a Rhod sequence. Rhod and COI sequences were concatenated with missing loci treated as missing data. A Bayesian search was conducted in MrBayes (ver. 3.2, RonquistandHuelsenbeck 2003), with four partitions representing the two genes and treating third positions separately from first and second positions for each gene. Taxonomic families and clades receiving ≥ 90% support on an acanthomorphsupertree (Near et al. 2013) were constrained to be monophyletic in our search. A search was conducted using 10 million steps (sufficient for convergence in the two chains: s.d. < 0.01) and a burnin of 2.5 million steps under a GTR + G + I model of evolution for each partition. The resultant consensus tree was transformed to a chronogram using non-parametric rate smoothing with penalized likelihood with the truncated Newton method (r8s ver. 1.7, Sanderson 2002) using fossil calibration points (Sanderson 2009).

References

Cowen RK, Sponaugle S (1997)Early life history and recruitment in fish populations. Chapman & Hall, New York.

Einum S, Fleming IA(2002 )Does within-population variation in fish egg size reflect maternal influences on optimal values? Am Nat 160:756–765

Froese R, Pauly D(2003)FishBase.

Leis JM (1991)The pelagic stage of reef fishes: the larval biology of coral reef fishes. In: Sale PF (ed.)Theecology of fishes on coral reefs. Academic Press,San Diego, pp 183–230

Li B, Dettaï A, Cruaud C, Couloux A, Desoutter-Meniger M, Lecointre G (2009) RNF213, a new nuclear marker for acanthomorph phylogeny. MolPhylogenetEvol50:345–363

Moser HG (ed.)(1996) The early stages of fishes in the California current region. California cooperative oceanic fisheries investigations,CalCOFI Atlas No 33

Near TJ, Dornberg A, Eytan RI, Keck BP, Smith WL, Kuhn KL, Moore JA, Price SA, Burbrink FT, Friedman M, Wainwright PC(2013) Phylogeny and tempo of diversification in the superradiation of spiny-rayed fishes. ProcNatlAcad SciU S A110:12738–12743

Okiyama M (ed.)(1988)Anatlas of the early stage fishes in Japan. Tokai University Press, Tokyo

Richards WJ (ed.)(2006) Early stages of Atlantic fishes: an identification guide for the Western Central North Atlantic, Vol 2. Taylor & Francis, United Kingdom

Riginos C, BuckleyYM, Blomberg SP,Treml EA(2014) Dispersal capacity predicts both population genetic structure and species richness in reef fishes. Am Nat 184:52–64

Ronquist F, Huelsenbeck JP (2003) MRBAYES 3: Bayesian phylogenetic inference under mixed models. Bioinformatics 19:1572–1574

Sanderson MJ(2002) Estimating absolute rates of molecular evolution and divergence times: a penalized likelihood approach.MolBiolEvol 19:101–109

Santini F, Harmon LJ,Carnevale G,Alfaro ME (2009)Did genome duplication drive the origin of teleosts? A comparative study of diversification in ray-finned fishes.BMC EvolBiol 9:194

Thresher RE (1984)Reproduction in reef fishes.T.F.H. Neptune City, New York

Ware DM (1975) Relation between egg size, growth, and natural mortality of larval fish.J Fish Res Board Can 32:2503–2512

Page 1