Risk and the evolution of embryonic cell cycles

Evolution, 56(4), 2002, pp. 708–720 RISK AND THE EVOLUTION OF CELL-CYCLE DURATIONS OF EMBRYOS RICHARD R. STRATHMANN,1 JENNIFER M. STAVER,2 AND JENNIFER R. HOFFMAN3 Friday Harbor Laboratories and Department of Zoology, University of Washington, 620 University Road, Friday Harbor, Washington 98250 1 E-mail: rrstrath@u.washington.edu 2 E-mail: pjhohenlohe@yahoo.com 3 E-mail: hoffrau@u.washington.edu Abstract. Embryos at low risk evolve slower development rates. In seven independent evolutionary contrasts for marine invertebrates (two in asteroids, three in gastropods, one each in phoronids and brachiopods) the more protected embryos had longer cell cycles from first to second cleavage than less protected planktonic embryos. Protected embryos had longer cell cycles even when protected eggs were smaller than planktonic eggs. In an eighth contrast, among tunicates, the embryonic cell cycle was unrelated to brooding and nearly proportional to egg size, but the literature provides examples of especially slow development in some brooding tunicates. The faster development of planktonic embryos is consistent with published estimates of greater mortality rates for planktonic larvae than for embryos in broods or egg masses. Examples from the literature for annelids, arthropods, holothuroids, and chordates also demonstrated longer embryonic cell cycles for more protected embryos with no consistent effect of egg size on cell-cycle duration. Longer cell cycles presumably reduce the benefits of protecting offspring because of longer exposure to whatever hazards remain, but slow development may permit compensating benefits. Hypothesized benefits of longer cell cycles include less maternal investment in rate-limiting materials, more or different transcription, and correction of errors. Such trade-offs are independent of feeding and growth and are influenced by parental protection. Key words. Brood, egg mass, embryo, mortality, plankton, protection, safe harbor. Received May 10, 2001. A function of embryos is to restore a multicellular organism from a unicellular bottleneck in the life history (Grosberg and Strathmann 1998). Embryos must put the right cells in the right places at the correct times. For many embryos, their absolute speed in doing this is also important (Karr and Mittenthal 1992). Speed reduces a period of vulnerability. On its own, an animal embryo initially lacks capabilities for escape, defense, or capture of food that depend on specialized types of cells. An embryo’s internal nutrient stores make it a choice food for predators. An egg contains more than five times the organic material of a diatom of the same size (Strathmann and Vedder 1977). Embryos receive differing sorts of protection and are therefore exposed to different risks. Some are held on or in a parent’s body, enclosed in tough envelopes or voluminous gel, or sequestered in safe locations (Lee and Strathmann 1998). Planktonic embryos, though often possessing extraembryonic envelopes (Strathmann 1987) and sometimes possessing toxins (Lucas et al. 1979; Lindquist 1996; McClintock and Baker 1997), are among the minimally protected embryos. For a single, unprotected embryo, the plankton may be safer than the benthos. Indeed, a possible risk for embryos released as single eggs or zygotes is deposition on the bottom before they develop the ability to swim. Nevertheless, there appears to be greater risk for single embryos in the water column than for protected aggregates of embryos on the bottom. Mortality rates for small planktonic larvae are greater than those estimated for protected benthic embryos (Strathmann 1985; Rumrill 1990), and vulnerability of planktonic embryos to a variety of predators is similar to that for larval stages, and often greater (Pennington and Chia 1984; Pennington et al. 1986). Thus, both the magnitude of protective materials and studies of mortality indicate greater vulnerability for planktonic embryos than for embryos aggregated in broods, capsules, or gel masses. Although planktonic em- Accepted January 2, 2002. bryos have some defenses, we shall refer to them as unprotected, in contrast to embryos receiving protection in benthic aggregations. Most major groups of animals include planktonic embryos in some lineages and protected aggregations of embryos in other lineages. These repeated evolutionary divergences provide opportunities to test hypotheses on the effect of risk on the evolution of embryos. Here we examine the effect of protection on the evolution of development rates, specifically the duration of early embryonic cell cycles. As much as possible, we have dissociated development rate and protection from possibly confounding factors. We present observations on egg size, because protected eggs are often larger than planktonic eggs. Berrill (1935) reported slower development from early cleavages through hatching for tunicates with larger eggs, and a positive correlation between egg size and time to hatching has been noted for many kinds of animals (McLaren 1966; Shine 1978; Duarte and Alcaraz 1989). Larger eggs need not divide more slowly in initial cleavages, however, and in numerous embryos, a lengthening of cell cycles during development depends on an increase in nucleocytoplasmic ratio as cell volumes diminish through successive cleavages (Kane and Kimmel 1993 and references therein). We wished to see if risk affects evolution of rates of development independently from egg size. Life-history models of trade-offs between number and size of offspring include assumptions on the dependence or independence of egg size and rates of embryonic development. For example, the safeharbor hypothesis predicts evolution of larger eggs in species with more protected eggs (Shine 1978, 1989). It is also possible that increased egg size should result in selection for increased parental care (Nussbaum and Schultz 1989). The causal connections depend on effects of egg size on a complex of traits: embryonic development rates, stage at hatching, 708 q 2002 The Society for the Study of Evolution. All rights reserved. 709 RISK AND THE EVOLUTION OF DEVELOPMENT RATE posthatching growth rates, and mortality rates at both embryonic and posthatching stages. Because development rate affects the period of risk for embryos, distinguishing effects of egg size from effects of protection refines evolutionary hypotheses on correlated life-history traits. Development rates at stages with no growth are more conservative than growth rates and subject to different tradeoffs. In experimental selection for larval and pupal development times of Drosophila melanogaster, there was proportionately more change in the duration of the larval stage (with feeding and growth) than in the duration of embryonic or pupal stages (Chippendale et al. 1994, 1997; Prasad et al. 2001). We wished to isolate evolution of development rate from the trade-offs for duration of life-history stages that arise from foraging, assimilation of food, or habitat selection. We therefore examined embryos, which do not ingest food or swim. Although most marine embryos investigated are capable of active transport of small organic molecules dissolved in the sea (Shilling and Manahan 1990; Shilling and Bosch 1994; Jaeckle 1995), these materials support little or no growth until a functional gut develops (Hart and Strathmann 1995). We used cell cycle duration from first to second cleavage as a measure of development rate. This measure has numerous advantages. It is easy to see. It permits comparisons among disparate kinds of embryos because most embryos progress from two to four nuclei with a synchronous second division. Also, although in some animal embryos some cell lineages have shorter cell cycles than first to second cleavage, as in leech teloblasts (Bissen and Weisblat 1989), rat primitive streak (MacAuley et al. 1993), or early cleavages of axolotl embryos (Hara 1977), the cell cycle from first to second cleavage is among the shortest in the life of most animals and includes the shortest cell cycles recorded for animals. Also, for embryos developing at different temperatures (Detlaff 1964) or for closely related animals (Schneider et al. 1992), the duration of other early cell cycles is nearly proportional to duration of this first cell cycle; thus the first cell cycle is representative of rates of early development. Other embryonic stages are less comparable among species and thus less suited to broad comparisons of development rates. Gastrulation, for example, occurs at quite different cell numbers and by quite different processes in different animals. Time to hatching has been used as a measure of development rate, but hatching and swimming often occur at later stages of development in species with protected embryos or larger embryos (Strathmann 1987, Kiørboe and Sabatini 1994). Thus, time to hatching confounds differences in stage with differences in rate. Use of a single, early cell cycle limits the phenomena that may be rate limiting. The early embryonic cell cycles’ duration depends in part on times required for replication of the DNA (S phase) and mitosis (M phase). There is some early transcription of the zygotic genome, the amount varying among embryos (Davidson 1986), but for many embryos, including those planktonic embryos that have been investigated, there are no measurable gap phases (G1 and G2) in the early cell cycles (Murray and Hunt 1993). A major advantage of embryonic cell cycles as a measure of rate is that differing cycle durations point to differing rates at the mo- lecular level and may facilitate eventual genetic analyses of fitness trade-offs for fast or slow development. MATERIALS AND METHODS Study Organisms For independent contrasts, we used eight evolutionary divergences within five phyla (Table 1; see Figs. 1–4). The species were selected because of availability, not from advanced knowledge of their cell-cycle durations. We then augmented this sample with additional contrasts from published observations. Phylogenetic inferences for the asteroids vary (Blake 1987; Lafay et al. 1995; Wada et al. 1996; Knott and Wray 2000), but inferences agree on two independent divergences between brooding and free spawning in our sample of species (Fig. 1). One clade combines the brooder Leptasterias hexactis with three species of free-spawners with feeding larvae: Evasterias troschelii, Orthasterias koehleri, and Pisaster ochraceus, all in the order Forcipulatida. Another combines two species of free-spawners with large eggs, Henricia leviuscula and Henricia sp. 1 (undescribed species with gray armpits), with the brooding Henricia sp. 2 (undescribed species with small mottled adult), all in the order Spinulosida. Inferences from fossils suggest that the four orders of asteroids represented in our sample of species diverged during the Mesozoic (Lafay et al. 1995). Thus, there has been ample time for divergence of embryos that have remained planktonic, if divergence were to occur for other reasons than benthic protection. The order Paxillosida, which includes Luidia foliolata, is inferred to be the sister group of the other asteroids in our sample (Lafay et al. 1995; Wada et al. 1996). Crossaster papposus represents yet another Mesozoic divergence. The pelagic embryos of Pteraster tesselatus were included because brooding relatives, the aboral chamber (with broods in other pterasterids), and absence of larval brachiolar arms and attachment disk indicate a brooding ancestry (Strathmann 1974; McEdward 1995). The free-spawning gastropods were the lottiid limpet Tectura scutum and the trochid snail Calliostoma ligatum. Gastropods with benthic embryos in capsules surrounded by a gelatinous matrix were the caenogastropods Lacuna sp. (either L. vincta or L. variegata) and Littorina sitkana and the cephalaspidean opisthobranchs Haminaea vesicula and Haminaea callidegenita. The clades that include T. scutum and C. ligatum are inferred to have diverged early and separately from the lineage leading to the remaining gastropods (Haszprunar 1988; Ponder and Lindberg 1997), which implies at least one evolutionary divergence between planktonic embryos and embryos in capsules or egg masses in our sample of gastropod species (Fig. 2). These phylogenetic inferences plus the fossil record (Sepkoski 1982) indicate divergence between planktonic embryos and benthic protected embryos in the Paleozoic. Littorina scutulata has a small planktonic egg capsule containing several embryos. Functionally, protection of embryos in L. scutulata falls between benthic protection in a mass and planktonic development singly. Protected benthic eggs as a synapomorphy in the clade that includes the caenogastropods and opisthobranchs is the simplest hypothesis consistent with 710 RICHARD R. STRATHMANN ET AL. TABLE 1. Development times for cell cycle from first to second cleavage and for first cleavage to first visible rotation within egg envelope (designated by an asterisk) or first swimming. Temperature 108C Taxon Egg diameter (mm) 2 to 4 cell (h) Echinodermata, Asteroidea Luidia foliolata Evaserias troschelii Orthasterias koehleri Pisaster ochraceus Leptasterias hexactis Crossaster papposus Pteraster tesselatus Henricia leviuscula Henricia sp. (gray armpit) Henricia sp. (brooder) 152 162 148 178 995 796 1175 1342 1087 1144 1.93 1.5 2 1.57 6.55 2.03 1.68 1.98 1.73 2.93 7 3 7 8 5 5 6 8 6 9 Mollusca, Gastropoda Tectura scutum Calliostoma ligatum Littorina scutulata Littorina sitkana Lacuna vincta or variegata Haminaea vesicula Haminaea callidegenita 138 233 100 190 101 82 250 0.87 1.57 3 4.13 3.85 3.77 10.17 Phoronida Phoronis pallida Phoronis vancouverensis 63 128 Brachiopoda, Articulata Terebratalia transversa Terebratulina unguicula Chordata, Tunicata Oikopleura dioica Corella inflata Ascidia paratropa Boltenia villosa Temperature 148C 2 cell to swim or rotate (h) n 39.92 32.23 39.73 33.95 10 5 7 8 57.12 63.32 72.53 65.12 5 3 8 6 10 10 3 5 11 14 9 19.77 41.32* 232.02 160.07* 146.6* 131.9* 10 9 3 5 12 13 0.55 1.03 1.68 2.8 2.6 1.98 4.8 9 6 4 6 14 14 14 14.72 26.13* 132.3 10 5 3 106.42* 65.32* 13 13 1.25 2.88 6 7 11.03 6 0.68 2.02 7 9 6.37 160 154 1.03 1.43 10 13 14.25 22.43 8 13 0.63 0.8 14 9 76 135 169 156 0.33 0.87 1.08 0.92 4 15 13 12 8.87 36.42 45.8 30.62 7 15 17 13 0.23 0.58 0.7 0.59 9 12 12 10 n phylogenetic inferences (Haszprunar 1988; Ponder and Lindberg 1997). If this is true, then L. scutulata provides a second evolutionary divergence, a reacquisition of planktonic development, within this clade (Fig. 2). The fossil record is ambiguous but the divergence could be as ancient as early Tertiary (Reid 1996). The simplest interpretation of the distribution of traits and inferred phylogeny in the genus Littorina is a reversal from planktonic capsules to benthic gel masses in evolution within the genus Littorina (Reid 1996), with secondarily benthic development for L. sitkana (Fig. 2). Thus, the distribution of traits and inferred phylogeny indicates three independent evolutionary divergences between planktonic embryos and benthic protected embryos in our sample of gastropod species. Independence of the three other evolutionary divergences is clear because they are in three different phyla (Table 1, Figs. 3, 4). A more informative sample of tunicates (Fig. 4) would have compared the brooding Corella inflata with the co-occurring nonbrooding congener (Lambert et al. 1981) or compared the free-spawning solitary ascidians with brooding colonial ascidians in the same family or order, but we were unable to obtain these embryos. The phylogenetic inferences in Figure 4 are from Swalla et al. (2000). We also obtained times from first to second cleavage for species from other 2 to 4 cell (h) n 2 cell to swim or rotate (h) Embryo location Feeding larva or not pelagic pelagic pelagic pelagic brooded pelagic pelagic pelagic pelagic brooded feeding feeding feeding feeding no feed no feed no feed no feed no feed no feed pelagic pelagic pel. caps. gel gel gel gel no feed no feed feeding no feed feeding feeding no feed 7 pelagic brooded feeding feeding 10.65 17.5 13 4 pelagic brooded no feed no feed 6 23.25 28.92 18.75 4 15 10 8 pelagic brooded pelagic pelagic no no no no n feed feed feed feed phyla and classes to augment comparisons based on published studies. Experimental Methods and Analyses Most methods of obtaining eggs and sperm or of obtaining zygotes are described in Strathmann (1987). For observation, embryos were distributed in a sparse monolayer so that development rates would not be oxygen limited (Strathmann and Strathmann 1995; Cohen and Strathmann 1996). Embryos in gelatinous masses were separated by forcing the mass through a nylon mesh with a mesh opening larger than the diameter of the capsule surrounding an embryo. Buoyant embryos were floated up against a submerged transparent ceiling for observation. The ceiling roofed a tripod and had a low retaining edge, the whole unit cut from plastic culture well plates. Development is faster at higher temperatures. We maintained embryos within 60.38C of 108C or 148C by holding them in beakers in a heating-cooling water bath. Seasonal and geographic comparisons indicate that although temperature tolerances vary within species, there is usually little acclimation (Fujisawa 1995; Nomaguchi et al. 1997) or adaptation (Bosch et al. 1987; Hoegh-Guldberg and RISK AND THE EVOLUTION OF DEVELOPMENT RATE FIG. 1. Inferred relationships among asteroids in the sample of species and minimal estimates of origins of planktonic (P) and benthic brooded (B) embryonic development. The circles are approximately proportional to egg sizes and the times are from first to second cleavage, in hours, at 108C. Pearse 1995) of development rates. Selection at two temperatures also produced little change in duration of pupal stages of fruit flies (Partridge et al. 1994). Nevertheless, temperature acclimation of development rates has been reported (Landry 1975; Hart and McLaren 1978; Tester 1985). We therefore compared rates at temperatures that would avoid effects of acclimation or adaptation by being within the range of temperatures encountered in the reproductive season. We reduced chances of erroneously concluding slower development for protected embryos by comparing winter brooding asteroids to spring and summer free-spawners at the same temperature. We also used animals adapted to one region. All animals were collected near the San Juan Archipelago, Washington. All the species are native to the northeast Pacific with the possible exception of H. callidegenita. Although H. callidegenita was first described from the northeast Pacific (Gibson and Chia 1989), we suspected that it was recently introduced because it was recently discovered, was restricted to warmer bays in the region, and had recently colonized areas where it was previously absent. We included it in the table for comparison, but excluded it from data analysis because 108C and 148C may be below the temperatures to which it is adapted. Its inclusion would further support the conclusions. Times of first and second cleavage and in some cases times of first swimming were recorded with video cameras mounted on dissecting microscopes and a time-lapse video recorder. Often a batch of embryos was divided and recorded at two or three temperatures. The recorder was switched from one 711 FIG. 2. Inferred relationships among gastropods in the sample of species and minimal estimates of origins of planktonic (P) and benthic protected (B) embryonic development. The first derivation of planktonic development is single embryos. The second derivation of planktonic development is small groups of embryos in egg capsules. The benthic embryos are all in gelatinous masses. The circles are approximately proportional to egg sizes and the times are from first to second cleavage, in hours, at 108C. FIG. 3. Divergence between planktonic (P) and benthic brooded (B) embryonic development for two phoronids and two brachiopods. The circles are approximately proportional to egg sizes and the times are from first to second cleavage, in hours, at 108C. 712 RICHARD R. STRATHMANN ET AL. FIG. 4. Inferred relationships among tunicates in the sample of species and minimal estimates of origins of planktonic (P) and benthic brooded (B) embryonic development. The circles are approximately proportional to egg sizes and the times are from first to second cleavage, in hours, at 108C. camera to the next at 45-sec intervals. Thus, the recording at worst was accurate to 90 sec. Differences in appearance of cleavage furrows with orientation and from first to second cleavage also limited accuracy. First swimming was usually recorded as the time the embryo swam from its position, even though some embryos begin to rotate in place before swimming away. For gastropods with benthic development and for the planktonically developing C. ligatum, first swimming was recorded as time of first rotation within the egg capsule, even though hatching occurred much later. For L. scutulata, first swimming was recorded as hatching from the planktonic egg capsule, as for other embryos with planktonic development. (Data for C. ligatum were omitted from calculations that would be affected by this inconsistency, which occurred before criteria were standardized for benthic and pelagic embryos.) We report medians for times of development. For observations of several different batches of embryos for one species, we report the mean of the medians for each run. Intraspecific variation was much smaller than interspecific variation. For our comparisons of cell cycle durations, each species provides one datum at a given temperature. RESULTS Cell-Cycle Durations The cell cycle from first to second cleavage was shorter in planktonic embryos than in more protected embryos in seven of eight comparisons, although differences ranged widely (Table 1, Figs. 1–4). The two brooding seastars had longer cell cycles than did those with planktonic embryos, including planktonic embryos developing from large eggs into nonfeeding larvae (Table 1, Fig. 1). The egg size of the brooding Henricia sp. fell between the sizes for the two Henricia spp. with planktonic embryos, although both planktonic embryos had a shorter cell cycle from first to second cleavage. The embryos of the brooding L. hexactis had the longest cell cycle of any seastar in our sample, which includes three species with larger but planktonic eggs. The eggs of L. hexactis were larger than the eggs of other sampled species in its family (Asteriidae), but for asteroids with planktonic embryos, early cell-cycle durations were about the same for species with large eggs (800–1340 mm diameter, 1.73–2.03 h) and small eggs (150–180 mm, 1.5–2 h), and egg diameter and cell cycle duration were not correlated (Spearman rank correlation, r 5 0.036, n 5 8, P . 0.5). Thus, greater protection was associated with longer cell cycles in two evolutionary divergences, and this difference is not explained by differences in egg size. In the sample of gastropod species, all species with embryos aggregated into protective gel masses or capsules had longer cell cycles than those in the two species with single, planktonic embryos, even though two of the species with benthic gelatinous egg masses (Lacuna sp. and H. vesicula) had smaller eggs than the two species with planktonic embryos (T. scutum and C. ligatum; Table 1, Fig. 2). The planktonic embryos of L. scutulata (with a few embryos in a small egg capsule) are secondarily planktonic, and their cell cycle was shorter than those in the Lacuna and Haminaea species, which continue to deposit benthic egg masses. All three of these species have eggs of similar size (Table 1). This reversion to planktonic development in small packages provides a second evolutionary divergence in protection within our sample of gastropods. We consider the benthic gelatinous egg mass of L. sitkana to be a third evolutionary divergence because planktonic egg capsules are the inferred ancestral trait for the genus Littorina (Fig. 2). The cell cycle for L. sitkana was longer than that for L. scutulata, but its egg was also larger (Table 1). Altogether, the three inferred evolutionary divergences in protection consistently resulted in a longer cell cycle for the more protected embryos, and egg size did not account for two of the three evolutionary divergences. Egg diameter and cell cycle duration were not positively correlated. At both 108C and 148C the Spearman rank correlation was 20.14 (n 5 6, P . 0.5; H. callidegenita omitted). In the phoronids, the brooder had a longer cell-cycle duration (Table 1, Fig. 3). Possible effects of egg size on cell cycle duration could not be separated from effects of protection in this comparison because the brooder had a larger egg. In the brachiopods, the brooder had a longer cell-cycle duration, although the difference was not great (Table 1, Fig. 3). The brooder had a slightly smaller egg, and thus slower development of the brooder could not be attributed to larger eggs. In our sample of tunicates, there was no evidence of an effect of protection on cell-cycle duration. The early cell RISK AND THE EVOLUTION OF DEVELOPMENT RATE cycle of the brooder (C. inflata) was within the range for planktonic embryos, and egg diameter was correlated with cell-cycle duration for the four species sampled (Table 1, Fig. 4). The Spearman rank correlation for cell cycle and egg size was 1 at both 108C and 148C. (The correlation was just significantly different from zero at P 5 0.05, but the sample was small, with n 5 4.) Thus, in our sample of species, a longer cell cycle from first to second cleavage was associated with greater protection of embryos (seven of eight evolutionary divergences). Although cell-cycle durations changed disproportionately among species from 108C and 148C (Table 1), cell cycle durations were ranked the same within gastropods, phoronids, brachiopods, and tunicates at each of the two temperatures. We conclude that greater protection of embryos often results in slower development and does not result in faster development. It is also clear that larger eggs do not necessarily impose longer cell cycles. In the eight evolutionary divergences examined, two had smaller eggs with greater protection, three had larger eggs with greater protection, and the remainder had similar egg sizes. Effects of egg size may differ among groups, however. In the asteroids and gastropods, no effect of egg size on duration of early embryonic cell cycles was apparent. In the four tunicate species, however, cell cycle duration increased with egg size. Berrill (1935) reported this effect of egg size for early cell cycles and also for later development of tunicate embryos. Berrill’s observations have been generalized to other groups, but it is clear that larger eggs do not necessarily impose longer embryonic cell cycles. The evolutionary history of protection and risk can have a large effect on rate of embryonic development that is independent from any effect of egg size. Age and Stage at First Swimming Our sample of planktonic and protected embryos confirmed that planktonic embryos usually begin swimming at both a younger age and earlier stage than more protected embryos. The trend indicates that the earlier swimming with planktonic development results not only from faster development (shorter embryonic cell cycles) but also from an earlier stage at first swimming. In the phoronids, the free-spawner (P. pallida) began swimming as a blastula in less than 7 h at 148C (Table 1), whereas the brooder (P. vancouverensis) is reported to first swim as a blastula in vitro at 36–45 h at 13–158C, and stage and age at first swimming are even greater for embryos in intact broods, with larvae released at 350 h at about the stage when the larvae first feed (Zimmer 1964, 1987). In the brachiopods, the free-spawning T. transversa began swimming as a blastula, but the brooding T. unguicula is reported to release larvae when they are nearly ready to settle (Long 1964). Embryos of T. unguicula swam earlier than this stage in vitro, but even in vitro the free-spawner swam before the brooder at both 108C and 148C. The ratio S/C 5 (time from first cleavage to swimming)/(time from first to second cleavage) was 13% greater for the brooder at 108C and 29% greater at 148C. Greater S/C for the brooded embryos suggests that they began swimming at a later stage even in vitro. 713 In the molluscs, the free-spawning limpet T. scutum began swimming as a trochophore before any of the gastropod embryos in capsules or gel masses began rotating (Table 1). Time to first swimming was not recorded for the free-spawning gastropod C. ligatum, which first swims as a veliger larva, but time to rotation within the egg envelope was less for C. ligatum than time to first rotation for the gastropod embryos in gel masses (Table 1) and also less than time to first rotation for embryos of L. scutulata, which was about 107 h at 108C and 64 h at 148C. Hadfield and Strathmann (1990) report 5– 7 days to hatching for C. ligatum at 128C, which indicates earlier swimming than for the hatching veligers of L. scutulata. Also, with S as the time to swimming for the freespawning T. scutum and time to rotation for the other species, the ratio S/C was less than 30 for the two free-spawners and greater than 30 for the embryos in capsules or gel (calculated from Table 1 for both temperatures). Thus, the longer early cell-cycle duration for protected embryos does not entirely account for the longer time to production of ciliary currents. The protected embryos are either at a later stage of development or have a disproportionately greater lengthening of cell cycles before they produce ciliary currents. In the asteroids, the free-spawners hatch at the blastula or gastrula stage, whereas the brooders hatch after formation of the brachiolar arms (structures used in larval attachment; Strathmann 1987). The brooded young do not leave the mother until they are crawling seastars, after several weeks of development. The exception to later stage at hatching or release was the brooding ascidian C. inflata, with stage of first swimming (simple tadpole larva) and time to first swimming (Table 1) similar to that of free-spawning solitary ascidians. The ratio S/C was similar for the brooding C. inflata and the freespawning A. paratropa. (Both are phlebobranch ascidians.) The lack of an effect of brooding on age and stage of first swimming of C. inflata is not general for ascidians, however. The brooded embryos of colonial ascidians develop into larger tadpoles with more muscle cells after a more prolonged development (Berrill 1935; Cavey 1973; Satoh 1994; and see below). In summary, for the inferred evolutionary divergences in our sample of species, most of the planktonic embryos began locomotion at both an earlier stage and younger age than protected benthic embryos. This was true in our sample of embryos of asteroids, gastropods, phoronids, and brachiopods. Time to hatching combines rate of development and stage at hatching into a single measure. For comparisons among species, times to hatching are not a clear indication of rates of development. Long times to first swimming in animals with planktonic embryos could be a legacy of benthic development in their ancestry. This appears to be the case for L. scutulata, with protected benthic embryonic development in its ancestry (Fig. 2). It has a planktonic egg capsule but a much longer development to first swimming than did the free-spawning limpet T. scutum. Ancestral benthic development may also account for the long time to first swimming for the freespawning C. ligatum (J. M. Staver and R. R. Strathmann, unpubl. data). 714 RICHARD R. STRATHMANN ET AL. Other Divergences in Protection and Cell Cycles of Embryos Comparisons of published and additional new data for other evolutionary divergences in protection support the conclusion that cell-cycle durations are usually longer for protected embryos. Spiralia. A comparison of cleavage schedules of two nereid annelids, the brooding Platynereis massiliensis and planktonically developing P. dumerilii, demonstrated longer cell cycles for brooded than for planktonic embryos from first cleavage to 62 cells (Schneider et al. 1992). The ratios of cell-cycle durations were similar for equivalent cells in embryos of the two species, about 3.7 times longer for the brooder. In this intrageneric comparison the brooded eggs are more than twice the diameter of the planktonic eggs, so that egg size is confounded with protection. Also in annelids, the leech Helobdella triserialis has a cell cycle from first to second cleavage of about 4.5 h at 258C (Weisblat et al. 1980), much longer than the 0.4, 0.45, 0.70, and 0.73 h that we recorded at 188C for four annelids with planktonic development (Owenia fusiformis, Serpula columbiana, Sabellaria cementarium, Arctonoe fragilis). The leech, however, has an egg diameter of about 500 mm, whereas the planktonic eggs were all less than 80 mm. The contrast extends to other spiralians, and as with the gastropods that we observed, unprotected embryos have shorter cell cycles even when larger than protected embryos. Polyclad flatworms deposit encapsulated embryos whose early cell cycles are slow. Times from first to second cleavage greatly exceeded 2 h at 9–118C in two polyclad flatworms from the San Juan Islands, Pseudoceros canadensis (egg diameter 115–120 mm) and Kaburakia excelsa (150–160 mm; Shinn 1987). In contrast, this cell cycle was shorter at 108C in all planktonically developing spiralian embryos that we sampled from this region, including those with similar or larger egg diameters: 2.2 h in the bivalve Acila castrensis (126 mm), 1.6 h in the gastropod Calliostoma ligatum (233 mm), and 0.9 h in the gastropod Tectura scutum (138 mm). Crustacea. Barnacles and copepods both develop to nauplius larvae, but the embryos in barnacle broods are protected within the adult’s shell, whereas the smaller and more vulnerable copepods brood embryos externally or release zygotes that develop individually in the plankton. We recorded 3.45 h from two cells to four cells for the barnacle Balanus glandula at 148C (egg diameter 150 mm), and the time from two to eight cells is 1.25 h for the copepod Calanus finmarchicus at 13–158C (145 mm; Marshall and Orr 1953). Echinodermata, Holothuroidea. Among sea cucumbers in the subfamily Cucumariinae, the brooded embryos have cell cycles longer than 2 h and the planktonic embryos cell cycles less than 2 h at about 118C (McEuen 1987; Table 2). The large ranges for some estimates may be the result of imprecise estimates of stages, but the difference between brooders and free-spawners is great. All species live near the San Juan Islands in the northeast Pacific. McEuen’s Cucumaria fallax is now identified as C. pallida (Lambert 1997). The phylogeny inferred for the species in Table 2 implies two independent evolutionary divergences between planktonic and brooded development (Arndt et al. 1996). McEuen observed four TABLE 2. Times from two-cell to four-cell stage at 10.5–11.58C for brooded and pelagic embryos in the subfamily Cucumariinae (McEuen 1987). Species Pelagic Cucumaria miniata Cucumaria piperata Cucumaria pallida Brooded Cucumaria pseudocurata Pseudocnus lubricus Egg diameter (mm) Time (h) 520 530 500 1.75 1 0.75 1050 970 5 2–11 other species from the same order, all with planktonic embryos and nonfeeding larvae, with egg diameters of 330–630 mm and the times from two to four cells between l h and 1.75 h. Thus, for all dendrochirote sea cucumbers sampled from this region and observed at similar temperatures, the brooded embryos had longer cell cycles than the planktonic embryos. Development was also slower for the brooders through later (gastrula) stages. In this case, the brooded eggs are larger than the planktonic eggs, but greater differences in egg size had little, if any, effect on cell-cycle durations in our sample of asteroids (Table 1). Tunicata. For planktonically developing ascidian embryos, Berrill (1935) found slower development with larger eggs for stages from early cleavages to release of tadpole larvae. For some of the brooded ascidian embryos, differences in development times could be attributed to differences in egg size, but even in Berrill’s data, there appears to be an additional effect of brood protection on rates of development for numerous species of brooding ascidians, and these cases represent several evolutionary divergences in parental protection. In the Molgulidae, a nonbrooding species developed to tadpole in the time expected for planktonically developing ascidians of that egg size, whereas the two brooding species had a longer development to tadpole, although their tadpole larvae were described as ‘‘equivalent in structure’’ to those of the free-spawners (Berrill 1935). A longer time to gastrulation is indicated for two Ecteinascidia species contrasted with other brooding, colonial ascidians of similar egg size. For two related brooders, times to gastrulation were similar despite a secondarily reduced egg size in Botrylloides sp. and a larger egg in Botryllus sp. Thus, although our sample of ascidians was insufficient to demonstrate an effect of protection on rate of development, Berrill’s (1935) sample does indicate that brooding is often associated with a longer period for development to an equivalent stage and that the longer development exceeds that attributable to egg size. Cavey (1973) provides an additional example of extraordinarily long cell cycles in a brooded ascidian embryo from the same region as our study. The time from first to second cleavage is 26 h at 12–148C for the colonial ascidian Distaplia occidentalis, with an egg diameter of 400 mm. Cephalochordata and Vertebrata. In the cephalochordatevertebrate clade, the especially long early embryonic cell cycles are found among the brooders (Table 3). Times from two to four cells are much greater for marsupial and placental 715 RISK AND THE EVOLUTION OF DEVELOPMENT RATE TABLE 3. Times from two-cell to four-cell stage, temperature, and egg diameter (blastodisc for chicken) for cephalochordate and vertebrate embryos differing in protection. Species, protection Branchiostoma belcheri tsingtauenso cephalochordate (amphioxus), unguarded Danio rerio teleost (zebra fish), unguarded Nothobranchius guentheri teleost (annual fish), unguarded, diapause in drying puddles Austrofundulus myersi teleost (annual fish), unguarded, diapause in drying puddles Fundulus heteroclitus teleost (mummichog), unguarded, benthic Cyprinus carpio teleost (carp), unguarded, benthic Abudefduf saxatilis teleost (sergeant major), guarded Ambystoma mexicanum urodele (axolotl), unguarded Xenopus laevis anuran, unguarded Ascapus truei anuran (tailed frog), unguarded, on rocks in streams Gastrotheca riobambae anuran, brooded Gallus gallus bird, guarded Antechinus stuartii marsupial mammal, brooded Sminthopsus macroura marsupial mammal, brooded Mus musculus placental mammal (mouse), brooded Egg diameter (mm) 2 to 4 cells (h) Temperature (8C) 124a 0.25–0.4 23–26 675b 0.25 28.5 1000c 1.64d 25 1800e 1 25 2000f 0.75 20 1200–1400g 1200–1400g 660i 0.28h 1.58h 0.42 26 12 24 1800–2300j 1.75 21 1300–1400k 0.5l 20–21 4000m 2700–3600o 3000q ;2n ;18o 0.3–0.4r 16.5 17–18?p 36–38 242s 11t 35 250t 8 35 ,100u 15 37 Hirakow and Kajita 1990 Kimmel et al. 1995 Wourms 1967 d van Haarlem et al. 1981 e Wourms 1972 f Trinkaus 1967 g Albrecht et al. 1977 h Chaillou et al. 1991 i Shaw 1955 j Hara 1977 k Elinson 1987 l Masui and Wang 1998 m Wernz and Storm 1969 n Brown 1975 o del Pino and Escobar 1981 p del Pino and Loor-Vela 1990 q Pattern 1971 r Olsen 1942 r Eyal-Giladi and Kochav 1976 s Selwood 1980 t Selwood and Smith 1990 u Johnson 1981 a b c mammals than for less protected bird embryos, and a brooding frog has longer early embryonic cell cycles than amphibians with less protected embryos. Not all guarded embryos have especially long cell cycles. The early cell cycles of the chicken are not especially long and are much shorter than those of mammals. The early cell cycles of guarded embryos of Abudefduf saxatilis are not especially long compared to those of unguarded fish embryos deposited on the bottom. Overall, however, the pattern for the sample of species in Table 3 supports the hypothesis that longer embryonic cell cycles have evolved in association with greater protection of embryos. Neither temperature nor egg size can account for the association of longer early cell cycles with brooded embryos in the two separate evolutionary divergences in this sample of vertebrate species. The mammals’ protected embryos are hot and small but nevertheless slow developers. General pattern. The comparisons indicate that selection has maintained short early embryonic cell cycles in less protected embryos, in contrast to the longer cell cycles in more protected embryos. The generality of the result is supported by the inclusion of independent evolutionary contrasts between disparate kinds of embryos in distantly related animals. Planktonic embryos have converged on short early embryonic cell cycles, about 2 h or less at 108C and 1 h or less at 148C, but there is much variation in the ‘‘or less.’’ Correlates of cell cycle duration and time to first swimming for 716 RICHARD R. STRATHMANN ET AL. a larger sample of planktonic embryos will be reported elsewhere (J. M. Staver and R. R. Strathmann, unpubl. ms.). DISCUSSION Constraints and Trade-offs for Evolution of Cell-Cycle Duration Although protection reduces risk, all embryos are exposed to some risk from predators or external physical hazards. Why has selection not minimized the risk during this vulnerable stage by pushing all embryos’ early cell cycles to the fastest limit? One hypothesis is that selection for fast development is balanced by mutations that slow development, with the weaker selection on more protected embryos permitting accumulation of alleles that slow development, even though the mutations are weakly deleterious. With weaker selection for fast development, cell cycles become longer. We doubt, however, that selection on development rates of protected embryos is generally so weak that this is a sufficient explanation of observed differences in development rates. In addition to risk of death for embryos, there are other possible costs to slow development. Brooding restricts parental activities. For many animals, long development in a brood reduces the number of possible broods and lifetime reproductive output. Also, some maintenance costs of cells must continue, presumably whether cell cycles are slow or fast (Wolpert 1990; Leong and Manahan 1997). A more attractive hypothesis is that longer cell cycles in more protected embryos result from costs of fast development. There are numerous possibilities. The following list is not exhaustive, nor are the hypotheses mutually exclusive. Indeed, some of the processes invoked may be interdependent. Greater maternal investment in substrates that limit rates of DNA replication and mitosis is one possible cost of fast embryonic development. These could be materials such as tubulin for assembling mitotic spindles, mRNAs for cyclins or other materials regulating the cell cycle, mRNAs for histones, or materials for replication forks (Woodland 1980; Cavalier-Smith 1985; Davidson 1986; Karr and Mittenthal 1992; Murray and Hunt 1993). Long cell cycles permit more transcription or different transcription at earlier stages in more protected embryos (Davidson 1986; del Pino and Loor-Vela 1990). The increase in transcription in the midblastula transition of embryos of D. melanogaster, Xenopus laevis, and sea urchins contrasts with a significant increase in transcription in the two-cell stage of the more protected mouse embryo (Schultz 1993). Within anuran amphibians, nucleoli appear much earlier in the marsupial brooder Gastrotheca riobambae than in X. laevis (del Pino and Loor-Vela 1990). Although transcription begins before the midblastula transition in sea urchins, X. laevis, and D. melanogaster (Yasuda and Schubiger 1992) and even before first cleavage in sea urchins (Davidson 1986), there are differences in degree, with major zygotic gene activation in the blastula stage of these less protected embryos. Earlier transcription could permit earlier inductive signaling between cells (Davidson 1986) and reduce dependence on maternal materials during early cleavages. In particular, mitosis can lead to abortion of nascent tran- scripts; short cell cycles can limit the size of transcription units and thus the size of genes that are expressed (Shermoen and O’Farell 1991; Ruden and Jäckle 1995; Ohsugi et al. 1997). Protected embryos with longer cell cycles could have access to a greater variety of gene products at earlier stages of development. Conversely, incorporation of introns into genes normally expressed early in embryogenesis could shift expression to a later stage (with longer cell cycles) if the genes were already close to the upper size limit for transcription during the short early embryonic cell cycles. Slower cell cycles give more scope for differential lengthening of cell cycles. Lengthening of cell cycles in only some lineages could change cell fates by allowing expression of genes that are not yet expressed in lineages with shorter cell cycles. Changes in embryonic cell cycle durations could provide a mechanism for heterochrony. The greater shifts in timing among cell lineages that are permitted by longer cell cycles also give additional scope for signaling between cells or other cell interactions in embryos. As an example, in many mollusc and annelid embryos one of the macromeres is specified to form the mesentoblast, then functions as an organizer of developmental fates of other blastomeres, and is thereby responsible for departures from a radial to a bilateral arrangement of cell fates (van den Biggelaar et al. 1997). The primary trochoblasts develop cilia in the absence of intercellular signaling that alters this fate. In free-spawners, embryos not only develop faster, their early cleavages remain nearly synchronous through more cell cycles, and the primary trochoblasts differentiate autonomously and develop cilia in a radial arrangement. With nearly synchronous cleavages, the mesentoblast is formed relatively late. In contrast, the protected embryos not only have longer cell cycles in general, the cells that will form trochoblasts and apical ectoderm also have (in many clades) especially long cell cycles, and the mesentoblast and dorso-ventral body axis are specified when the embryo has fewer cells (van den Biggelaar et al. 1997). As a consequence, trochoblasts and other ectodermal cells can depart from a radial arrangement of cell fates and diverge in differentiation when there are fewer cells, as has occurred in encapsulated gastropod embryos. For example, fewer cells form prototrochal cilia and some instead form the head vesicle in pulmonate gastropods (van den Biggelaar 1993). Rapid development may increase risk of errors (Karr and Mittenthal 1992). As an example, cells with checkpoints are delayed in their cell cycles to ensure completion of DNA replication or events in mitosis before moving to the next stage (Hartwell and Weinert 1989; Paulovich et al. 1997; Epel et al. 1999). Embryos without checkpoints in early cell cycles may lose cells because of uncorrected errors from uncompleted events in DNA replication and mitosis (Sullivan et al. 1993; Anderson et al. 1997), whereas those that do have checkpoints may be afflicted with abnormal asynchronies among cell lineages when some cell cycles are delayed. Slower development could reduce the frequency of either of these potential errors in development. Embryonic cell cycles may be subject to pleiotropic effects other than those mentioned above, with slow development associated with other advantages. The possibility is suggested by the discovery of additional functions for proteins involved RISK AND THE EVOLUTION OF DEVELOPMENT RATE in regulation of the cell cycle, such as activating estrogen receptors or certain transcription functions (Jacks and Weinberg 1998). Also, a correlation between DNA content and fecundity, body size, and longevity is reported for amphibians, with low metabolic rates perhaps associated with larger cells, slow development, and energy economy (Nevo and Beiles 1991). However, we observed a shorter cell cycle from first to second cleavage in the brachiopod Terebratalia transversa (0.63 h at 148C) than in the mussel Mytilus sp. or scallop Chlamys hastata (0.75 h and 1.04 h at 148C), and brachiopods are slower in feeding currents and in metabolic rates (LaBarbera 1981; Thayer 1981). Living slow as a juvenile and adult does not necessarily entail slower embryonic development. Changes in cell-cycle controls during development (Edgar 1995) along with other changes in gene expression suggest possible means of reducing pleiotropic effects of selection for fast early cell cycles. Finally, genome size has been implicated in rates of development, but usually in association with growth, with the hypothesized limit being the correlation between genome size and cell size and hence the growth required for doubling cell size (Cavalier-Smith 1985). There are other hypotheses for longer cell cycles with larger genomes, however, such as late replication of satellite-rich heterochromatin in large genomes (MacGregor and Sessions 1986) or prolonged accumulation of cyclin or reduced expression of cyclin with larger genomes (Gregory 2001). Genome size is correlated with times to blastula or hatching in amphibians (Horner and MacGregor 1983; Jockusch 1997) and with prehatching development times of calanoid copepods (McLaren et al. 1988, 1989). There is not a simple relation between genome size and early cell cycles, however. In the crustacea, genome sizes of copepods can exceed those of cirripedes (Lécher et al. 1995), yet cirripedes have much slower embryonic development in early cell cycles (see above) and a much longer time to hatching. If larger genomes impose longer early cell cycles (which is as yet undemonstrated), then protected embryos could acquire larger genomes because of weaker selection against deleterious effects of junk or selfish DNA or because of benefits from larger genomes (Jockusch 1997; Gregory 2001). Adaptation of Growth Rates Constraints on rates of growth differ from constraints on rates of embryonic development (Cavalier-Smith 1985). Nevertheless, if there are trade-offs for cell-cycle durations that are independent of growth, as suggested by our comparisons among embryos, then these trade-offs might also affect rapidly growing cells and therefore be generally important in the evolution of rates of growth and development. Adaptation of Developmental Rates to Temperature Shorter cell cycles in embryos with less protection suggest that selection can alter cell-cycle durations, but there appears to be little adaptation of development rates to compensate for temperature. Low temperatures slow development and thereby extend the period of risk to embryos. Development of planktonic embryos in polar seas is very slow indeed (Bosch et al. 1987; Hoegh-Guldberg and Pearse 1995; Stanwell-Smith and Peck 1998). Because rates of development 717 do evolve in response to risk at a given temperature, why cannot selection for reduced risk produce fast development at low temperatures? It is possible that selection for fast development at low temperatures imposes greater costs, so that trade-offs between risk during long development and costs of fast development result in longer cell cycles at low temperatures. To test this hypothesis, it will be first necessary to identify what limits the durations and then compare these limiting features in minimally protected embryos adapted to different temperatures. Costs and Benefits of Parental Protection Protection of embryos results in evolutionary changes in development that alter the costs and benefits of parental protection. A protected embryo experiences a lower mortality rate per unit time, but its slower development exposes it to risk for a longer period at a given stage. The decrease in instantaneous mortality rate for protected embryos is therefore an inadequate measure of the gain in safety. Under a hypothesis of weak selection for fast development and accumulation of deleterious mutations, the initial benefits from evolution of greater embryonic protection could erode over time. Under the hypothesis of greater costs with shorter cell cycles, however, greater protection confers benefits associated with slower development, which may be diverse, and a gain in safety that is partially lost by a longer period of risk. Our comparisons suggest costs of short cell cycles or benefits of long ones that combine with risk to embryos to affect evolution of cell-cycle durations and thus rates of development. Such trade-offs are a complication omitted from previous life-history hypotheses. For example, hypotheses for correlations between parental protection and propagule size have included size-specific growth and mortality rates as factors affecting evolution of the size and stage at which propagules are released from protection but not evolution of slow development rates as a result of protection (Nussbaum and Schultz 1989; Shine 1989). Our comparisons indicate that cell-cycle durations and thus rates of development can be unaffected by egg size. This independence means that a correlation between egg size and development rate can be omitted as an assumption in models of life-history evolution. However, this independence does not exclude two other consequences of larger egg size: less growth to reach a given juvenile size and often a need for more cell divisions or differentiation to reach an optimal stage for release from protection. Evolutionary Transitions in Parental Protection Evolutionary changes in development rate may also bias evolutionary transitions in parental protection. Fast development of planktonic embryos does not preclude a change to greater embryonic protection in benthic broods or egg masses, but slowly developing benthic embryos would experience a longer period of risk if released into the plankton. This inferred bias has not entirely prevented reacquisition of planktonic embryonic development. An evolutionary transition from benthic protection to planktonic development has been inferred within the Pterasteridae and the Littorinidae, in the clades examined in this study (Strathmann 1974; 718 RICHARD R. STRATHMANN ET AL. McEdward 1995; Reid 1996). Although the barrier is not absolute, we infer that the effect of protection on cell cycle durations reduces the range of conditions in which planktonic development of individual embryos can evolve from development in a protected aggregation of embryos. ACKNOWLEDGMENTS National Science Foundation grants OCE-9633193 and IBN-0113603 and the Friday Harbor Laboratories of the University of Washington supported this research. We are grateful to D. O. Duggins, E. Edsinger-Gonzales, S. Y. Henderson, M. B. Hille, E. V. Iyengar, J. B. Marcus, D. K. Padilla, S. Santagata, C. Staude, M. F. Strathmann, M. von Dassow, and others at the Friday Harbor Laboratories for expert advice and help with obtaining and rearing a wide variety of embryos. R. A. Cameron, R. A. Cloney, E. H. Davidson, D. Epel, V. E. Foe, G. Freeman, B. W. Frost, G. M. Odell, J. A. M. van den Biggelaar, G. A. Wray, G. A. Wyngaard, R. L. Zimmer, and many others provided information and advice. LITERATURE CITED Albrecht, M. L., S. Fritzsche, H. Jaehnichen, and W. Steffens. 1977. Versuche zur Gewinnung lebensfaehiger Karpfenbrut unter Warmwasser bedingungen im Seitraum von Januar bis Mai. Z. Binnenfisch 24:371–376. Anderson, J. A., A. L. Lewellyn, and J. L. Maller. 1997. Ionizing radiation induces apoptosis and elevates cyclin A1-Cdk2 activity before but not after the midblastula transition in Xenopus. Mol. Biol. Cell 8:1195–1206. Arndt, A., C. Marquez, P. Lambert, and M. J. Smith. 1996. Molecular phylogeny of eastern Pacific sea cucumbers (Echinodermata: Holothuroidea) based on mitochondrial DNA sequence. Mol. Phylogenet. Evol. 6:425–437. Berrill, N. J. 1935. Studies in tunicate development. III. Differential retardation and acceleration. Philos. Trans. R. Soc. Lond. B 225: 255–326. Bissen, S. T., and D. A. Weisblat. 1989. The durations and compositions of cell cycles in embryos of the leech, Helobdella triserialis. Development 105:105–118. Blake, D. B. 1987. A classification and phylogeny of post-Palaeozoic sea stars (Asteroidea: Echinodermata). J. Nat. Hist. 21: 481–528. Bosch, I., K. A. Beauchamp, M. E. Steele, and J. S. Pearse. 1987. Development, metamorphosis, and seasonal abundance of embryos and larvae of the Antarctic sea urchin Sterechinus neumayeri. Biol. Bull. 173:126–135. Brown, H. A. 1975. Temperature and development of the tailed frog, Ascaphus truei. Comp. Biochem. Physiol. 50A:397–405. Cavalier-Smith, T., ed. 1985. The evolution of genome size. John Wiley and Sons, New York. Cavey, M. J. 1973. Differentiation of the larval caudal muscle cells in the compound ascidian Distaplia occidentalis. Ph.D. diss., University of Washington, Seattle, WA. Chaillou, C., B. Lefèbvre, D. Sjafei, and P. Roubaud. 1991. L’Embryon de carpe en segmentation: un modèle expérimental pour l’étude toxicologique du cycle cellulaire. Cybium 15: 303–314. Chippendale, A. K., D. T. Hoag, P. M. Service, and M. R. Rose. 1994. The evolution of development in Drosophila melanogaster selected for postponed senescence. Evolution 48:1880–1899. Chippendale, A. K., J. A. Alipaz, H.-W. Chen, and M. R. Rose. 1997. Experimental evolution of accelerated development in Drosophila. 1. Developmental sped and larval survival. Evolution 51:1536–1551. Cohen, C. S., and R. R. Strathmann. 1996. Embryos at the edge of tolerance: effects of environment and structure of egg masses on supply of oxygen to embryos. Biol. Bull. 190:8–15. Daniel, J. C., Jr. 1964. Early growth of rabbit trophoblast. Am. Nat. 98:85–98. Davidson, E. H. 1986. Gene activity in early development. 3rd ed. Academic Press, Orlando, FL. del Pino, E. M., and B. Escobar. 1981. Embryonic stages of Gastrotheca riobambae (Fowler) during maternal incubation and comparison of development with that of other egg-brooding hylid frogs. J. Morphol. 167:277–296. del Pino, E. M., and S. Loor-Vela. 1990. The pattern of early cleavage of the marsupial frog Gastrotheca riobambae. Development 110:781–789. Detlaff, T. A. 1964. Cell divisions, duration of interkinetic states, and differentiation in early stages of embryonic development. Adv. Morphog. 3:323–362. Duarte, C. M., and M. Alcaraz. 1989. To produce many small or few large eggs: a size-independent reproductive tactic of fish. Oecologia 80:401–404. Edgar, B. 1995. Diversification of cell cycle controls in developing embryos. Curr. Opin. Cell Biol. 7:815–824. Elinson, R. P. 1987. Change in developmental patterns: embryos of amphibians with large eggs. Pp. 1–21 in R. A. Raff and E. C. Raff, eds. Development as an evolutionary process. Alan R. Liss, New York. Epel, D., K. Hemela, M. Shick, and C. Patton. 1999. Development in the floating world: defenses of eggs and embryos against damage from UV radiation. Am. Zool. 39:271–278. Eyal-Giladi, H., and S. Kochav. 1976. From cleavage to primitive streak formation: a complementary normal table and a new look at the first stages of the development of the chick. I. General morphology. Dev. Biol. 49:321–337. Fujisawa, H. 1995. Variation in embryonic temperature sensitivity among groups of the sea urchin, Hemicentrotus pulcherimmus, which differ in their habitats. Zool. Sci. 12:583–589. Gibson, G. D., and F.-S. Chia. 1989. Description of a new species of Haminoea, Haminoea callidegenita (Mollusca: Opisthobranchia), with a comparison with two other Haminoea species found in the northeast Pacific. Can. J. Zool. 67:914–922. Gregory, T. R. 2001. Coincidence, coevolution, or causation? DNA content, cell size, and the C-value enigma. Biol. Rev. 76:65–101. Grosberg, R. K., and R. R. Strathmann. 1998. One cell, two cell, red cell, blue cell: the persistence of a unicellular stage in multicellular life histories. Trends Ecol. Evol. 13:112–116. Hadfield, M. G., and M. F. Strathmann. 1990. Heterostrophic shells and pelagic development in trochoideans: implications for classification, phylogeny and palaeoecology. J. Mol. Stud. 56: 239–256. Hara, K. 1977. The cleavage pattern of the axolotl egg studied by cinematography and cell counting. Wilhelm Roux’ Arch. Dev. Biol. 181:73–87. Hart, M. W., and R. R. Strathmann. 1995. Mechanisms and rates of suspension feeding. Pp. 193–221 in L. McEdward, ed. Ecology of marine invertebrate larvae. CRC Press, Boca Raton, FL. Hart, R. C., and I. A. McLaren. 1978. Temperature acclimation and other influences on embryonic duration in the copepod Pseudocalanus sp. Mar. Biol. 45:23–30. Hartwell, L. H., and T. A. Weinert. 1989. Checkpoints: controls that ensure the order of cell cycle events. Science 246:829–634. Haszprunar, G. 1988. On the origin and evolution of major gastropod groups, with special reference to the Streptoneura. J. Mol. Stud. 54:367–441. Hirakow, R., and N. Kajita. 1990. An electron microscopic study of the development of amphioxus, Branchiostoma belcheri tsingtauenso: cleavage. J. Morph. 203:331–344. Hoegh-Guldberg, O., and J. S. Pearse. 1995. Temperature, food availability, and the development of marine invertebrate larvae. Am. Zool. 35:415–425. Horner, H. A., and H. C. MacGregor. 1983. C value and cell volume: their significance in the evolution and development of amphibians. J. Cell Sci. 63:135–146. Jacks, T., and R. A. Weinberg. 1998. The expanding role of cell cycle regulators. Science 280:1035–1036. RISK AND THE EVOLUTION OF DEVELOPMENT RATE Jaeckle, W. B. 1995. Transport and metabolism of alanine and palmitic acid by field-collected larvae of Tedania ignis (Porifera, Demospongiae): estimated consequences of limited label translocation. Biol. Bull. 189:159–167. Jockusch, E. L. 1997. An evolutionary correlate of genome size change in plethodontid salamanders. Proc. R. Soc. Lond. B:597– 604. Johnson, M. H. 1981. The molecular and cellular basis of preimplantation mouse development. Biol. Rev. 56:463–498. Kane, D. A., and C. B. Kimmel. 1993. The zebrafish midblastula transition. Development 119:447–456. Karr, T. L., and J. E. Mittenthal. 1992. Adaptive mechanisms that accelerate embryonic development in Drosophila. Pp. 95–108 in J. E. Mittenthal and A. B. Baskin, eds. Principles of organization in organisms. Addison-Wesley, Reading, MA. Kimmel, C. B., W. W. Ballard, S. R. Kimmel, B. Ullmann, and T. F. Schilling. 1995. Stages of embryonic development of the zebrafish. Dev. Dyn. 203:253–310. Kiørboe, T., and M. Sabatini. 1994. Reproductive and life cycle strategies in egg-carrying cyclopoid and free-spawning calanoid copepods. J. Plankton Res. 16:1353–1366. Knott, K. E., and G. A. Wray. 2000. Controversy and consensus in asteroid systematics: new insights to ordinal and familial relationships. Am. Zool. 40:382–392. LaBarbera, M. 1981. Water flow patterns in and around three species of articulate brachiopods. J. Exp. Mar. Biol. Ecol. 55:185–206. Lafay, B., A. B. Smith, and R. Christen. 1995. A combined morphological and molecular approach to the phylogeny of asteroids (Asteroidea: Echinodermata). Syst. Biol. 44:190–208. Lambert, G., C. C. Lambert, and D. Abbot. 1981. Corella species in the American Pacific Northwest: distinction of C. inflata Huntsman, 1912 from C. willmeriana Herdman, 1898 (Ascidiacea, Phlebobranchia). Can. J. Zool. 59:1493–1504. Lambert, P. 1997. Sea cucumbers of British Columbia, southeast Alaska and Puget Sound. Univ. of British Columbia Press, Vancouver. Landry, M. R. 1975. Seasonal temperature effects and predicting development rates of marine copepod eggs. Limnol. Oceanogr. 20:434–440. Lécher, P. D. Defaye, and P. Noel. 1995. Chromosomes and nuclear DNA in Crustacea. Invert. Reprod. Dev. 27:85–114. Lee, C. E., and R. R. Strathmann. 1998. Scaling of gelatinous clutches: effects of siblings’ competition for oxygen on clutch size and parental investment per offspring. Am. Nat. 1551:293–310. Leong, P. K. K., and D. T. Manahan. 1997. Metabolic importance of Na1/K1-ATPase activity during sea urchin development. J. Exp. Biol. 200:2881–2892. Lindquist, N. 1996. Palatability of invertebrate larvae to corals and sea anemones. Mar. Biol. 126:745–755. Long, J. A. 1964. The embryology of three species representing three superfamilies of articulate Brachiopoda. Ph.D. diss., University of Washington, Seattle, WA. Lucas, J. S., R. J. Hart, M. E. Howden, and R. Salathe. 1979. Saponins in eggs and larvae of Acanthaster planci (L.) (Asteroidea) as chemical defenses against planktivorous fish. J. Exp. Mar. Biol. Ecol. 40:155–165. MacAuley, A. M., Z. Werb, and P. E. Mirkes. 1993. Characterization of the unusually rapid cell cycles during rat gastrulation. Development 117:873–883. MacGregor, H. C., and S. K. Sessions. 1986. The biological significance of variation in satellite DNA and heterochromatin in newts of the genus Triturus: an evolutionary perspective. Philos. Trans. R. Soc. Lond. B 312:243–259. Marshall, S. M., and A. P. Orr. 1953. Calanus finmarchicus: egg production and egg development in Tromsø Sound in spring. Acta Borealia A. Sci. 5:1–21. Masui, Y., and P. Wang. 1998. Cell cycle transition in early embryonic development of Xenopus laevis. Biol. Cell 990:537–548. McClintock, J. B., and B. J. Baker. 1997. Palatability and chemical defense of eggs, embryos, and larvae of shallow-water Antarctic marine invertebrates. Mar. Ecol. Prog. Ser. 154:121–131. McEdward, L. R. 1995. Evolution of pelagic direct development in 719 the starfish Pteraster tesselatus (Asteroidea: Velatida). Biol. J. Linn. Soc. 54:299–327. McEuen, F. S. 1987. Phylum Echinodermata, class Holothuroidea. Pp. 574–596 in M. F. Strathmann, ed. Reproduction and development of marine invertebrates of the northern Pacific Coast, Univ. of Washington Press, Seattle, WA. McLaren, I. A. 1966. Predicting development rate of copepod eggs. Biol. Bull. 131:457–469. McLaren, I. A., J.-M. Savigny, and C. J. Corkett. 1988. Body sizes, development rates, and genome sizes among Calanus species. Hydrobiologia 167/168:275–284. McLaren, I. A., J.-M. Savigny, and B. W. Frost. 1989. Evolutionary and ecological significance of genome sizes in the copepod genus Pseudocalanus. Can. J. Zool. 67:565–569. Murray, A., and T. Hunt. 1993. The cell cycle. Oxford Univ. Press, New York. Nevo, E., and A. Beiles. 1991. Genetic diversity and ecological heterogeneity in amphibian evolution. Copeia 1991:565–592. Nomaguchi, T. A., C. Nishijima, S. Minowa, M. Hashimoto, C. Haraguchi, S. Amemiya, and H. Fujisawa. 1997. Embryonic thermosensitivity of the ascidian, Ciona savignyi. Zool. Sci. 14: 511–515. Nussbaum, R. A., and D. L. Schultz. 1989. Coevolution of parental care and egg size. Am. Nat. 133:591–603. Ohsugi, K., D. M. Gardiner, and S. V. Bryant. 1997. Cell cycle length affects gene expression and pattern formation in limbs. Dev. Biol. 189:13–21. Olsen, M. W. 1942. Maturation, fertilization, and early cleavage in the hen’s egg. J. Morphol. 70:513–533. Partridge, L., B. Barrie, K. Fowler, and V. French. 1994. Thermal evolution of pre-adult life history traits in Drosophila melanogaster. J. Evol. Biol. 7:645–663. Patten, B. M. 1971. Early embryology of the chick. 5th ed. McGrawHill, New York. Paulovich, A. G., D. P. Toczyski, and L. H. Hartwell. 1997. When checkpoints fail. Cell 88:315–321. Pennington, J. T., and F. S. Chia. 1984. Morphological and behavioral defenses of trochophore larvae of Sabellaria cementarium (Polychaeta) against four planktonic predators. Biol. Bull. 167: 168–175. Pennington, J. T., S. S. Rumrill, and F. S. Chia. 1986. Stage-specific predation upon embryos and larvae of the Pacific sand dollar, Dendraster excentricus, by 11 species of common zooplanktonic predators. Bull. Mar. Sci. 39:234–240. Ponder, W. F., and D. R. Lindberg. 1997. Towards a phylogeny of gastropod molluscs: an analysis using morphological characters. Zool. J. Linn. Soc. Lond. 119:83–265. Prasad, N. G., M. Shakarad, D. Anitha, M. Rajmani, and A. Joshi. 2001. Correlated responses to selection for faster development and early reproduction in Drosophila: the evolution of larval traits. Evolution 55:1363–1372. Reid, D. G. 1996. Systematics and evolution of Littorina. Ray Society, London. Ruden, D. M., and H. Jäckle. 1995. Mitotic delay dependent survival identifies components of cell cycle control in the Drosophila blastoderm. Development 121:63–73. Rumrill, S. S. 1990. Natural mortality of marine larvae. Ophelia 32:163–198. Satoh, N. 1994. Developmental biology of ascidians. Cambridge Univ. Press, Cambridge, U.K. Schneider, S., A. Fischer, and A. W. C. Dorresteijn. 1992. A morphometric comparison of dissimilar early development in sibling species of Platynereis (Annelida, Polychaeta). Wilhelm Roux’ Arch. Dev. Biol. 201:243–256. Schultz, R. M. 1993. Regulation of zygotic gene activation in the mouse. BioEssays 15:531–538. Selwood, L. 1980. A timetable of embryonic development of the dasyurid marsupial Antechinus stuartii (Macleay). Aust. J. Zool. 28:649–668. Selwood, L., and D. Smith. 1990. Time-lapse analysis and normal stages of development of cleavage and blastocyst formation in the marsupials the brown antechinus and the stripe-faced dunnart. Mol. Reprod. Dev. 26:53–62. 720 RICHARD R. STRATHMANN ET AL. Sepkoski, J. J., Jr. 1982. A compendium of fossil marine families. Milw. Public Mus. Contrib. Biol. Geol. 51:1–125. Shaw, E. S. 1955. The embryology of the sergeant major, Abudefduf saxatilis: Copeia 1955:85–89. Shermoen, A. W., and P. H. O’Farell. 1991. Progression of the cell cycle through mitosis leads to abortion of nascent transcripts. Cell 67:303–310. Shilling, F. M., and I. Bosch. 1994. ‘Pre-feeding’ embryos of Antarctic and temperate echinoderms use dissolved organic material for growth and metabolic needs. Mar. Ecol. Prog. Ser. 109: 173–181. Shilling, F. M., and D. T. Manahan. 1990. Energetics of early development for the sea urchins Strongylocentrotus purpuratus and Lytechinus pictus and the crustacean Artemia sp. Mar. Biol. 106: 119–127. Shinn, G. L. 1987. Phylum Platyhelminthes with an emphasis on marine Turbellaria. Pp. 114–128 in M. F. Strathmann, ed. Reproduction and development of marine invertebrates of the northern Pacific Coast. Univ. of Washington Press, Seattle, WA. Shine, R. 1978. Propagule size and parental care: the ‘‘safe harbor’’ hypothesis. J. Theor. Biol. 75:417–424. ———. 1989. Alternative models for the evolution of offspring size. Am. Nat. 134:311–317. Stanwell-Smith, D., and L. S. Peck. 1998. Temperature and embryonic development in relation to spawning and field occurrence of larvae of three Antarctic echinoderms. Biol. Bull. 194: 44–52. Strathmann, M. F. 1987. Reproduction and development of marine invertebrates of the northern Pacific Coast. Univ. of Washington Press, Seattle, WA. Strathmann, R. R. 1974. Introduction to function and adaptation in echinoderm larvae. Thalassia Jugosl. 10:321–339. ———. 1985. Feeding and nonfeeding larval development and lifehistory evolution in marine invertebrates. Annu. Rev. Ecol. Syst. 16:339–361. Strathmann, R. R., and M. F. Strathmann. 1995. Oxygen supply and limits on aggregation of embryos. J. Mar. Biol. Assoc. UK 75:413–428. Strathmann, R. R., and K. Vedder. 1977. Size and organic content of eggs of echinoderms and other invertebrates as related to developmental strategies and egg eating. Mar. Biol. 39:305–309. Sullivan, W., D. R. Daily, P. Fogarty, K. J. Yook, and S. Pimpinelli. 1993. Delays in anaphase initiation occur in individual nuclei of the syncytial Drosophila embryo. Mol. Biol. Cell 4:885–896. Swalla, B. J., C. B. Cameron, L. S. Corley, and J. R. Garey. 2000. Urochordates are monophyletic within deuterostomes. Syst. Biol. 49:52–64. Tester, P. A. 1985. Effects of parental acclimation temperature and egg-incubation temperature on egg-hatching time in Acartia tonsa (Copepoda: Calanoida). Mar. Biol. 89:45–53. Thayer, C. W. 1981. Ecology of living brachiopods. Pp. 110–126 in J. T. Dutro Jr. and R. S. Boardman, eds. Lophophorates: notes for a short course. Univ. of Tennessee, Knoxville, TN. Trinkaus, J. P. 1967. Fundulus. Pp. 113–122 in F. H. Wilt and N. K. Wessells, eds. Methods in developmental biology. Thomas Y. Crowell, New York. van den Biggelaar, J. A. M. 1993. Cleavage pattern in embryos of Haliotis tuberculata (Archaeogastropoda) and gastropod phylogeny. J. Morphol. 216:121–139. van den Biggelaar, J. A. M., W. J. A. G. Dictus, and A. E. van Loon. 1997. Cleavage patterns, cell-lineages and cell specification are clues to phyletic lineages in Spiralia. Cell Dev. Biol. 8:367–378. Van Haarlem, R., R. van Wijk, and A. H. M. Fikkert. 1981. Analysis of the variability in cleavage times and the demonstration of a mitotic gradient in cleavage stage of Nothobranchius guentheri. Cell Tissue Kinet. 14:285–300. Wada, H., M. Komatsu, and N. Satoh. 1996. Mitochondrial rDNA phylogeny of the Asteroidea suggests the primitiveness of the Paxillosida. Mol. Phylog. Evol. 6:97–106. Weisblat, D. A., G. Harper, G. S. Stent, and R. T. Sawyer. 1980. Embryonic cell lineages in the nervous system of the glossiphoniid leech Helobdella triserialis. Dev. Biol. 76:58–78. Wernz, J. G., and R. M. Storm. 1969. Pre-hatching stages of the tailed frog, Ascaphus truei Stejneger. Herpetologica 25:86–93. Wolpert, L. 1990. The evolution of development. Biol. J. Linn. Soc. 39:109–124. Woodland, H. R. 1980. Histone synthesis during the development of Xenopus. FEBS Lett. 121:1–7. Wourms, J. P. 1967. Annual fishes. Pp. 123–137 in F. H. Wilt and N. K. Wessells, eds. Methods in developmental biology. Thomas Y. Crowell, New York. ———. 1972. Developmental biology of annual fishes. Stages in the normal development of Austrofundulus myersi Dahl. J. Exp. Zool. 182:143–168. Yasuda, G. K., and G. Schubiger. 1992. Temporal regulation in the early embryo: Is MBT too good to be true? Trends Genet. 8: 124–127. Zimmer, R. L. 1964. Reproductive biology and development of Phoronida. Ph.D. diss., University of Washington, Seattle, WA. ———. 1987. Phylum phoronida. Pp. 476–485 in M. F. Strathmann, ed. Reproduction and development of marine invertebrates of the northern Pacific Coast. Univ. of Washington Press, Seattle, WA. Corresponding Editor: R. Burton