INTRODUCTION
Like many aquatic colonial invertebrates, members of the bryozoan family Cupuladriidae are able to produce new colonies by both sexual and asexual processes (Hughes & Jackson, Reference Hughes and Jackson1980; Winston, Reference Winston1988; O'Dea et al., Reference O'Dea, Herrera-Cubilla, Fortunato and Jackson2004). Sexual propagation occurs via the formation of a founding larva (McKinney & Jackson, Reference McKinney and Jackson1989) while asexual or clonal propagation occurs via colony breakage, autofragmentation or budding (O'Dea et al., Reference O'Dea, Jackson, Taylor and Rodríguez2008). Recruitment in some species is entirely sexual and in others it is almost entirely clonal, while the majority of species make use of the two modes of recruitment in varying proportions (Baluk & Radwanski, Reference Baluk and Radwanski1977; Winston, Reference Winston1988; Thomsen & Håkansson, Reference Thomsen and Håkansson1995; Håkansson & Thomsen, Reference Håkansson, Thomsen, Jackson, Lidgard and McKinney2001; O'Dea et al., Reference O'Dea, Herrera-Cubilla, Fortunato and Jackson2004, Reference O'Dea, Jackson, Taylor and Rodríguez2008). The modes of propagation are recorded in the form of calcified skeletons of colonies (O'Dea et al., Reference O'Dea, Herrera-Cubilla, Fortunato and Jackson2004). It is therefore possible to examine how life histories vary amongst living populations and measure directly the evolution of reproductive life histories through geological time using fossil assemblages (Thomsen & Håkansson, Reference Thomsen and Håkansson1995; Håkansson & Thomsen, Reference Håkansson, Thomsen, Jackson, Lidgard and McKinney2001; O'Dea et al., Reference O'Dea, Herrera-Cubilla, Fortunato and Jackson2004, Reference O'Dea, Jackson, Taylor and Rodríguez2008; O'Dea, Reference O'Dea2006).
Despite advances in our understanding of the many processes of cloning in cupuladriids (Marcus & Marcus Reference Marcus and Marcus1962; Winston, Reference Winston1988; Thomsen & Håkansson, Reference Thomsen and Håkansson1995; Håkansson & Thomsen, Reference Håkansson, Thomsen, Jackson, Lidgard and McKinney2001; O'Dea et al., Reference O'Dea, Herrera-Cubilla, Fortunato and Jackson2004, Reference O'Dea, Jackson, Taylor and Rodríguez2008; O'Dea, Reference O'Dea2006) little is known about several aspects of sexual reproduction including embryonic brooding. This dearth of information hinders the scholarly potential of life history studies in the Cupuladriidae because it is not known if increased cloning is reciprocated by a reduced investment in sexual reproduction, as has been previously proposed (Håkansson & Thomsen, Reference Håkansson, Thomsen, Jackson, Lidgard and McKinney2001).
Unlike many cheilostome bryozoans, cupuladriids do not brood embryos and larvae in specialized ovicells (McKinney & Jackson Reference McKinney and Jackson1989) nor do they appear to possess sexually dimorphic zooids as some other free-living bryozoans do (Cook & Chimonides, Reference Cook and Chimonides1978, Reference Cook and Chimonides1983; Chimonides & Cook, Reference Chimonides and Cook1981). To estimate investment in sexual reproduction it is therefore necessary to make qualitative and quantitative observations of embryos in living colonies (Jackson & Wertheimer, Reference Jackson, Wertheimer, Nielsen and Larwood1985, Herrera et al., Reference Herrera, Jackson, Hughes, Jara and Ramos1996).
Several important questions remain unanswered regarding sexual reproduction in cupuladriid bryozoans: (1) How much do species invest in gamete formation and sexual reproduction?; (2) Does investment in sexual reproduction decrease in species whose populations are dominated by clonal propagation?; (3) Are there patterns of embryonic brooding within cupuladriid colonies?; and (4) What is the process of embryonic brooding in colonies and how does it vary amongst taxa?
The last of these questions is addressed in a separate study (Ostrovsky et al., in press) while the aim of the present study is to begin to address the first two and to make observations pertinent to the third. To do so, we make collections of live colonies from populations of four cupuladriid species from the tropical eastern Pacific, make observations on brooding embryos in colonies, their frequency, colour, position and morphology, and their relationship with morphology and age of the colony. We examine the relationships between all these aspects of embryonic brooding and frequency of clonal propagation. And finally, we consider these results in an evolutionary context.
MATERIALS AND METHODS
Study organisms
Cupuladriid bryozoans are common members of the soft-bottomed sand fauna in tropical and sub-tropical shelf seas (Cadée, Reference Cadée1975; Winston, Reference Winston1988; Cook & Chimonides, Reference Cook and Chimonides1983, Reference Cook and Chimonides1994; O'Dea et al., Reference O'Dea, Herrera-Cubilla, Fortunato and Jackson2004). Larvae normally settle upon a grain of sand or other small particle (Cook, Reference Cook1963, Reference Cook1965; Lagaaij, Reference Lagaaij1963; O'Dea et al., Reference O'Dea, Jackson, Taylor and Rodríguez2008), metamorphose and grow incrementally through the sequential iteration of auto and vibracula zooids until the colony overhangs the particle. At this point the colony becomes free-living (Cook, Reference Cook1965). Vibracula zooids act as hinge and muscle for long tapered setae, which, at the colony's margin, curve outwards, and downwards to support the colony above the sediment surface. Movement of setae is used to remove depositing sediment and restrict epibiotic growth and enables colonies to ‘walk’ over the sediment surface and up through the sediment to the sea floor if they become buried (Cook, Reference Cook1963; Lagaaij, Reference Lagaaij1963; O'Dea, in preparation).
The prevalence of cloning amongst cupuladriids is correlated with a suit of morphological variables. Decreased cloning is observed in species whose colonies are small, heavily calcified and domed shaped, while conversely, elevated cloning is observed in species that produce large, thinly calcified and flat colonies (O'Dea et al., Reference O'Dea, Herrera-Cubilla, Fortunato and Jackson2004). Thus, colony morphology may be used as an adaptive strategy to control levels of clonal reproduction. In support of this idea, several species possess highly specialized morphologies that increase cloning by either enhancing the likelihood of fragmentation (e.g. peripheral fragmentation) or are designed specifically to undergo colony fission (e.g. colony budding and autofragmentation) (Marcus & Marcus, Reference Marcus and Marcus1962; Håkansson & Thomsen, Reference Håkansson, Thomsen, Jackson, Lidgard and McKinney2001; O'Dea et al., Reference O'Dea, Jackson, Taylor and Rodríguez2008).
Collections
Colonies of cupuladriids were collected from near Las Perlas Archipelago, in the Gulf of Panama, in the tropical eastern Pacific using dredge-sampling methods from the RV ‘Urraca’ (Figure 1). In total, 27 dredges were made ten of which contained living cupuladriid colonies at a variety of depths (Table 1). All dredge material was washed on board through a 2 mm sieve. Four species of two genera of cupuladriid bryozoans were collected. During their examination, living colonies were kept in open seawater tanks with water sourced from the Gulf of Panama.
Fig. 1. Area of study in the Gulf of Panama, tropical eastern Pacific. Numbers relate to samples presented in Table 1.
Table 1. Location of collections of four species of tropical eastern Pacific cupuladriid colonies with the abundance of living colonies and number and percentage of colonies observed with embryos in each site for each species and the entire cupuladriid assemblage.
Morphologies and demography of colonies
All living colonies were sorted into species and their median diameters measured. Diameter was used to calculate colony surface area by modelling the colony on a flat circle (πr2). Diameter is linearly related to the age of the colony (O'Dea & Jackson, Reference O'Dea and Jackson2002) while area is linearly related to number of zooids. No Pacific species produces highly domed colonies (O'Dea et al., Reference O'Dea, Herrera-Cubilla, Fortunato and Jackson2004) and thus the flat circle model is a reliable estimate of colony surface area.
Prevalence of embryos
Identified colonies were first separated into clonal or sexual as described in O'Dea et al. (Reference O'Dea, Herrera-Cubilla, Fortunato and Jackson2004, Reference O'Dea, Jackson, Taylor and Rodríguez2008) and the percentage of sexual colonies for each species' population estimated. Colonies were then separated into those with brooding embryos (termed ‘fertile’) and those with no visible brooding embryos. We found that in all species, early and medium-sized oocytes tended to be of a yellow colour and therefore liable to be concealed by the yellow or brown cuticle of the zooidal frontal wall and sometimes heavy algal epibiotic growth. It was therefore only possible to ensure that large mature oocytes and embryos brooded in internal sacs were counted. In Discoporella marcusorum the coloration of the frontal cuticle is so intense that the oocytes and embryos were best observed through the semitransparent calcified basal wall of the colony where their presence could easily be detected.
The numbers of visible embryos in each ‘fertile’ colony were counted. For each species, colony area data were used to determine the mean number of embryos per mm2.
RESULTS
Species occurrences and abundances
A total of 1040 living cupuladriid colonies were collected from the ten dredges (Table 1). Four species were collected: Cupuladria exfragminis, Discoporella marcusorum, D. cookae (Herrera et al., Reference Herrera-Cubilla, Dick, Sanner and Jackson2006, Reference Herrera-Cubilla, Dick, Sanner and Jackson2008) and Discoporella sp. nov. P1. The most abundant species' were D. marcusorum and D. cookae. Fewer colonies of C. exfragminis were found while Discoporella sp. nov. P1 was rare (Table 1).
Morphologies and demography of colonies
Mean colony diameters and mean colony areas of species and the percentage of sexually-produced colonies in populations are presented in Table 2. Data are consistent with results found by O'Dea et al. (Reference O'Dea, Herrera-Cubilla, Fortunato and Jackson2004). For example, species with large colonies produce the lowest percentage of sexually-produced colonies.
Table 2. Colony morphometrics, frequency and density of embryos, and percentage of clonal propagation in three species of tropical eastern Pacific cupuladriids. SD, standard deviation. * Note that only two colonies of Discoporella nov. sp. P1 were collected meaning data on morphologies are unlikely to be representative, but ** was corroborated by O'Dea et al. (Reference O'Dea, Jackson, Taylor and Rodríguez2008) who also found 100% sexual colonies with a much larger sample size.
Frequency histograms of colony size (area) show a multimodal distribution in the sampled population of C. exfragminis (Figure 2A), corroborating the multimodal demographic distribution discovered by O'Dea (Reference O'Dea2006) for the same species in the same region. The sampled population of D. cookae shows a bimodal size distribution (Figure 2B), while D. marcusorum is unimodal (Figure 2C).
Fig. 2. Frequency histograms of colonies of estimated colony area (mm2) in three species of tropical eastern Pacific cupuladriids.
Prevalence of embryos
In each of the four species at least one colony was found to be fertile. However, the percentage of fertile colonies varied considerably among species, ranging from 5% to 100% (Table 1). The frequency of embryos in colonies also varied greatly among species (Figure 3A, B). The maximum number of embryos observed in colonies of both D. marcusorum and D. cookae was around 60 (Table 2) but D. marcusorum had, on average, significantly greater numbers of embryos per colony (t = 7.139, df = 962, P < 0.001). This was also true for the density of embryos per mm2 of colony (t = 7.645, df = 962, P < 0.001), despite D. marcusorum having smaller zooids (Herrera-Cubilla et al., Reference Herrera-Cubilla, Dick, Sanner and Jackson2008). Cupuladria exfragminis had on average small numbers of embryos per colony (Table 2). Only two colonies of Discoporella sp. nov. P1 were found, and although both were fertile, the sample size was considered too small to conduct accurate statistical analyses.
Fig. 3. Embryo arrangement (top), number (middle) and density (bottom) in colonies of Cupuladria exfragminis (left), Discoporella cookae (centre) and Discoporella marcusorum (right). (A–C) Photographs with arrows labelling brooding embryos within colonies that illustrate the typical arrangement of embryos in each species (see text for details); (D–F) colony area versus total number of embryos observed within individual colonies. Arrows indicate size of smallest fertile colony; (G–I) colony area versus density of embryos (number of embryos per mm2) observed within individual colonies.
Plots of colony area against number of embryos observed in colonies of C. exfragminis, D. marcusorum and D. cookae show the demography of embryo brooding in the three populations at the time of sampling. The size of the smallest fertile colony was similar in both Discoporella species at approximately 10 mm2 (Figure 3E, F). In general, the most embryo-rich colonies occurred amongst the largest colonies in both D. cookae and D. marcusorum, but D. marcusorum showed a decline in the numbers of embryos in the very largest colonies (Figure 3E, F).
Figure 3G–I show how embryo density varied with colony size. Embryo density in the population of D. cookae revealed a very different pattern than absolute frequency, with density tending to decrease as colonies get larger (Figure 3H). In D. marcusorum density tracked absolute number, although the fall-off in embryo brooding in larger colonies is slightly exaggerated (Figure 3I).
Brooding and clonal propagation
Colonies of species that predominantly propagate clonally (C. exfragminis and D. cookae) were found to have fewer embryos than predominantly sexually-reproducing species (D. marcusorum) and all had many less than Discoporella sp. nov. P1 which has never been observed to clone (Figure 4A, B; Table 2). This pattern appears to be valid even if the poorly-sampled Discoporella sp. nov. P1 is removed.
Fig. 4. Relationship of per cent of sexually-produced colonies and mean number (A) and mean density (B) of embryos in colonies of four species of tropical eastern Pacific cupuladriids. Error bars represent 95% confidence intervals. Fitted lines are based on the least squares method.
Within each of the three species that are able to clone (C. exfragminis, D. cookae and D. marcusorum), clonally-produced colonies were found to have fewer embryos per colony and lower embryo densities than sexually-produced colonies (Table 2). This tendency was significant for C. exfragminis (t = 7.194, df = 72, P < 0.001 and t = 7.317, df = 72, P < 0.001 respectively) and D. marcusorum (t = 3.150, df = 519, P < 0.01 and t = 3.068, df = 519, P < 0.01 respectively), but non-significant for D. cookae (t = 0.525, df = 433, P = 0.600 and t = 0.3454, df = 433, P = 0.730 respectively).
Observations on arrangement of embryos within colonies
The position of brooding embryos in colonies appeared to vary considerably between species. The following tendencies were observed. In the few fertile colonies of C. exfragminis that were collected, embryos tended to be scattered throughout the colony somewhat randomly (Figure 3A), while in D. marcusorum embryos were predominantly located in zooids at the peripheral margins of colonies (Figure 3C). In contrast, embryos in D. cookae were most often observed in zooids in the central parts of colonies (Figure 3B). This was true in both cloned and sexually-produced colonies of this species.
DISCUSSION
This study aimed to better understand patterns of embryonic brooding within and between cupuladriid species, with particular focus on the relationship between frequency of embryo brooding and cloning. Although the study was restricted to a small number of species, we observed striking differences in brooding biology within and between species populations that may help clarify the nature of reproductive life history variation among cupuladriid bryozoans.
Embryo frequency and clonal propagation
Cupuladriids do not possess specialized brood chambers (ovicells) unlike many other cheilostome taxa (Ostrovsky et al., in press). In those ovicellate cheilostomes that have so far been examined, a negative relationship tends to exist between the amount of cloning and the frequency of ovicells suggesting a direct trade-off between investment in clonal and sexual modes of propagation. In a study of encrusting reef corals, Jackson & Wertheimer (Reference Jackson, Wertheimer, Nielsen and Larwood1985) showed that species that readily clone produce, in general, fewer embryos. Likewise, by collating data from fossil and Recent erect, encrusting and free-living bryozoans over millions of years Thomsen & Håkansson (Reference Thomsen and Håkansson1995) and Håkansson & Thomsen (Reference Håkansson, Thomsen, Jackson, Lidgard and McKinney2001) showed a strong negative relationship between frequency of clonal propagation and the numbers of ovicells. They suggested that as phyletic lineages invest more heavily in cloning through geological time they invest proportionally less energies into sexual reproduction. Essentially, they thought that energy for sexual reproduction is diverted into vegetative growth and enhancement of cloning by fragmentation. Indeed, there is good evidence that among bryozoan species there is often a clear trade-off in the relative investment into vegetative growth and sexual reproduction (Herrera et al., Reference Herrera, Jackson, Hughes, Jara and Ramos1996).
Our embryo survey data from four species of cupuladriids support this hypothesis, suggesting a negative relationship between frequency of embryos and the proportion of cloned colonies. Cupuladria exfragminis was found to brood the lowest overall number of embryos and is a species that propagates almost entirely clonally, has thinly-calcified colonies and undergoes autofragmentation (O'Dea, Reference O'Dea2006). The maximum mean number of embryos and maximum embryo densities were observed in colonies of Discoporella sp. nov. P1, despite only two colonies collected and analysed. This species also propagates entirely sexually and produces a thickly calcified skeleton to prevent fragmentation (O'Dea et al., Reference O'Dea, Herrera-Cubilla, Fortunato and Jackson2004, Reference O'Dea, Jackson, Taylor and Rodríguez2008). Within these two extremes of approach to propagation, the other two species analysed also fit this pattern: D. cookae has lower mean number of embryos and lower mean embryo densities than D. marcusorum and produces thinner, less calcified colonies and propagates considerably more of its colonies clonally.
Cheetham et al. (Reference Cheetham, Jackson and Sanner2001) also observed an increase in clonal reproduction over several millions of years in some Caribbean populations of Metrarabdotos but unlike the inference of Thomsen & Håkansson (Reference Thomsen and Håkansson1995) and Håkansson & Thomsen (Reference Håkansson, Thomsen, Jackson, Lidgard and McKinney2001), the authors concluded that the trends were purely an ecophenotypical response to the occurrence of upwelling that could support clonal propagation through increased vegetative growth of colonies. This raises the question of whether life history variation in embryo frequency and prevalence of cloning is controlled to a greater extent by environmental variations (ecophenotypically) or by the genotype.
It is clear that in cupuladriids frequency of cloning is controlled to a great extent by the genotype. This is demonstrated by the variety of distinct morphologies that are likely to have evolved in cupuladriids to either enhance or prevent fragmentation and clonal recruitment (O'Dea et al., Reference O'Dea, Jackson, Taylor and Rodríguez2008). As a result, strategies of life history in cupuladriids must be under selective pressure, and it is therefore necessary to consider the evolutionary benefits of one over the other at any particular time or environment (see O'Dea et al., Reference O'Dea, Jackson, Taylor and Rodríguez2008). Nonetheless, there does appear to be limited evidence that both the amount of cloning and the relative investment into embryonic brooding is ecophenotypically variable within species populations. For example, autofragmentation in C. exfragminis occurs most frequently during times of increased planktonic productivity during upwelling in the Gulf of Panama that results in a seasonal pattern of clonal recruitment (O'Dea, Reference O'Dea2006). Similarly, populations of C. biporosa have greater prevalence of cloning in regions of higher nutrients along the Caribbean coast of Panama (O'Dea et al., Reference O'Dea, Herrera-Cubilla, Fortunato and Jackson2004). Future study is however required to gauge the plasticity of investment into clonal and aclonal reproduction within species, and appreciate its potential importance in the evolution of cupuladriid life histories.
The relationship between the frequency of embryonic brooding and clonal propagation observed between cupuladriid species is mirrored by variation that occurs within species. Our data show that clonally-produced colonies have on average fewer embryos than sexually produced colonies. This could be explained in a number of ways. Firstly, it is known that immediately following fragmentation colonies increase marginal growth rate four-fold, probably in order to rapidly attain a ‘normal’ discoidal shape and repair fractured zooids (O'Dea, Reference O'Dea2006). Nutrients that are used for reparative growth are therefore unavailable for gamete formation and/or embryo brooding, and brooding frequency may consequently be reduced. Secondly, cloned colonies may have a different tempo of sexual reproduction than sexually-produced colonies, and our single sampling session failed to observe alternative seasonal investments in sex within clonal colonies. Thirdly, the process of fragmentation could alter colony morphology to such an extent that the ability of a colony to brood embryos is itself compromised. This may occur if fragmentation hinders the movement of energies around the colony that are needed to support embryonic brooding, or if oocytes are destroyed during fragmentation. And finally, because the zooids that comprise clones are likely to be old, having originated from the fragment of another colony, senescence at the zooidal level may reduce the ability of the colony as a whole to brood (Palumbi & Jackson, Reference Palumbi and Jackson1983).
Arrangement of embryos within colonies
Comparative observation of the arrangement of embryos in two species (Discoporella cookae and D. marcusorum) suggests that the location of oocytes in a colony is also related to the frequency of clonal propagation. The predominantly sexually-reproducing of the two species (D. marcusorum) tends to brood its embryos at the periphery of a colony while the predominantly clonal species (D. cookae) tends to brood in the central portion of the colony. If fragmentation is more likely to occur it may be advantageous for a colony to ensure that assets such as embryos are located in those regions less likely to be damaged. The central zone in colonies of D. cookae may be safer compared to marginal areas that may be more susceptible to fragmentation. Indeed, most embryos in cloned colonies of D. cookae were observed in the very central part of the colony, which normally remains intact during any subsequent fragmentation either because of greater calcification of older zooids or the preferential breakage along lines of previous fracture (O'Dea, Reference O'Dea2006). Additionally, the peripheries of regenerated fragments are always devoid of embryos. Discoporella marcusorum, on the other hand can take advantage of the lower risk of fragmentation and brood its embryos in the periphery of colonies. Brooding in the peripheral zone of colonies should be more advantageous given that zooids at the periphery are often more efficient feeders (Cook, Reference Cook1977; Cook & Chimonides, Reference Cook and Chimonides1978; McKinney & Jackson, Reference McKinney and Jackson1989; Okamura et al., Reference Okamura, Harmelin, Jackson, Jackson, Lidgard and McKinney2001).
Similar patterns are observed in a number of other free-living bryozoan groups, which are unrelated to the cupuladriids, suggesting a universal explanation. In members of the genus Selenaria reproductive zooids are located at the peripheral margins of colonies and members of this genus have never been observed to clone via fragmentation. Colonies of Selenaria therefore reach sexual maturity only when the colony reaches maximum size and this trait may only be feasible because the colony is not at risk of fragmentation. In D. marcusorum the ontogeny of sexual reproduction is not as strict as in Selenaria because brooding zooids were also observed in very small colonies (Figure 3), however low rates of fragmentation may permit the potentially advantageous delay in brooding until colonies reach a certain size.
Both Otionella tuberosa and Petasosella moderna regularly clone by fragmentation and have their reproductive zooids scattered across the colony (Cook & Chimonides, Reference Cook and Chimonides1985). This is similar to the situation seen in D. cookae and may be beneficial if peripherally brooded embryos are more prone to damage during fragmentation. Despite these parallel patterns, the location of reproductive zooids and its association with clonal propagation has yet to be assessed in most free-living bryozoans, but one exception to this general pattern is Otionellina squamosa that propagates clonally but also has peripheral zones of reproductive zooids (Cook & Chimonides, Reference Cook and Chimonides1984; Ostrovsky et al., in press).
Do colonies or zooids senesce?
Counting absolute number of embryos allows the exploration of the reproductive effort of the population as a whole whereas measuring the densities of embryos within colonies better reflects the ability of a unit area of bryozoan colony or a species population to reproduce sexually. Combining results from these two approaches reveals that patterns of reproductive investment amongst the species studied are highly variable (Figure 3D–I), and helps to reveal a species' life history strategy.
In D. cookae, a clear decline in embryo density occurs as colonies increase in size (Figure 3H) demonstrating that the increase in total number of embryos as colony size increases seen in Figure 3E is simply a product of larger colonies having more space to brood, and that in fact the most densely-brooding colonies are small. This may be in part because zooids of old fragments seem to be unable to support brooding in D. cookae. It is often observed in many cupuladriid species that zooids located in the central regions of colonies may die off or completely seal shut frontal openings by secondary calcification. This senescence of individual zooids may be unavoidable (Palumbi & Jackson, Reference Palumbi and Jackson1983; Jackson & Hughes, Reference Jackson and Hughes1985) or it may have functional significance, particularly to assist in the formation of expellent currents (Silén & Harmelin, Reference Silén and Harmelin1974; Okamura et al., Reference Okamura, Harmelin, Jackson, Jackson, Lidgard and McKinney2001) or perhaps more importantly in free-living species, to strengthen colonies.
In D. marcusorum, which on the whole reproduces sexually, the very oldest colonies also show a decline in embryonic density, although the pattern is unlike that seen in D. cookae (Figure 3H, I). One explanation is that colonies themselves are senescing and as they reach a determinate size/age (Herrera-Cubilla et al., Reference Herrera-Cubilla, Dick, Sanner and Jackson2008) reproductive ability or investment decreases, as occurs in solitary organisms. Although modular organisms are generally thought to persist by modular turnover (Jackson & Hughes, Reference Jackson and Hughes1985) many cupuladriids, including D. marcusorum, are perhaps more similar in their life histories to solitary organisms than other colonial invertebrates. They are free-living, reproduce sexually and have determinate growth and one can apply the same selective pressures that account for such processes in solitary organisms, such as increased predation, accident and disease probability to explain why reproductive effort per unit size should reduce ontogenetically (Kirkwood, Reference Kirkwood1977).
Discoporella cookae, which propagates mostly clonally, cannot be treated in the same way because it has indeterminate growth with fragments playing an essential role in the founding of new colonies. Given its capacity to clone through vegetative growth, senescence in D. cookae is probably restricted to the level of the zooid, meaning clones are potentially able to propagate eternally as is the case in most bryozoans (Palumbi & Jackson, Reference Palumbi and Jackson1983). Testing this could involve comparing the capacity to brood of newly budded zooids from old clones with those from new clones of species that normally reproduce sexually. Reduced capacity to brood in old clone zooids could suggest senescence.
Summary
This preliminary study explored the reproductive life histories in free-living cupuladriid bryozoans by conducting surveys on embryos in four species. The data, although limited in scope, highlight several important aspects of cupuladriid reproductive biology:
1. Frequency of embryos is highly variable both within and between species.
2. Investment in embryonic brooding appears to be negatively correlated with the incidence of clonal propagation amongst species.
3. Cloned colonies brood fewer embryos than sexually-produced colonies of the same species. This could be due to the diversion of energies away from gamete formation to growth, morphological disruption that physically limits brooding ability, or zooid senescence.
4. The arrangement of brooding embryos within colonies is variable and distinct amongst species, and may depend upon whether the species produces most of its colonies sexually or if it tends to clone.
5. Colonial senescence may cause the ontogenetic reduction of embryo density in species that preferentially propagate sexually. The life histories of such species may be more easily understood if they are considered as solitary rather than colonial organisms.
6. Future work is required to expand and clarify these data and hypotheses. More species need to be analysed, and it is crucial that spatial and temporal replication be incorporated to elucidate seasonal variations in brooding and the effects of different environments upon embryonic brooding.
ACKNOWLEDGEMENTS
Jeremy Jackson assisted with many aspects of the production of this paper. Rachel Collin kindly provided laboratory space, microscope and camera, and also helped collect material. Andrés Gómez, Etelyn González, Graciela Quijano, Marissa Quintero, Aileen Terrero and the crew of the RV ‘Urraca’ helped collect material. Amalia Herrera made comments to the manuscript. Anonymous referees gave very helpful comments on the manuscript.
The Secretaría Nacional de Ciencia, Tecnología e Innovación (SENACYT) funded the project. A. O'Dea was supported by the Tupper fellowship programme at STRI and National Science Foundation grant EAR03-45471. F. Rodríguez was supported by the Smithsonian Tropical Research Institute. A. Ostrovsky would like to thank Austrian Science Fund (FWF, grant P19337-B17) and Russian Foundation for Basic Research (RFBR, grant 07-04-00928a) for financial support. We would also like to thank Eldredge Bermingham and Harilaos Lessios from STRI for additional funds.
