INTRODUCTION
Sponge reproduction has been a topic of study over several decades (e.g. Delage, Reference Delage1892; Meewis, Reference Meewis1936; Leveaux, Reference Leveaux1941; Lévi, Reference Lévi1956; Bergquist, Reference Bergquist, Lévi and Boury-Esnault1979; Reiswig, Reference Reiswig1983; Amano, Reference Amano1986; Sarà, Reference Sarà, Adiyodi and Adiyodi1993; Mariani et al., Reference Mariani, Uriz and Turon2000, Reference Mariani, Uriz, Turon and Alcoverro2006; Ereskovsky, Reference Ereskovsky2005; de Caralt et al., Reference DeCaralt, Uriz, Ereskovsky and Wijffels2007; Maldonado & Riesgo, Reference Maldonado and Riesgo2009) but there are still reproductive processes that remain to be investigated, or are unconfirmed (Ereskovsky, Reference Ereskovsky2010). Some sponge species are gonochoristic; however, most sponges are hermaphroditic, and this is also true of the marine haplosclerids (Bergquist, Reference Bergquist1978; Ayling, Reference Ayling1980; Ilan & Loya, Reference Ilan and Loya1990a; Sarà, Reference Sarà, Adiyodi and Adiyodi1993; Maldonado & Riesgo, Reference Maldonado and Riesgo2009). Viviparous and oviparous reproductive strategies are found in sponges (Bergquist, Reference Bergquist1978). These traits were used to separate the marine Haplosclerida into two suborders: Haplosclerina (viviparous) and Petrosina (oviparous) (Lévi, Reference Lévi1956; Maldonado & Bergquist, Reference Maldonado, Bergquist, Young, Sewel and Rice2002). Recently, molecular systematic approaches have indicated that both suborders are polyphyletic, suggesting that these traits have been acquired and lost over time in the marine haplosclerids (e.g. Borchiellini et al., Reference Borchiellini, Chombard, Manuel, Alivon, Vacelet and Boury-Esnault2004; Raleigh et al., Reference Raleigh, Redmond, Delahan, Torpey, van Soest, Kelly and McCormack2007; Redmond et al., Reference Redmond, van Soest, Kelly, Raleigh, Travers and McCormack2007, Reference Redmond, Raleigh, van Soest, Kelly, Travers, Bradshaw, Vartia, Stephens and McCormack2011). The adult skeleton in haplosclerids forms the basis of the current classification, but is particularly simple leading to many problems in taxonomy and classification in the group. Molecular data suggests that some species are misplaced in the current classification (Raleigh et al., Reference Raleigh, Redmond, Delahan, Torpey, van Soest, Kelly and McCormack2007; Redmond et al., Reference Redmond, Raleigh, van Soest, Kelly, Travers, Bradshaw, Vartia, Stephens and McCormack2011) and it is also likely that species have been misidentified and assigned to the wrong family or genus. A review of reproductive patterns across the group is now highly recommended in parallel with classical morphological and molecular approaches given the new indications of phylogenetic relationships in this group.
Haplosclerid reproduction generally takes place during the warmer summer months. However, spawning often takes place during the winter months for oviparous haplosclerids such as Petrosia ficiformis (Maldonado & Riesgo, Reference Maldonado and Riesgo2009) and some viviparous haplosclerids such as Xestospongia testudinaria and Chalinula sp. produce sperm year round (Fromont, Reference Fromont, Choat, Barnes and Borowitzka1988; Ilan & Loya, Reference Ilan and Loya1990a). Haplosclerids have parenchymella larvae (Ereskovsky, Reference Ereskovsky1999), which in Haliclona as well as freshwater sponges at least, commonly possess a fringe of longer flagella that encircle a strongly pigmented posterior pole that is otherwise devoid of flagellated cells. Haplosclerid larvae have been described as possessing a dense bundle of oxeas located at the posterior end (Bergquist et al., Reference Bergquist, Sinclair and Hogg1970; Simpson, Reference Simpson1984; Wapstra & Van Soest, Reference Wapstra, Van Soest, Boury-Esnault and Vacelet1987; Woollacott, Reference Woollacott1993; Fromont, Reference Fromont1994; Ereskovsky, Reference Ereskovsky1999). The presence of choanocyte chambers inside larvae is uncommon, but has been documented in two marine, haplosclerid species, Haliclona limbata and Chalinula sp. (Meewis, Reference Meewis1939; Ilan & Loya, Reference Ilan and Loya1990a). Once released, free-swimming larvae can remain mobile hours to weeks before they settle and develop into sessile juveniles (Maldonado et al., Reference Maldonado, Giraud and Carmona2008).
The origin of spermatogonia of freshwater haplosclerids is from flagellated choanocyte chambers (Efremova & Papkovskaya, Reference Efremova and Papkovskaya1980; Sukhodolskaya & Papkovskaya, Reference Sukhodolskaya and Papkovskaya1985; Paulus & Weissenfels, Reference Paulus and Weissenfels1986; Paulus, Reference Paulus1989; Weissenfels, Reference Weissenfels1989) but whether the suborder Spongillina belongs in the order Haplosclerida has yet to be conclusively determined (Redmond et al., Reference Redmond, van Soest, Kelly, Raleigh, Travers and McCormack2007; Sperling et al., Reference Sperling, Petersen and Pisani2009). However, male gametes are also reported to develop from choanocytes in Petrosia ficiformis, a marine haplosclerid (Maldonado & Riesgo, Reference Maldonado and Riesgo2009). Female gametes of haplosclerids are thought to develop from choanocytes, or archaeocytes (e.g. Tuzet, Reference Tuzet1932; Meewis, Reference Meewis1936; Leveaux, Reference Leveaux1941; Brien, Reference Brien1967; Saller & Weissenfels, Reference Saller and Weissenfels1985; Saller, Reference Saller1988; Weissenfels, Reference Weissenfels1989; Ereskovsky, Reference Ereskovsky1999), and are associated with nurse cells; this is a greater specialization compared to other demosponge orders (Ereskovsky, Reference Ereskovsky1999).
Reproductive biology and/or the form and behaviour of larvae have been described for quite a number of marine haplosclerid species, e.g. H. oculata and H. xena (Wapstra & Van Soest, Reference Wapstra, Van Soest, Boury-Esnault and Vacelet1987), H. cinerea (Meewis, Reference Meewis1941), H. amboinensis and H. cymiformis (Fromont, Reference Fromont1994), Haliclona spp. (Fromont, Reference Fromont1999; Whalan et al., Reference Whalan, de Nys, Smith-Keune, Evans, Battershill and Jerry2008), H. loosanoffi (Fell, Reference Fell1976), Callyspongia paralia, Niphates rowi, Petrosia elephantotus and Amphimedon chloros (Ilan et al., Reference Ilan, Gugel and van Soest2004; Ilan & Loya, Reference Ilan and Loya1990a, Reference Ilan and Loyab), H. permollis (Elvin, Reference Elvin1976), Xestospongia testudinaria, X. bergquistia, and X. exigua (Fromont & Bergquist, Reference Fromont and Bergquist1994), Petrosia ficiformis (Maldonado & Young, Reference Maldonado and Young1996), Amphimedon queenslandica (Leys & Ereskovsky, Reference Leys and Ereskovsky2006), and some of these studies provide excellent detail on the morphology throughout development, while others are limited in scope, or old and/or brief, providing considerably less insight, including the publications on H. indistincta (Bowerbank, Reference Bowerbank1866; Lévi, Reference Lévi1956). Furthermore, reports on the reproductive biology and form of the larva differ for some species (reviewed in Wapstra & Van Soest, Reference Wapstra, Van Soest, Boury-Esnault and Vacelet1987) suggesting the possibility of species misidentification in some cases. This is especially likely in species of the order Haplosclerida due to their notoriously simple skeleton and morphological plasticity.
We describe here the reproductive cycle and characteristics of the larva of H. indistincta, a species commonly found in the intertidal regions around Ireland and elsewhere (DeWeerdt, Reference DeWeerdt1986). The species name is derived from its variable form and the scarcity of defining characters (Bowerbank, Reference Bowerbank1866). Its growth form includes thin sheets or cushions, and the colour ranges from light tan to bright purple. Its most defining characteristic is its sticky mucous consistency (Lévi, Reference Lévi1956; DeWeerdt, Reference DeWeerdt1986). Lévi's (Reference Lévi1956) description of reproduction in H. indistincta is focused upon the embryonic and larval stages and, though informative, is very brief. Our aim is to carry out a more detailed investigation of reproduction and development in this species and here we present data based upon histology and light microscopy.
MATERIALS AND METHODS
Material
To study the annual reproductive cycle of this species, Haliclona indistincta specimens (Figure 1A, identified via comparison of morphological characters with type material (DeWeerdt, Reference DeWeerdt1986) provided by the Zoological Museum University of Amsterdam) were collected from Corranroo, County Clare, Ireland (latitude 53° and longitude 9°). The sampling area was approximately 2000 m2. The tidal amplitude was a maximum of 5 m during spring tides and approximately 2.5 m at neap tides. In 2009–2010 the average surface water temperature (Malin Head station, 55°N and 7° W) ranged from 6.4°C in February to 15°C in August (http://www.met.ie/marine/marine_climatology.asp). The most prevalent rock types found in the area are granite and limestone, with limestone and sheets of algae as the preferred substrates of this species at this site. It can be an opportunistic settler as multiple specimens were also discovered growing upon the shells of bivalves.
Fig. 1. Haliclona indistincta; (A) adult specimen in situ; (B) adult specimen with developing larvae (arrows) throughout the mesohyl. Scale bar: 5 mm.
Ten complete specimens ranging from 19 cm3 to 30 cm3 were collected monthly from September 2009 to September 2010 with an additional twenty specimens collected between April and July. Specimens were randomly sampled from just below the waterline during the low spring tides. The distance between sampled individuals was not measured but varied from approximately 30 cm (individuals on the same rock) to 300 m (individuals separated distantly on the same region of shoreline). From April onwards, multiple specimens were dissected to identify whether larvae were present and to observe the distribution of larvae throughout the parent sponge.
Histological staining
Given that mature embryos/pre-released larvae were found distributed throughout the sponge bodies (Figure 1B), with a higher concentration seen in the mid-section, subsamples (5 mm3) were taken from the middle area of each specimen and fixed in Bouin's solution, and then transferred to 70% ethanol. The excised sections were washed with 70% ethanol until the Bouin's solution had been completely removed, and placed in histology cassettes in a histological tissue processor that moved the sections through a series of alcohol and Histoclear baths. The tissue was impregnated and blocked in paraffin wax. Successive serial sections (4 and 7 µm) were made with a microtome from each subsample, mounted on microscope slides, and stained with haematoxylin and eosin. Slides were examined using light microscopy (Olympus BX51). Each specimen slide was examined for reproductive elements (oocytes, eggs, spermatic cysts, embryos and pre-released larvae). Ten representatives of each of the reproductive elements present (per slide) were measured where possible. All measurements were taken from the longest diameter of each structure using imaging software Cell^D (Olympus). The average size and number of the various reproductive elements of all the specimens per month was then obtained. Summary statistics and graphs were generated in Microsoft Excel.
Larvae
Collection of larvae took place in the middle of the day during the low spring tides through the months May–August. Two methodologies were successfully employed to secure free-swimming larvae. Eight mature sponges were housed in situ by placing pyramid-shaped larval traps (made from nylon mesh with two collection tubes for larvae to swim into) over the rock containing the sponge. The larval collection tubes were checked at midday, five days a week during the last week of June and the first week of July, and once a week from the second week of July until the end of August. Collection tubes were replaced, and removed tubes were taken back to the laboratory where the contents were placed in 50 ml Petri dishes, and examined under a dissection microscope. We also induced release of larvae directly from parent sponges by cutting them open, allowing larvae to swim out into a container of water. Initially, larvae were held in varying densities in 25 ml Petri dishes in seawater. Subsequently, 69 larvae were individually maintained in 25 ml Petri dishes filled with filtered seawater. Specimens were allowed to settle and develop further. All stages were observed and photographed using a dissecting microscope (Olympus SZX16) and video recorded using a Panasonic 3CCD attached to a light microscope (Nikon SMZ-U).
RESULTS
Reproductive cycle
There was no evidence of reproductive elements in specimens collected in September, December and January (Table 1; Figure 2A). All individuals collected between February and May contained reproductive material; this number dropped to 70% of individuals collected in July but was 100% in August (Figure 2A). The first sign of oogenesis was seen in October when one amoeboid oocyte (29 µm in maximum diameter), without the presence of nurse cells, was observed in one specimen. Early oocytes were also seen in low numbers in November (one in each of two specimens, 28 and 32 µm). There were no early oocytes present in December and January. Of the 58 early amoeboid oocytes observed (and measured) in specimens from October through to August, the highest numbers were observed in March (14), April (18) and May (9). Early oocytes were still present in June–August, but were very few in number with three being present in June (sections from three specimens had one early oocyte each), one each in two July specimens, and four in August in three separate specimens.
Fig. 2. (A) Graph showing (per month) the percentage of individuals sampled that were reproductive and non-reproductive. Also showing a plot of the water temperature over the same time period; (B) graph showing the average diameter of reproductive elements (standard deviation (error bars) included). Also showing a plot of the water temperature over the same time period; (C) graph showing the percentage frequency of each reproductive element of Haliclona indistincta over a twelve-month period, using 10 replicate individuals per month, e.g. 71% of all late oocytes/eggs were seen in June. The numbers of each reproductive element counted were: spermatic cysts N = 2166; oocytes N = 2133; late oocyte/egg N = 92; embryo N = 13; larvae (i.e. pre-released larvae) N = 23. For ‘oocyte' both early and mid stages were included, and for ‘spermatic cyst' both immature and mature stages were included.
Table 1. Numbers of Haliclona indistincta specimens per month showing those that were found to be hermaphrodites (H), male (M), female (F) and those which had embryos and larvae of different stages. 0:10 = no specimen out of a total of ten specimens surveyed; 10:30 = ten specimens out of a total of thirty specimens surveyed.
Reproductive elements increased in size from February through to August, in parallel to the rising water temperature (Figure 2B, C). The diameter of oocytes (Figure 3A) ranged from 29 µm to 78 µm (N = 1013 mean 58 µm ± 27 µm standard error (SE)). Choanocyte chambers filled with cells and surrounded by a ring of choanocyte cells, suggestive of early spermatozoa (Supplementary Figure 1A), were seen from February through to the end of August. Fully developed spermatozoa were only seen during the months of May and June (Table 1) in one-sixth of the specimens surveyed.
Fig. 3. (A) Early oocyte; (B) egg; (C) developed embryo, longitudinal-section. Scale bar: 10 µm.
Of a total of 200 mature Haliclona indistincta specimens examined throughout the year, 55 were hermaphrodites, 84 had female gametes only, while four had male reproductive tissue only (Table 1). A number of sponges contained developing oocytes and what we presume were early spermatozoa from February, and female-only sponges were also present at this time. By May, four sponges contained both fully developed spermatic cysts and late ooctytes, while others still contained early spermatogonia along with early and late oocytes. Mature male-only sponges (i.e. with fully developed spermatic cysts, Supplementary Figure 1B) were seen first in June.
By the end of May, late oocytes (eggs) were seen in four out of the thirty specimens sampled and no embryos or larvae were seen. Eggs (Figure 3B), were circular and had an outer layer composed of yolk granules with an average size of 9 µm (N = 10 ± 1.7 µm (SE)) with nurse cells contained within the cytoplasm, as is typical of haplosclerids, with an average maximum diameter of 3 µm (N = 10 ± 1 µm (SE)), and a large amount of non-cellular material. The diameter of the eggs ranged from 186–220 µm (N = 261, mean 206 µm ± 45 µm (SE)). Embryos were similar to the eggs in shape as they were also spherical but ranged in size from 239–252 µm (N = 64, mean 245 µm ± 22 (SE)). They did not possess the prominent layer of yolk granules on the exterior, and the internal cavity showed significantly less non-cellular material, being filled instead with dense clusters of cells. As the embryos matured inside the adult body, they attained a more elongated oval shape and, internally, the non-cellular material present formed a layer that separated an outer layer of epithelial cells and an internal area that contained a number of cell types (Figure 3C and Supplementary Figure 2). No spicules were present. The highest number of specimens containing mature embryos was in July, and by August, mature embryos were found in only one individual (Table 1). Mature embryos/pre-released larvae occurred in groups of varying numbers and were not held in defined brood chambers but were heavily surrounded by mucus (Figure 1B). The size of free-swimming larvae ranged from 414 µm to 477 µm (N = 41, mean 469 µm ± 20 µm (SE)).
Larvae
Six larvae were retrieved from one larval trap on 29 June and over the six days between 1 July and 8 July a total of 67 larvae were collected from eight traps (i.e. eight sponges). The numbers of larvae collected in this way varied per day ranging from 0 larvae present in the traps to a maximum of 26 larvae present in a single trap. Only three sponges, out of the eight, released larvae over this time period. No larvae were present in the traps on 10 August. Most of the larvae used for observations were obtained through induced release in the laboratory. Three distinct free-swimming larval stages were observed. Figure 4A shows the initial free-swimming, mobile stage, which was oval and showed anterior–posterior orientation. The anterior end pointed upward when stationary, and at a slight angle when in motion (in the direction of movement). Larvae were found to swim in an anticlockwise, corkscrew pattern. The larvae were uniformly ciliated and the pink pigment associated with this species showed a higher concentration at the posterior pole. Figure 4B, shows a later stage where the shape of the larvae had changed to being compact and angular. The pale anterior end diminished in size, but the anterior–posterior orientation was still apparent, especially when swimming. In a third stage (Figure 4C) the body was nearly perfectly circular, rotating on its axis, the pink colour of the free-swimming stage still apparent.
Fig. 4. Haliclona indistincta larvae: (A) free-swimming stage; (B) compact angular stage showing a diminishing anterior end; (C) semi-spherical ball stage; (D) free-swimming larvae fusing with two other already combined larvae. Scale bar: 100 µm.
A fusing behaviour among the larvae was observed multiple times throughout this study; the highest number of events taking place in two separate (25 ml) Petri dishes holding 29 and 16 larvae, respectively. Of the dish containing 29 larvae, four separate fusing events were observed, with one fusion comprising three larvae (Figure 4D), and the remaining three events involving two larvae. In the dish containing 16 larvae, two separate fusion events each involving two larvae occurred. A third distinct fusing event was observed between two larvae when they were placed together in a drop of seawater (10 µl) for ten minutes. In all cases, larvae were observed to fuse and continue to develop to settlement stage.
Variability was seen in the amount of time it took for each larva to progress through the mobile stages. The shortest time from release to settlement was 25 hours (11/69 larvae), and longest time was 285 hours (3/69 larvae), with an average time of 161 hours (7 days). When about to settle, larvae ceased their horizontal forward corkscrew movement, though still spinning anti-clockwise, and eventually remained stationary, attached via the posterior pole and settled. Two hours after settlement, specimens began to flatten onto the substrate. After 24 hours, specimens were almost completely flattened. Subsequent days produced ostia formation (Figure 5A), the development of excurrent channels (Figure 5B) and oscular opening formation (Figure 5C). Sponge larvae, and newly settled sponges, were observed to be the dietary target of a water mite that could not be removed by filtering the water. No sponge survived beyond 39 days after release.
Fig. 5. Post-settlement stages of Haliclona indistincta: (A) juvenile 37 hours post-settlement showing surface becoming punctated (arrows); (B) juvenile 39 hours post-settlement showing excurrent canals (arrows); (C) juvenile 115 hours post-settlement, formation of oscules and excurrent canals (arrows), colour change associated with addition of Phaeodactylum tricornutum (diatom used in feeding juvenile sponges).
DISCUSSION
We have added to the current knowledge about the reproductive cycle and early development of Haliclona indistincta from that presented by Lévi (Reference Lévi1956). We show the species to be primarily gonochoric but we cannot rule out that this species may be sequentially hermaphroditic as individual adult sponges were not sampled throughout the season. Female-only sponges appeared first and male-only sponges appeared later (in June), while hermaphrodites were present between February and August. Males were very few in number (4/200). It is possible that the focus on the middle section of the sponges resulted in missing male gamete production in the specimens surveyed. The asynchronous gamete production observed, a strategy commonly found in viviparous sponges (Simpson, Reference Simpson1980; Tarin & Cano, Reference Tarin and Cano2000), can result in extended larval release, as also seen in this species (mature larvae present from the end of June to early August).
The first appearance of identifiable oocytes showed amoeboid cells, which were very similar to archaeocytes though larger. Their form, and given that they were not found associated with choanocyte chambers, leads us to suggest an archaeocyte origin for oocytes, which is also consistent with other Haplosclerida (Leys & Ereskovsky, Reference Leys and Ereskovsky2006). The identification of male reproductive elements was difficult, however, we suggest that spermatogenesis in this species is consistent with other haplosclerids. While the identity of mature sperm in the spermatic cysts were confirmed via transmission electron microscopy (TEM) (Supplementary Figure 1C), it was not possible to confirm the identity of developing spermatozoa in the same way. However, the presence of choanocyte-filled chambers (Supplementary Figure 1A) is consistent with the manner in which spermatic cysts develop in other sponges including the marine haplosclerid Petrosia ficiformis (Maldonado & Riesgo, Reference Maldonado and Riesgo2009), and other haplosclerid species, as previously described (e.g. Efremova & Papkovskaya, Reference Efremova and Papkovskaya1980; Sukhodolskaya & Papkovskaya, Reference Sukhodolskaya and Papkovskaya1985; Weissenfels, Reference Weissenfels1989; Ilan, et al., Reference Ilan, Gugel and van Soest2004), as well as other orders (Halicondrida, (Barthel & Detmer, Reference Barthel and Detmer1990) and Dictyoceratida (Kaye & Reiswig, Reference Kaye and Reiswig1991)).
We observed fewer embryos in the months June–July than oocytes or larvae. It is possible that oocytes were reabsorbed into the sponge body and did not contribute to the reproductive effort. It is also possible that through the sampling and sectioning approach adopted, some embryos were missed. In a similar manner to the embryos described by Lévi (Reference Lévi1956), the embryos seen during the current study showed the presence of non-cellular material associated with the oocytes, embryos and larvae from an early stage (February). This material made it difficult to document cleavage and early development using the approaches employed here and we are currently pursuing TEM for more clarity. Unidentified, dense, granular material has been found to be associated with developing oocytes of a number of sponge species including the marine haplosclerid Haliclona cinerea (Tuzet, Reference Tuzet1947), and Stelletta grubii (Liaci & Sciscioli, Reference Liaci and Sciscioli1967), Suberites massa (Diaz et al., Reference Diaz, Connes and Paris1975), Aplysina cavernicola (Gallissian & Vacelet, Reference Gallissian and Vacelet1976), Hippospongia lachne, Spongia barbara, S. graminea and S. cheiris (Kaye, Reference Kaye1991) and, though its function and composition is still unknown, it has been suggested to be important to oocyte development (Simpson, Reference Simpson1984; Kaye, Reference Kaye1991). We made initial attempts to identify the material (Masson's trichrome with Gomori's aldehyde fuchsin stain to identify collagen and PAS (periodic acid-Schiff) to identify carbohydrate) with no success; further work is ongoing on this material.
Increase in the numbers of reproductive individuals, as well as the diameter of reproductive elements (e.g. eggs and embryos) corresponded to an increase in water temperature (Figure 3A–D), a particularly important dynamic for initiating reproductive processes as shown by other authors (e.g. Fromont & Bergquist Reference Fromont and Bergquist1994; Witte et al., Reference Witte, Barthel and Tendal1994; Ereskovsky, Reference Ereskovsky2000; Whalan et al., Reference Whalan, Battershill and de Nys2007; Maldonado & Riesgo, Reference Maldonado, Riesgo, Durfort and Vidal2008). The timing of larval release described here is shorter than described in Lévi (Reference Lévi1956), as we did not find any mature larvae until the end of June while he suggested a period from May to July. It is possible that the geographical difference between the two study sites (Clare, Ireland versus Brittany, France) may have contributed to this difference. The size of the reproductive elements also differed between the studies; Lévi (Reference Lévi1956) recorded an egg size of 300–350 µm, which is larger than the eggs recorded in our study and the largest embryo we measured was 252 µm. Lévi (Reference Lévi1956) recorded a maximum length of larvae in Brittany to be 500 µm, while the corresponding value for the Clare population is 477 µm. The geographical difference could also contribute to this discrepancy in the size of reproductive elements. However, we have also observed two morphotypes currently ascribed to H. indistincta. Details of the morphology of the specimen used in Lévi's study are not provided (we investigated the cushion form), and it is possible that the larvae from the two studies are from different morphotypes. Further work on the two morphotypes of this species and on specimens from different populations would shed light on this matter.
We have also observed the fusion of larvae in this species as has been described for other sponges (e.g. Ophlitaspongia seriata (Fry, Reference Fry and Crisp1971), Crambe crambe (Van der Vyver, Reference Van deVyver1970), Halichondria panicea, Hymeniacidon perlevis, Pachymatisma johnstonia, Clathrina (Leucosolenia) coriacea, Iophon hyndmani, and Dercitus bucklandi (Fell, Reference Fell, Giese and Pearse1974)), including haplosclerids (e.g. Chalinula sp. Ilan & Loya (Reference Ilan and Loya1990b), Petrosia ficiformis, Maldonado & Riesgo (Reference Maldonado and Riesgo2009), Haliclona sp. (McGhee, Reference McGhee2006) but it is still relatively little studied.
All fusing larvae from H. indistincta seen in this study were from the same parent sponge. Only one-sixth of adult specimens collected in May and June possessed both spermatic cysts and oocytes, therefore, while self-fertilization may be possible in these individuals (and yield more genetically similar offspring) it is unlikely to be the case for the entire population. Without genotyping adults and larvae from such fusion events it is difficult to predict whether they are chimeras of siblings that are more genetically similar than expected from sexual reproduction, or not. Ilan & Loya (Reference Ilan and Loya1990b) and McGhee (Reference McGhee2006) have shown that larva fuse indiscriminately with other genetically distinct larvae (and juveniles), but that adults will only fuse with fragments from genetically identical individuals, suggesting that the mechanism allowing distinction of self from non-self develops sometime after settlement (Ilan & Loya, Reference Ilan and Loya1990b). Fusion of sponge larvae and early juveniles may be an advantage as a higher rate of survivorship has been demonstrated for juvenile sponges that had previously formed fused groups (e.g. Fry, Reference Fry and Crisp1971; Connell, Reference Connell, Jones and Endean1973; Highsmith, Reference Highsmith1982; Ilan & Loya, Reference Ilan and Loya1990b). Also, a larger body size gained through fusion may allow an individual to be sexually mature sooner (Connell, Reference Connell, Jones and Endean1973; Highsmith, Reference Highsmith1982; Ilan & Loya, Reference Ilan and Loya1990b).
However, fusing appears to be more likely to occur if larvae are under some sort of stress, e.g. forcefully moved, overcrowded or restricted (Ilan &Loya, Reference Ilan and Loya1990b; McGhee, Reference McGhee2006) as was also seen during this study for H. indistincta. If close proximity of larvae induces the fusion of individuals, which our findings would suggest to be the case, then it could also be a competitive mechanism in environments where resources are limited (McGhee, Reference McGhee2006).
Ereskovsky (Reference Ereskovsky2010) described the early presence of a dense bunch of spicules at the posterior pole of haplosclerid larvae as a common feature of this group, and reported this in a number of Haliclona species, e.g. H. aquaeductus, H. cinerea, H. ecbasis and some fresh water sponges (Ereskovsky, Reference Ereskovsky1999, Reference Ereskovsky2010). However, spicules were not evident in the larvae of H. indistincta and were also not mentioned by Lévi (Reference Lévi1956) in his work on this species. Erevskovsky (Reference Ereskovsky2010) also indicated that Haliclona larvae possess a ring of longer cilia at an otherwise bare posterior pole, however H. indistincta were uniformly ciliated (also described by Lévi, Reference Lévi1956). The external characteristics of H. indistincta larvae were also not reminiscent of the Haliclona sp., Chalinula sp. or C. limbata larvae illustrated by Mariani et al. (Reference Mariani, Uriz and Turon2005), lacking the longer posterior cilia and/or the pointed anterior pole. Because we observed this uniform ciliation both in larvae whose release from the parent sponge was induced, and from larvae collected in larval traps, the external morphology observed is not likely to be due to immaturity. Further, Maldonado & Bergquist (Reference Maldonado, Bergquist, Young, Sewel and Rice2002) also describe the larva of H. petrosioides as being uniformly ciliated.
Bergquist (Reference Bergquist, Lévi and Boury-Esnault1979) reported two larval types in the Haplosclerida. While she described Callyspongia, Adocia and Haliclona as having a ring of long cilia and a bare posterior pole, Reniera, which is now a synonym of Haliclona and Chalinula larvae both lacked this ring of longer cilia. While Wapstra & van Soest (Reference Wapstra, Van Soest, Boury-Esnault and Vacelet1987) predicted that Bergquist's Reniera would belong to Halichondrida and Chalinula to the Poecilosclerida this was never subsequently determined to our knowledge. In any case, recent molecular phylogenies suggest that the family Chalinidae and the genus Haliclona are polyphyletic and that species of what were Reniera (e.g. H. fulva) are not closely related to the type species H. oculata (Redmond et al., Reference Redmond, Raleigh, van Soest, Kelly, Travers, Bradshaw, Vartia, Stephens and McCormack2011).
Another unusual aspect of H. indistincta larvae is that they settled on the posterior pole rather than the anterior pole, which was also described by Wilson (Reference Wilson1935) for the poecilosclerid Mycale syrinx but appears to be otherwise uncommon in sponges. Langenbruch & Jones (Reference Langenbruch and Jones1990) suggested that adult H. indistincta had a different morphological structure from other Haliclona species after histological examination of the structure of the choanoctye chambers in particular. Given that recent molecular data suggest that Haliclona probably contains a large diversity of distantly related species (e.g. Raleigh et al., Reference Raleigh, Redmond, Delahan, Torpey, van Soest, Kelly and McCormack2007; Redmond et al., Reference Redmond, van Soest, Kelly, Raleigh, Travers and McCormack2007, Reference Redmond, Raleigh, van Soest, Kelly, Travers, Bradshaw, Vartia, Stephens and McCormack2011), our results may also support the suggestion that H. indistincta is not closely related to other typical Haliclona species (e.g. the type species H. oculata).
ACKNOWLEDGEMENTS
We thank Paul Casburn, Jack Darcy and Ken Maher of the MRI Carna facility for culturing the diatoms and algae used in this study. We also thank Mark Canny from the Anatomy Department, National University of Ureland (NUI) Galway for technical assistance and the anonymous referees for suggesting improvements to the manuscript. This project has been funded through an NUI Galway PhD fellowship awarded to Kelly Stephens.
Supplementary materials and methods
The supplementary material referred to in this paper can be found online at journals.cambridge.org/mbi.
