Cryptobiota associated to dead Acropora palmata ( Sde:ractinia : Acroporidae ) coral , Isla G:rande , colombian Caribbean

Cryptobiota of dead fragments of five branehes in live position and five fallen pieces of the coral Acropora palmata each one of approximate Idm3, eovered by filamentous algae were extraeted from the north reef erest of Isla Grande (Colombian Caribbean), in April 1991. There were three groups of organisms according lo size and position (on and wi!hin the coral): 1) mobile epibenthos, mainly microcrustaceans ¡hat ¡ive among the fIlamentous algae 2) boring microcryptobiota, located in lhe layer between ¡he epilithic organisms and the coral skeJeton itself and, 3) perforating macrocryptobionts Ihat bore <LTld penetrate the coral skeleton. Polychaetes, sipunculíds, mollusks and crustaceans were mos! abundant in !he las! group. There were no dífferences in macrocryptobiont composition between standing dead branches and fallen fragments. There was a large variation in total biomass and type and density of macrocryptobionts, possibly associated to stochastic factors such as placement and thickness of branches and small scale variations in recruitment.

The structure and shape of reefs are the result of a dynamic system of constructive and destructive forces that constantIy transform them (Schuhmacher 1978).The speed at which a reef is formed depends on rates of skeletongenesis and consolidation, whíle destruction rates depend upon physical, chemical and biological eroding processes (Risk andSammarco 1982, Hutchings 1986).
In a healthy reef, bioerosion is very important since it increases the complexity of substrates, constantly creating free spaces for coral ]arvae and other organisrns, thus maintammg diversity in species and renovating populations (Hutchings 1986); however in sorne zones s1Ich as the reef crest of Isla Grande where coral growth is limited, bioerosion can eventual1y kilI the reef.
The purpose of this paper is first to describe the bioeroding cryptic fauna associated to skeletons of dead Acropora palmata, responsible for the bioeroding action.Cryptic fauna means organisms living in small pores, holes or cavities within the rock or inside organisms that cover them (Hutchings 1978, Peyrot-Clausade 1981, Scott 1987).Second, to establish whether differences exist between the cryptobiota inhabiting colony branches in life position and broken fallen fragments.

MA1ERIALS AND METHODS
Area: Rosario islands' (1008'-10015' N, 75040'-75°48' W) reef complex ís formed by approximately 28 coralline islands, 45 kilometers southwest of Cartagena (Colombia); most of the islands are.fringed by reefs and seagrass meadows, which are protected within the Rosario Coral Natural National Park (PNNCR) (Sarmiento et al. 1989).Isla Grande is the largest of the islands; paralJel to its northem margin, and separated by a small reef channel, Hes a large fringíng reef reaching 950 m in length (Penereiro 1988).Initial studies in this zone describe a hea1thy coralline community (Pfaff 1969, Werding andSánchez 1979).However, from 1982 onwards, a drastic reduction of coral cover, mainly of Acropora palmata, was detected, reaching a values as less as 0.2% of cover of living tissue in 1988 (Garzón-Ferreira and Kielman 1993).The origin and evolution of the degradatíng process are not yet well understood.
Methodology: coral fragments covered by filamentous algae were coHected in April 199 1.Five branches in live position and five fallen fragments, approximate volume of ldm3 each, of A. palmata were evaluated (Brock andSmith 1983, Grajal andLaughling 1984).Each fragment was covered with a tightly sealed cotton bag, then broken into small pieces with hammer and chisel over a thick blanket, preventing any 10ss of material.Macroscopic organisms were narcotized for six hours using magnesium sulfate, then fixed in 10% formaldehyde neutralized with borax and finaIly preserved in alcohol or formaldehyde, depending from the organims.The volume of each fragment was determined by water displacement (Young 1986).
The sediments and material contained in the bags, as well as the remains obtained after macroscopic sorting with hammer and chis el were meshed with a 100 micron mesh.The resulting material was observed with a stereoscope and the organisms found were preserved in 70% alcohol.Mathematical analysis of macrocryptobiota was performed on a family level, the lowest taxonomic unit cornmon to all organisms found.Bray-Curtis dissimilarity between fragments was calculated with density data (number af individuals per dm3).
As an approximation to the total biomass (epibiont and cryptobiont), six píeces of coral were fragmented, then dried and calcinated until constant weight was reached.Results were expressed as percent of a coraJ's dry weight.

RESULTS
In a transversal cut of a branch of A. patmata three horizons can be ciearly dístinguished.The supe110r conformed by a film of fílamentous algae and its associated organisms.Then, the middle part of the branche, that is skeleton, and the inferior horizont, formed by a crust of coralinaceous algae.
Then, dead Acropora palmata's skeletons supports tbree types af organisms: Mobile epibenthon: Composed by the fauna obtained from the examined sediments, living in assodations over and under comlline surfaces covered by filamentous algae.Most abundant were small crustaceans, micropolychaetes and foraminífers.Gammarids and caprellids were common amphipods both found in all states of development, tanaidacea was very cornmon and freqllently found with eggs.Conceming isopoda, asellots were the most abundant in number while most of the species were f1abellifers (Dinamenella, Geocerceis and Exocoralla genus).The most cornmonly found foraminifers belong to the genera Triloculina and Qinqueloculina, (Myliolidae) although individuals of the Cyclogyra planorbis, Articulina mucronata and Sphaeroina bulloides were also frequent.
Associated molluscs Tricolia affi nis, Anachis pulchela, A. obesa, Volvarina sp" Hyalina pallida, Cyclostremiscus sp. and various .. species--nf Caecum were. the species commonest found.It's important to.notice the presence of the gastropod lviarissa cornuaretis, a freshwater mollusc frequently found in the fragments; mainly as empty juvenile shells that are possibly brought to the sea attached lo the roots of "taruyas", Eichornia crassipes.This "taruyas" comefrom the Dique channel, whích Hes 45 kilometers apart from theR osario islands, and that Connects the Magdalena fiver with the Cartagenabay.
showed varyiug• thickness•• qepending on ag� and sitewherethe colony was extra!.:ted.111is skeleton seemed perforateg by tubes and holes in an random manner.Cryptobiota, localízed in the body of the COraUinf¡ skeleton were the organisms quantitatively evaluated, measuring 5 millimeters or more.Polychaets were the most abundant fol1owed by sípunculids, mollúsks and crustaceans (Table 1   of polychaetes with few individual s that were not ídentifiable.Polychaetes were frequently associated to clionid sponges and foraminifers, a common sÍtuation in these habitats (McCloskey 1970, Lipps andRonan 1974).
The second largest component in number of organisms were the sipunculids, reaching a density of 213 JJdm3 (Table 1).The population was formed mainly by three genera.Aspidosiphon, type 2, and Phascolosoma, present in all fragments.Behavior is similar to that of polychaetes.Big populations were always present in the fragments but with large variations between them.Various sípunculíd perforations communicated directly with the galleries of Alpheus schmitti, and not to the exterior.Tms was a frequent finding.
The vermetid Dendropoma corrodens is the most abundant mollusc (89 I1dm3) found in the fragments extracted from the north reef crest of Isla Grande (Table 1).Bívalves were the next group in abundance.Caves belonging to bivalves showed intensive infestation by Siphonodyction coralliphagum sponge.Other mollusc individuals were poorly represented in the samples.These were Fissurella angusta, and chitons like Acanthochitona sp., lschnochiton striolatus and Choneplax lata, of scraping habits.Others were the filtrating bivalves lsognomon bicolor, 1. radiatus and free living gastropods such as Engina turbinella and Leucozonia ocellata.
We coUected a total of 255 macrocrustaceans distributed in sev�n families and twelve species (Table 1).Eíght Thunor simus cavities in the superior pru1: of a 14 x 14 cm rock were found, each with an eigth shape.They measured 1.32cm maximum mean width and 3.42cm in length.Each of the chambers had a paír of adults; females, always with eggs, and using the larger chamber while the majes the smaller.These cavitíes had no galleries or tunnels interconnecting them, and organisms appear to be completely covered by the skeletal crust of coraUinaceous algae.
The most common boring species found in the A. palmata scaffold were Cliona cf.celata, Cliona spp., Siphonodictyon coralliphagum, Geodia cf.gibberosa and Aka sp.At a glance of dead branches, presence of any organism is nothing but conspicuous, for their openings to the exterior are through tiny pores.The biggest observed perforatíons belonged to Geodia cf.gibe rosa reaching a length of 5 centimeters.Those of Siphonodictyon coralliphagum were of intermediate size and Cliona sp. were the smallest.
Anemones of the Lebrunia coralligens and severa} species of fish were also found.Sorne sampling showed a Gobiessocidae Gobiesox punctulatus and the blennoid Hypleurochilus springeri; another two samples had an Acanthemblemaria aspera, and A. rivasi respectively, all of them inhabiting cavities built by other organisms (Greenfield and Greenfield 1982).
Representativity evaluation, based on density data of different families, showed that an acceptable sample is that of four sampling uruts of one dm3 each.A dissimilarity dendrogram between fragments from both coral positions (Fig. 1) done according to quantified family densities, showed no difference between macrocryptobiota inhabiting líve position and faUen fragment corals.

DISCUSSION
The tbree kínds of described organisms are found in dead coral blocks (Brawley and Adey 1981, Rice and McIntyre 1982, Grajal and Laugbling 1984, Seott and Risk 1988) as weH as in rocky shores and submerged rocks, and are not restricted to this habitat and to this particular specíes (McCloskey 1970, Hutchings 1983, ----------------- Mobile epibenthon: Organisms found here (microcrustaceans, micromolluscs and foraminifera) are the most common elements of such habitats (Bell and Deblin 1983,  Hutchings 1983, Klumpp et al. 1988).This order of abundance changes between reefs and zones, and depends upon bioJogical cycles of the species (Klumpp et al. 1988) These abundant kinds of organisms plays a very important roH acting as links between filamentous algae, very abundant in this kind of habitats (Navas et al. in press) and important as primary productors, and reef secondary consumers as fishes, crustaceans larvae and criptobiontes, etc (Brock and Smith 1983).
These organisms are microperforators that allow the entrance of others due to their capacity of chemical dissolution of coral skeletons and are solid bases for microbenthic nourishment (Hutchings 1986 andMay et al. 1982).
Macrocryptobiota: First of all we must consider that fauna diversity and abundance i8 seriously affected by: l.
Hammer and chisel method destroys an important amount of organisms.2. Some very abundant individuals such as polychaetes were counted on a family level.3. Organisms with few individuals impossible to identify were not considered for the quantitative analysis.Classification analysis showed that there was no differences between• the two coral positions (live and fragmented) this could happen because there 1S no differential colonization due to 10calization, or because the coUected fallen fragments were not fixed to substrata so probably they had recently fallen from their branches.
The results of the homogeneity of the fragments and the reduced volume of coralline rock needed to evaluate the community were valid only on a family leve!.Variatíons of biomass, density and number of identified organisms on a more specific ground were too ample.This difference between the fragments was due to multiple factors such as: 1.The characteristic position of each fragment plays an important role in the determination of the type of cryptobiont larvae that colonize (Obrebski 1979, Hutchings 1981).2. The monopolizatíon of various sectors of the substratum by species of closed recruitment and asexual reproduction (Hutchings 1978).3. Morphologic characteristics of each branch, thickness in particular, seem inversely related to the density of perforating organisms.4. The type of epibiota, which favors, for filamentous algae, or reduced, for incrustíng or carnivorous organisms, the settlement of the eroding larvae (Klumpp et al. 1988).5. Territorial behavior of fish Jike Micro:,pathodon chrysurns and Stegastes dorsopunicans.They are abundant in the crest, and proteet sections of the coral where filamentous algae grow.This at last favor cryptofaunal colonization, particularly by sponge and sipunculids due to a probable ¡ncrease in larval settlement, reduction in herbivory and an ¡nerease in food supply (Risk andSammareo 1982, Winston andJaekson 1984).6. Pioneering colonization during dry or wet season, which determines the ulterior composition of the community (Coles 1980, Hutchings 1981).
The biomass values showed intrinsic differences between each piece of coral, random location of individuals and their occasional presence give variable results.
A. palmata mortality in Isla Grande occurred in mass between 1982 and 1983, probably in assoclatlOn with high temperatures (Solano et al. 1992).Until1992 there has not been a significant coral recolonization.What we have observed, though, during this survey, is the result of bioerosion and colonization processes after a period of approximately 8 years.In this succession stage and even though sponges population was not quantified, they seemed mostly responsible for bioerosion of A. palmata skeletons in Isla Grande.These organisms are the most important live or dead coral macroborers (Highsmith 1980, Moran andReaka 1988).The present organisms community, specially polichaetes, according to Davies and Hutchings (1983), are important features in reefs under active degradation processes.
In the other hand sorne of the secondary colonizers such as vermetids Dendropoma corrodens, coralline algae and sorne sponges "fix" coral branches (Wilkinson 1983, Marshall 1983) and make the bioerosion process slower.
Finally, if antropogenic stressers such as pollution, tourism, high temperatures, freshwater inputs and sedimentation keep on acting over Rosario Islands' reefs and avoiding recolonization in affected areas, reef structures such as that found in the north crest of Isla Grande are sentenced to disappear.ACKNOWLEDGMENTS Special thanks to Clara Osorio for field activities support, Camilo Garcia for his help with statistic software, Rafael Lemaitre and J. M. Diaz for their help with identification and confirmation of species.
Fig. 1.Affinity dendrogram of cryptofauna between sampling units using Bray-Curtís index based on family densities data.SU: sampling units FF: Fallen fragments LP: Live position.
), when 7574 individuals were counted and identified in 12.35 dm 3 of sample.Serpulidae families showed low densities.Furthermore there Were fifteen types