Successional and seasonal events in the fouling of naturally occurring substrate in a subtropical estuary

Successional and seasonal events in the fouling of naturally occurring substrate in a subtropical estuary

Material Information

Successional and seasonal events in the fouling of naturally occurring substrate in a subtropical estuary
Cuba, Thomas R.
Place of Publication:
Tampa, Florida
University of South Florida
Publication Date:
Physical Description:
xi, 145 leaves : ill. ; 29 cm


Subjects / Keywords:
Marine fouling organisms ( lcsh )
Dissertations, Academic -- Marine science -- Doctoral -- USF ( FTS )


General Note:
Thesis (Ph. D.)--University of South Florida, 1984. Bibliography: leaves 127-137.

Record Information

Source Institution:
University of South Florida
Holding Location:
University of South Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
030077519 ( ALEPH )
12772639 ( OCLC )
F51-00165 ( USFLDC DOI )
f51.165 ( USFLDC Handle )

Postcard Information



This item is only available as the following downloads:

Full Text


SUCCESSIONAL AND SEASONAL EVENTS IN THE FOULING OF NATURALLY OCCURRING SUBSTRATE IN A SUBTROPICAL ESTUARY by Thomas R. Cuba A Dissertation submitted in partial fullfillment of the requirements for the degree of Doctor of Philosophy in the Department of Marine Science in the University of South Florida Decembe r 1984 Major Professor: Norman J Bl a ke


Graduate Council University of South Florida Tampa, Florida CERTIFICATE OF APPROVAL Ph.D. Dissertation This is to certify that the Ph.D. Dissertation of Thomas R. Cuba with a major in Marine Science has been approved by the Examining Committee on September 21, 1984 as satisfactory for the dissertation requirement for the Ph. D. degree. Examining Committee: c Johfi1C: Briggs < < > Susan S. Bell


c Thomas R Cuba 1 984 All Rights reserved


In Nomine Patris ii


ACKNOWLEDGEMENTS I am most grateful to Dr. Norman J. Blake for both the gentle nudges and the swift kicks that were needed to produce this work. I also thank the other members of the committee, Drs. Jack Briggs, Susan Bell, Joseph Torres, and John Paul for their special insights and encouragements. I am indebted to Tom Perkins, Pam Muller, and Doug Parker who assisted in the indentification of the Polychaeta and Foraminifera. I am especially grateful to Dr. Harold Humm for his "sa voir fa ire" and general advice. Bruce Barber and Rob Erdman deserve mention for their comraderie and assistance throughout my time as a graduate student. I must also thank B Alle n Patrick, who assisted in my introduction to digital computers, Tony Greco, for his p atience and aid in my work with the Scanning Electron Microscope, and Deborah Walton for the drafting of the figures contained herein. Lastly I would like to thank my wife, Terri, who responded to every situation, both difficult and pleasant, with the utmost grace and composure. iii






LIST OF TABLES Table 1. Meiofaunal colonization. Sum of counts for all samples. 30 Table 2. The fourteen most abundant species. N = Total number of individuals; Freq = the number of replicates in which the species was found; % = the percent composition with respect to the data set after deletion of those individuals not identified to the species level. 50 Table 3. Interactions between species. + = facilitation, = inhibition, and o = tolerance. Species #1 was also totally preempted by open space. 104 Table 4. Initial composition of the larval pool (abundance 1) and the abundance of each species in the pool after alteration (abundance 2). notation indicates range. 106 Table 5. Similarity levels among six developing simulated assemblages. C2 models were allowed to develop normally. C3 and C4 models underwent changes in larval availability at 20% and 50% site saturation levels respectively. The upper right sector contains similarity values at 10% site saturation and the lower left, at 100% saturation. 111 Table 6. Similarity values between successive pairs of model assemblages (M series); values between each model assemblage and the actual assemblage (A); and the accumulation of species/sample 145 vi


LIST OF FIGURES Figure 1. The study site. Flag indicates the location of the sampling apparatus. Bridges have been omitted for clarity. 18 Figure 2. Salinity and temperature from to a. Salinity in parts per thousand. b. Temperature in degrees C. Range bars indicate high and low for previous 24 hrs. 21 Figure 3. Colonization curves for microalgae (*) and bacteria (o). Average of 4 sample sets. 26 Figure 4. Pennate diatom, bacteria, and microdetritus on surface of a scallop shell immersed in seawater for 48 hrs. 28 Figure 5. Microalgal species colonization curves. Number of species of microalgae per sample (o) and % of microalgal species encountered by the corresponding sample period(*). 33 Figure 6. Clusters formed by dendrogramming sample similarities where solid line (commencing at 0.5) indicates significant clustering. Sample numbers consist of exposure time in hrs (numeric) and sample series (Cyrillic). 35 Figure 7. a. Number of species/sample. b. Number of new species/sample. Set A(.), Set C (x). 52 Figure 8. S factor of diversity (ordinate) plotted against the X factor (abcissa). Set A data are open circles, set Care 's. 55 Figure 9. Sequential dissimilarity levels. a. Set A. b. Set c. Based on species and their distribution (.) and species presence/absence alone (*). 57 Figure 10. Dendrogram of sample to sample similarities where solid line (commencing at 0.5) indicates significant clustering. Sample numbers consist of exposure time and experiment code. 59 Figure 11. Dendrogram of sample to sample similarities considering species presence or absence only. Solid lines indicate significant clustering. Sample numbers consist of exposure time and experiment code. 62 Figure 12. Dendrogram produced by shape transformed interspecific affinities (occurrence similarities). Solid lines indicate significant clusters. 65 vii


Figure 13. Population curves for H. rapax. Mean number of individuals/sample with range bars: Top, Set A; Bottom, Set C. Figure 14. Frequency of with respect to carapace b.) Sample 12wkc; n=68. d.) Sample 20wkc; n=617. occurrence (%) of females of H. rapax length. a.) Sample 8wkc; n=19. c.) Sample 16wkc; n=276. Figure 15. Frequency of occurrence (%) of H. rapax with respect to carapace length. a.) Females sample 24wkc; n=274. b.) Females, sample 28wkc; n=521. c.) Females, sample 32wkc; n=456. 78 81 d.) Males, pooled from all samples. n=75. 83 Figure 16. Frequency of occurrence (%) of females of H. rapax from sample 32wkc with respect to carapace length and separated by reproductive status. a.) class; no apparent ovarian activity. n=189. b.) Preparatory and 2 and Copulatory ovaries active, eggs mature or immature, but always internal. n=230. c.) Copulatory females with eggs in the marsupium. n=33. d.) Copulatory females with a set of eggs developing internally and brooding juveniles in the marsupium. n=6. 85 Figure 17. of histological sections of Hargeria rapax. a. Longitudianl section centered about suture between the carapace and the first pereonal segment. b. Cross section of segment. Ov; ovary: C; Caeca: Oo; Oocyte: G e ; Germinal epithelium: Do; Developing oocyte. 89 Figure 18. Theoretical mode l flow diagram. Initially the habitat is equivalent to the substrate. When a single larva encounters the habitat, it may attempt to settle or depart pursuant to the fit of the habitat and niche hypershape. Facilitation and inhibition may induce settlement or rejection of the new arrival. The presence of thfe new species alters the habitat hypershape and such a n alteration may or may not affect the fit with respect to the niche hypershape of the subsequent larva l arrival. 101 Figure 19. Dendrogram of similarites produced by run of the computer simulation. The sample code is co mposed of a letter corresponding to the replicate and a number corresponding to the level of saturation in percent sites occupied. 109 Figure 20 The site matrix as filled by the computer simulation run The shaded area are sites occupied by species #5 (self-facilitating). Other species of particular interest (#8 and #10) are annotated with their species codes. The blank areas annotated with an M are single individuals or mixtures of the remaining 7 species. Figure 21. Species-Area curve (triangles), curve (middle curve; cir c les), and model to actual assemblage similarity curve (lower curve; boxes). r2 = 0.96, 0.80, 0.93 11 3 respectively. 144 viii


SUCCESSIONAL AND SEASONAL EVENTS IN THE FOULING OF NATURALLY OCCURRING SUBSTRATE IN A SUBTROPICAL ESTUARY by Thomas R. Cuba An Abstract Of a dissertation submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy in the Department of Marine Science in the University ofsouth Florida December 1 984 Major Professor: Norman J. Blake ix


The development of a marine fouling assemblage in Tampa Bay, Florida was traced using the shells of the bivalve mollusc, Argopecten irradians, as the fouling substrate. Sample shells for microbial analysis were collected at exposure times ranging from one hour to one week and quantitatively examined using a scanning electron microscope. Attachment of bacteria and diatoms, particulate adhesion, and adsorption of dissolved materials are identified as the earliest events in the fouling sequence. These events occur independently, at first, converging into an interdependent assemblage after two to three days. Samples used in macrofaunal analysis were taken at exposure times ranging from 2 to 32 weeks. Seasonally induced synchronization of initially asynchronous but parallel successional pathways is documented. Recruitment and initial colonization was probabilistic and early stages of development were found to be similar with respect to exposure time. Fall temperatures were found to alter development such that post=-event samples became more similar with respect to collection date. Seasonal events are identified not as successional events but rather as potential causative factors of either aspection or redirection dependent upon their magnitude. Hargeria rapax (Tanaidacea), arriving early in the sequence, became one of the structurally and numerically dominant species. Carapace length measurements and histological examinations of rapax were used to describe its reproductive status during population development. Population density was determined to be environmentally controlled while X


presence or absence was successionally controlled. A computer simulation of the initial stages of fouling was produced based upon behavioural traits as well as semi -stochastic distributional parameters of larvae. The simulation reproduces many colonization pat terns seen in both primary and secondary succession and emulates pat tern alterations brought about by changes in larval abundances. In doing so, the biotic interactions are held constant and only patch size and larval availability are altered. The flexibility and applicability of the simulation lead to the conclusion that the myriad patterns seen in early successional colonizations are the synergistic result of known parameters. Succession is seen to be deterministic while seasonal events (availability of larvae) are interpreted as disruptions of the seral sequence. Abstract approved: J .. = Professor, Marine Science Date of approval xi


GENERAL INTRODUCTION The concept of succession as described by Odum is founded on three premises: Order, control, and stability (Odum, 1971). The order is directional and predictable. The control is interactive between biological and physical factors. Stability is low in pioneer seres and high in well developed ones. This basic notion has been overlain with numerous theories and postulates of researchers attempting to identify and explain the processes and functions affecting observed changes in assemblage structures over time. The resulting generalizations have spawned contradiction and controversy leading to an overall reduction in comprehension (Mcintosh, 1980). The cornerstone of the thoery is that a successional chain of events is a predictable sequence of assemblages. This represents a great simplification while providing a manageable tenet upon which to build a functional model. In redeveloping the more expanded concept of succession, certain implications of this foundation must be considered. Each developmental sequence must be acknowledged as a discrete chain of events. Being directional and predictable, the order of the occurrence of seres in an individual developmental process is rigid. That order must, therefore, be reproducible (Odum, 1971). Control of the developmental process must be recognized as a mixture of biological effects and physical constraints (Margalef, 1968). In general, the biological factors steer the development within limits set


2 by the physical habitat. Stability must be recognized as a qualitative property, quantifications of which have inherent bias May, 1976). Arguments can be made that the time base can be anything from minutes to millennia or as a function of life span or length of reproductive cycle. The stability of the assemblage must also be seen to be radically different from, although moderately dependant upon, the stability of individual populations. In reference to the marine environment, the nebulosity of sera! stages, paucity of climax assemblages, uncertain assemblage stability, as well as questions regarding the ability of the assemblage to perpetuate itself and the apparent chronological flexibility of the postulated seres have led marine ecologists to question the frequency of the occurrence of classical succession in the sea. There is even doubt as to whether c lassical succession occurs at all (see Literature Review). Sutherland (1974) postulated that marine seres need not culminate in any climax assemblage but that they may be involved i n a shift from one stable point to another. This hypothesis was proposed in order to exp lain the apparent absence of order and climax assemblages in many marine systems. Marine ecologists have delineated only a few marine systems which exhibit a substantial number of successional attributes (Connell, 1978). Of these coral reefs and sessile epifaunal based systems (fouling) have been singled out as t h e best examples. A large part of the problem in identifying succession in the marine fouling process is the width of the size spectrum of the organisms involved. The size range of the fouling organisms presents problems in sampling which had been insurmountable in the past and denied the direct tracing of the link between the microbial sequence and the establishment


3 of the macrofauna. Operationally, the research has been divided into the study of microbial fouling and macrofouling. Methods of macrofaunal analyses extensivel y utilized various opaque subtrates. Analytical methods fall into two general categories: Non-disruptive visual censusing and the destruct! ve taking of samples. The first limits the acquired data set to organisms recognizable in vivo ---as a particular species. The data gathered in this fashion include information restricted to the larger members of the assemblage. Destruct! ve sampling results in a more complete data set but raises questions of equivalency between the developmental phases of samples taken at different exposure times. These objections were effectively nullified, with respect to fouling, by Schoener and Green (1981) who found that censusing procedures actually perturb the assemblage development Other studies of succession in the sea have been directed at assemblage development processes overlain by seasonal processes. These studies have subsequently included seasonal mollification of successional sequences but have termed the results as successional in nature instead of seasonal. The term 'seasonal succession' has resulted. This vein of reasoning has led to much of the apparent indefinability of marine seres. As new seasons arrive, the assemblage is altered in a non""predi ctable fashion. The arrival of each new season and the concomitant changes in the physical attributes of the habitat act on a different assemblage each year and thus produce different changes in the assemblage each season. Thu s if considered successional, the rigid predictability is lost and the assemblage seems to be developing and changing in a more stochastic fashion.


4 If, in the examination of the developing assemblage, the seasonal aspects can be viewed as a physical alteration causing the seral track to shift from one successional sequence to another, differing track, the process of successional development ultimately definable (Patten, 1971). well be the prime reason for the assemblages. becomes much more complex but In fact, this seasonal shifting may paucity of definable climactic The present study, based on the hypothesis that seasonal events do not qualify as successional events, is directed primarily at resolving the difference between the two and their effects on the development of a sessile assemblage. To accomplish this goal, the nature of the developmental process of a fouling assemblage as it occurs on a natural substrate was traced and the results analysed with respect to exposure time and temperature, The development of microbial populations and the subsequent colonization by macrofauna is traced on like substrate using the joint methodology of Scanning Electron Microscopy and total destructive censusing. The latter yielded a data set with members from protozoa to vertebrata. The thoroughness of sampling and the detail of the data set satisfies the objectio ns to fouling studies as incomplete or superficial (Schoener, 1974; Mcintosh, 1980). Multiple replicates are used primarily to determine sampling sufficiency but also have been examined as a tool for determining the degree of reproducability of the seral development. Classification of events as successional or physically mediated phenomena was attempted throughout.


5 LITERATURE REVIEW H.C Cowles and F.E. Clements fabricated successional hypotheses into formal ecological theory in the early 1900's and as yet it remains one of the most confounded of ecological concepts." (Mcintosh, 1980). The two major sources of disagreement stem from the deterministic as opposed to the stochastic viewpoints and the inclusion of seasonally controlled events in perceived successional pathways. (1934), Hewatt ( 193 5) Newcombe (1935), Aleem (1957), a nd Haines and Maurer (1980) all found substantial evidence for deterministic succession but others (Sutherland, 1 974; Dean, 1977 ; Osman, 1977; Sutherland and Karlson, 1 977; Smedes 1979) recognized no evidence for classical succession and often suggested alterations in the theory to account for their local observations. McDougall (1943) went beyond the alteration of theory in stating "There is doubt whether succession leading to climax occurs in marine communities." The conflict arises from attempts at global generalization prior to sufficient comprehension of local phenomena. Perhaps the most pertinent local inconsistency regards the problems presented to the successional theorist by temporal variances in superficially mature assemblages. Variations exhibited by established assemblages have been noted by McDougall ( 1943) and others. Fitting these observations of seasonal variance into a generalized successional scheme has contributed to the rift among the 'invisibl e colleges' (Mcintosh, 1980) of successional


6 theory. The Gleasonian approach to succession is often favored in this instance because it allows, with a minimal amount of interpretational dilemma, the effects of the changing seasons to be easily categorized as successional events (Mcintosh, 1980). Knatz ( 1978) and Lee and McAlice (1979) interpreted seasonal changes in planktonic copepod assemblage structure as successional progressions. This type of approach, while allowing easy interpretation, leads to the definition of successional sequences which are seldom reproducible and legitimizes the dominance of stochastic processes in successional function. Adherents to the Clementsian concepts object strongly to non-deterministic developmental schemes. Sheer ( 1945) and Aleem ( 1957) both felt that successional and seasonal processes should be recognized as discrete mechanisms. "The basic problem in the developmental sequence of communi ties in a limited environment is that of distinguishing between seasonal progression and true succession." (Sheer, 1945). Wells (1961) characterized the oyster reef fauna of North Carolina as having a year round fauna overlain by a pattern of subdominant seasonal transients. Calder and Brehmer (1967) found a seasonally cyclic pattern in the hydrozoan members of a fouling assemblage in Virginia waters. This alteration was reproduced yearly and they clearly state that the changes were not successional but due to alterations in the thermal regime. Sutherland and Karlson (1977) and Mook (1983) described a seasonally tuned sloughing of Styella plicata populations. The respecti ve authors classified these events as seasonal and successional. Thus the same phenomenon can be classified differently depending upon the


1 theoretical bent of the investigator. Shelford ( 1935) was well aware of seasonal variation in assemblage structure and suggested the use of the term 'aspection' when the alteration was relatively uniform from year to year. Glasser ( 1982) clearly agreed with the separation of season and succession when he wrote "Seasonal change in species composition need not be succesional (1934, 1935) described a cyclic alteration wherein Balanus balanoides is overgrown by Mytilus edulis. The latter is then heavily preyed upon by Thais lapillus bringing about the of the barnacle as the dominant species. This cycle is not true aspection as it is not seasonally attuned. Alterations in habitat other than thermal are at times closely associated with seasonal change. Santos and Simon (1980) examined the yearly recolonization of Hillsboro Bay, Florida, following recurring summer hypoxic disturbances. The disturbance itself was not seasonal, but rather seasonally induced just as ice formation is in higher latitudes. Seasonality also affects developing assemblages. Kawahara ( 1965) qualitatively observed that the time to impossibility of useful visual census techniques in fouling studies is appreciably shorter in summer months. Schoener (1974) subjected published data to analysis using the MacArthur and Wilson (1963) model and found that the colonization curves varied significantly with the season of immersion. In fouling studies the season of substrate emplacement can significantly affect the developmental sequence observed. McDougall (1943), Sutherland and Karlson (1973), Mook (1981), and Keough (1983) all noted that larvae of sessile epifauna occurred seasonally in the plankton


8 and that this greatly affected the composition of the early assemblage. McDougall (1943), working on pilings in the vicinity of Beaufort, North Carolina, found the effect significant enough to warrant delineation of the breeding season of many of the organisms prior to reaching any conclusions. Kitching (1937) and Kawahara (1965) found definite differences in the biofouling assemblages concomittant with the season of immersion; the latter wrote that the recolonization process was haphazard and determined largely by the order of arrival of larvae and spores. Coe and Allen ( 1937) state that the available larvae change the pattern of the initial assemblage with regard to the density of individuals. Osman (1977) states that the initial assemblage is a direct reflection of larval abundance. Sutherland and Karlson (1977) found that larval recruitment created different patterns in assemblage development both within and between years. They also reported that plates submerged at a given date always developed similarly. Their final conclusions included the statement that develpment is not orderly and directional because larval recruitment is variable. Osman (1977) found that the variance in larval settlement affected the resulting assemblage and included the individual history of the substrate as a factor in successional theory. Sutherland (1974), in basic agreement with Osman, argues that each of these initial assemblages will ultimately result in differing stable assemblages and that "A separate model must be built for each stable point." Dean (1977) and S medes (1979) both recorded larval variability in fouling studies in Delaware. Dean concluded that "Succession can only be described as a stochastic change in community structure with time that is


9 chiefly controlled by competitive and preemptive mechanisms." The work of Cairns and Henebry ( 1982) also supports stochastic control theory. Working with protozoa they postulated that many of these organisms occupy the same niche and that the initial assemblage structure is stochastic, elimination not occurring until densities reach levels where competition becomes the dominant structural force. Thus, they say, the population occupying any given niche at any given time is the result of historical accident in combination with the environment. A modest compromise between the stochastic and deterministic models was proposed by Paine and Levin (1981). Working on succession in a Mytilus californianus dominated area of Washington state, they noted a variance in larval availability but were faced with the same resulting assemblage. They observe that random factors of environment and larval densities become magnified as the early, stochastically determined, species influence later colonization but that these factors are eventually, at that site, competitively reduced and Mytilus californianus eventually dominates the assemblage. Therefore, the site exhibits a probabilistic development, stochastic within a certain boundary in its initial stages and ultimately resulting in a predictable, stable, M. californianus dominated assemblage. Excellent agreement between their probabilistic model of patch recovery and observed recovery patterns lends a great deal of credence to this proposition as predictive models are rarely successful in the field of ecological succession. The season itself is not necessarily the controlling factor in larval abundance. Warm winters resulting in reduced phytoplankton blooms subsequently reduce the food supply of planktotrophi c larvae. The subsequent decrease in the numbers of larvae available alters the initial


10 development of sessile assemblages (Visscher, 1927). By that pathway, it is apparent that the developmental sequence of sessile epifaunal assembalges is inextricably linked to planktonic ecosystems as well as physical variablity. Although the research cited above demonstrates that in many systems (much of the above work was carried out in temperate zones) the larval availbility is at least partially stochastic, there is much other work that demonstrates a certain predictablity in larval and algal colonization. Northcraft (1948) and Aleem (1957) found that a gelatinous coating of diatoms always precedes macroalgal colonization. Fager (1971), with regard to the same observation, felt that this was a purely physical event suggesting that the spores became trapped in the slime layer. Horn (1974) following a Markovian procedure presents arguments for stochastic individua l by individua l replacement in secondary systems and it can be argued that the same type of process functions, to a degree, in primary systems. The distribution of algal spores, however is more haphazard than that of the larvae of benthic invertebrates. While planktonic, the larvae exhibit behavioral traits which affect their distribution (Thorson, 1966; Chia, 1977). The behavior is triggered mos t often by thermal, haline, tactile, gravimetric and photic parameters. Larvae in general prefer rough to smooth surfaces and dark to light colored surfaces (Visscher, 1927; Crisp and Barnes, 1954). The effects of this, and other behaviors, are seen in colonization sequences. Crisp a nd Ryland (1960) determined that barnacle cyprids settle more readily when presented with a microbially fouled surface. Scheer (1945) found that bryozoa prefer to settle on surfaces covered with diatoms. Rocks immersed in summer were heavily colonized by this taxa the next


11 winter but rocks made available in winter were not, even though the larvae were present. The same author also noted that Mytilus only settles where there are established populations of either bryozoans or Styella. Standing ( 1976) deduced that the bryozoa tactily mimic the byssal threads of the mussel. The mussel larvae do tend to recruit to existing beds (Paine and Levin, 1981) and 1977). structural mimics (Dean, Dean ( 1977) and Smedes ( 1 97 9) state that the diversity of motile species in a sessile epifaunal assemblage increases with increases in biogenic structure. Wells et al. (1964); Haines and Maurer (1980); and Gallagher et al. (1983) all cite instances where established species which contribute structure (habitat diversity) facilitate the establishment of other members of the assemblage. In a classical example of the experimental removal technique, Standing (1976) croppe d Obelia dichotoma, an inhibitor of Balanus crenatus and facilitator of Ascidia ceratodes and found that the population of the barnacle increased and the ascidian settled less frequently. Many larvae cue on factors other than structure in their choice of habitat. O n e the more obvious cues is the availability of food (Zobell and Feltham, 1 937) Harpacticoids feeding on diatoms and bacteria will not immigrate if these energy sources are lacking (Brown and Sibert, 1977). The same has been proposed for macrofauna! larvae (Jones, 1950; Scheltema, 1974). Microbes have also been shown to be involved in site selection behavior when they are not utilised as an energy source. Microciona prolifera requires a microbially fouled surface for settlement (Calder and Brehmer ( 1967). Bugula neritina exhibits site selectivity behavior based on the synergistic result of wettability and


12 microbial populations (Mihm, et al., 1981). The larvae of the gregarious polychaete Spirorbis borealis was found to prefer microbially fouled surfaces to clean ones ( 1951, 1953a, 1953b; Wisely, 19 60 ; DeSilva, 1962). Meadows and Williams (1963) refined the knowledge finding that these larvae exhibited a preference for surfaces having a mixture of bacteria and the diatom Navicula. Nelson (1979) noted that the gregarious larvae of Janua brasiliensis require a diatom mat to induce settling. The more recent work of Kirchman, et al (1982) determined that these larvae prefer a surface which has a mixed population of bacteria and diatoms. Monospecific bacterial populations were not as attractive to the larvae as were mixed populations and the diatom Nitzschia was actually found to inhibit settling. These authors felt that the cue was chemical and could be found either in the surface chemistry of the microbes or in their exudates. The existence of chemical cues in larval site selection is not novel. Standing (1976) found nudibranchs to respond to chemicals found in the bryozoa upon which they feed. Crisp a nd Meadows (1963) found that Balanus balanoides respondes to an arthropodin class chemical found in cirri peds Haliotis and the cuticle of other arthropods larvae exposed to acid (see Lewis, 1977). (an algal extract) immediatley undergo settlement and metamorphosis (Morse, et al., 1979). Eckelbarger (1977) and others (see Crisp, 197 4) have identified the sensory structure involved in site selection in Sabellariid larvae. The knowledge of these cues is extensive (Gray, 1966; Crisp, 1973; Grant and Mackie, 197 4; Brewer, 197 8; Neumann, 1979; see Meadows and Campbell, 1972; Crisp, 1974; Scheltema, 1974; Chia and Bickell, 1977;


13 Hadfield, 1977, for reviews) and represents precise knowledge of deterministic relationships often overlooked in successional studies. Many studies have recognized the importance of microbial fouling to epifaunal and macroalgal assembalge development (Hewatt, 1935 ; Zebell and Allen, 1935; Thorson, 1957; Calder and Brehmer, 1967). The evidence for succession is not limited to relationships but can also be found within the microbial fouling sequence. The majority of the literature, when the distinction is made, states that bacterial colonization precedes the establishment of microalgal and protozoan populations (Aleem, 1957; Corpe, 197 3 and 1974; Caron and Sieburth, 1981a). Sieburth (1968) casts algae in the role of inhibitor due to their toxic exudates. Investigations into the process of bacterial colonization have shown that the sequence occurs in three phases (Floodgate, 1972; Corpe, 1974 ) First the bacteria are brought near the surface by random motion functions. Adsorption occurs and the bacteria are in a state of reversible sorption. During this phase the bacterium can either initiate irreversible sorption (phase two) or, should the substrate prove uns atisfactory, it may depart the surface. Irreversible sorption takes place when the bacterium becomes firmly attached by pili, microfimbrils, or exopolymers. The last phase growth and multiplication is self explanatory. The physiochemistry of the attachment process has now been compiled into the DLVO theory (Jones and Isaacson, 1983). One item conspicuously underaddressed in the literature is the fact that bacteria cannot be considered as a homologous group while maintaining the basic tenets of bacterial taxonomy. The bacteria microbially preparing the surface are undoubdtedly of many diverse


1 4 species and to consider them as a single entity represents a gross oversimplification. Each species metabolizes different compounds and produces different waste products and exudates. That bacterial taxonomy leans so heavily on the metabolic function of the organism reinforces the fact that larvae are presented with a highly variable suite of chemical stimuli in their search for an attachment site. Corpe (1974) apparently stands alone in his observation that rods were replaced by Caulobacter and Saprospira in the maturation of the bacterial assemblage. As stated above, in phase one of the bacterial colonization process, the bacterium exhibits an instinctive choice in proceeding to phase two. The biological reactions leading to irreversible sorption depends a great deal on the character of the surface offered. Zebell and Anderson (1936), Zebell (1943), Kriss (1959), Fenchel (1970), Jannasch and Pritchard (1972), and Goulder, et al. ( 1981) hav e noted and described increased bacterial activity at solid liquid interfaces. This is largely, if not totally, du e to the adsorbed organic material on the surface. Corpe (1973) realized that the bacteria were utilizing this adsorbed layer and stated that the diatoms a nd protozoa did not colonize the surface until after the bacteria had become established. Baier e t al. (1968) could find no instance where bacterial colonization preceded the adsorption of dissolved organic materials. Fletcher ( 1980), in treating the problem of charge distribution (most surfaces as well as most colonizing bacteria have net negative surface charges) notes that, due to organic adsorption, the surface the bacterium "sees" is radically different from the one immersed (Niehoff and Loeb, 1973, 1974). Marshall and Cruickshank ( 197 3) describe another method noting that many rod shaped bacteria are actually bipolar, bacterial adsorption occurring as


15 the result of the hydrophobic pole in the vicinity the solid""liquid interface. The discrete charge properties of the substrate also influence the character of the organic microlayer which in turn influen-ces the character of the colonizing organisms (Dexter etal, 1975; Fletcher and Loeb, 1979). The microlayer by itself, however, is not sufficient for larval colonization (Mihm etal, 1981); the proper microbes are required.


16 THE STUDY SITE The Study site is located near the mouth of Tampa Bay, 10 m west of Bonne Fortune Key, Ft. DeSoto County Park, Pinellas County, Florida (fig. 1). The embayment formed by that key, St. Jean Key and a portion of Mullet Key (all now joined by road fill) is a roughly rectangular area 1.5 x 0.5 Km, and only a few meters deep at its deepest point. The eastern portion (of St. Jean Key, largely denuded of trees and covered only by thin grass, constitutes a major watershed. The only hard substrate of considerable size is an intertidal oyster bar along the eastern shore of the key (approximately 100 m2 ) but smaller beds and various concrete and limestone blocks can be found scattered throughout the cove. The remainder of this shoreline is composed of moderately clean to muddy sand. The segment of Mullet Key facing the embayment is well covered with mangroves, cabbage palm, and sparse forest. Bonne Fortune Key is covered by grass in its southern third. The remainder is used as a nursery by the park service but is ringed about the shoreline by a 10 to 30 m wide band of mangrove and forest. The bottom of the cove is covered with thick growths of seagrasses (largely Ruppia maritima) growing in a sandy mud bottom (0.3 to 0 5 m thick) which overlies a harder, sandy layer. The area is closed to clam and oyster fishing (there is a sewage treatment plant at the NE tip of Madelaine Key) but remains a popular spot for obtaining blue crab and


Figure 1. The study site. Flag indicates the location of the sampling apparatus. Bridges have been omitted for clarity.


ST. PETERSBURG : . ;. : . MADELAINE . : : : : "( .:.;. '"\ .= ...... : ;.:. : ... .. . .. : ... BONNE FORTUNE KEY :. : ; . 18


19 mullet. The location for the emplacement of samples was not randomly chosen. The frame was positioned at the shoreward boundary of seagrass beds off of Bonne Fortune Key in 0.3 to 1.3 m of water. A small oyster bed (10 m2 ) was located 15 m to the north in an otherwise sandy subtidal area. To the south, both intertidally and subtidally, were several small pieces of concrete bridge debris. The concrete, as well as the roots of the red mangroves were covered with barnacles and only a few oysters while the oyster bed was relatively free of barnacles. Libinia dubia and Callinectes sapidus (Crustacea: Reptantia) were common in the area as were the Molluscs Melongena corona, Fasciolaria lilium, and Modulus modulus (Gastropoda) Bursatella leachii (Gastropoda) swarms in this area during the winter breeding season laying strings of eggs in the grasses. The most commonly observed fishes were Mugil cephalus, Eucinostomus gula, and Archosargus probatocephalus. With the closure of the channels between Mullet Key and the two smaller keys, circulation has been restricted to tidal flux and wind generated flow. The area can be as much as 60 to 75% exposed during winter storms combined with extreme low tides. The duration of such exposure is short and only occurs times each year. Salinity ranges from a low of around 29 ppt to highs near 34 ppt. Temperatures were recorded as low as 13C in January and as high as 40C in August (fig. 2).


Figure 2. Salintiy and temperature from 27F-May .. .qg78 to 3;.:.Mar;.:1979, a Salinity in parts per thousand. b. Temperature in degrees C. Range bars indicate high and low for previous 24 hrs.


SALINITY (%o) 3 MAR 3 FEB 6 JAN 9 DEC 11 NOV 14 OCT 16 9 2 SEP 20 1 9 12 5 AUG 29 22 1 5 8 1 JUL 2 4 1 7 10 3 JUN 1-----:'----1 . TEMPERATURE (C) 2 1 3 MAR 3 FEB 6 JAN 9 DEC 1 1 NOV 1 4 OCT 1 6 9 2SEP 26 19 12 SAUG 29 22 15 s 1 JlU1l 2'41 n lllll ::ll JlUJ!il


THE INITIAL DEVELOPMENT OF A MARINE FOULING ASSEMBLAGE ON A NATURAL SUBSTRATE IN A SUBTROPICAL ESTUARY. 22 The degree of microbial fouling that a submerged surface has undergone very strongly affects the subsequent fouling of that surface by members of the meiofauna and larvae of the macrofauna. Caron and Sieburth (1981a) term microbial fouling as a simple precursor of macrofaouling, while Knight..:Jones (1951, 1953a) and others place microbial fouling in the category of a true successional requirement. Many particulars in the relationships between various phases of the fouling sequence have been established (see Wood, 1967; Bitton and Marshall, 1979). Much of this work has been accomplished using glass slides as a fouling substrate (Zobell and Allen, 1935, and others) which allows direct examination of the microbial assemblage as well as permitting direct bacterial counts to be made. Staining techniques such as the use of in conjunction with epifluorescent microscopy have further enhanced the precision of this method and increased the reliability of the results. The advent of scanning electron microscopy has permitted the direct examination of opaque substrates and enabled direct counts to be made on manmade or naturally occurring materials. Dempsey ( 1981 a, b) examined the microbial assemblages occurring on several types of anti-fouling paints and Sieburth (1975) examined established assemblages on a variety of substrates ranging from rocks to various living materials such as algae and the shells of living gastropods. This study traces the


23 development of an assemblage on a naturally occurring substrate from initial colonization through the initiation of macrofouling. Methods The shells of freshly collected live bay scallops (Argopecten irradians Lamarck 1819) were used as the fouling substrate and examined using a s canning electron microscope which enables an analytical technique analogous to previous studies using glass slides and light microscopy. Two hundred and eight 1 cm2 chips were cut from the shells and heated to 500 C for 30 min to remove organic residues. These chips were then cleaned with a brush and glued to plexiglass strips with silicone sealant. Four strips were bolted to a frame of PVC piping which held the samples horizontally and approximately 0.33 m above the bottom of the bay. At the start of each sampling period 70 chips were simultaneously emplaced i n a shallow, low energy, inlet at Mullet Key located near the mouth of Tampa Bay, Florida. Water over the samples varied with the tides from 20 to 100 em. The four sampling periods (labelled A,B,C, and D) began 28 days apart, commencing on May 27, 1978. Salinity and temperature measurements were taken using an optical refractometer and a hand held thermometer. The 70 chips of each period were divided into 7 samples and 10 replicates each. Samples were collected after exposure times of 1, 5, 24, 48, 72, 96, and 168 hrs., fixed in a 3% buffered glutaraldehyde solution and prepared for scanning electron microscopy (SEM).


24 Subsampling of shell chips was carried out at a magnification of 5000 x and transects were oriented perpendicular to the shell rugosity. The total area examined per shell was 0.259 mm2 and replicates were examined until a total area per sample of 1.036 mm2 had been quantified. These areas were controlled by the results of Species-Area analyses (Preston, 1962) Liquids used in fixation and dehydration were filtered to retain deciduous meiofauna (0. 067 mm filter). These as well as the microalgae were identified to species where possible. Similarities in the developmental process of the replicate sets were examined using the calculation (Bray and Curtis, 1957) and constructing a dendrogram using group averaging ordination (Cody 1974). Other shells were used to examine the process of adsorption of dissolved matter. These shells were prepared as before and divided into controls and sample chips. The sample chips were placed in a flask of filtered seawater and let stand for 24 hrs. The controls were prepared and processed for the SEM, as were the samples, and examined immediately. Results The numerically dominant forms throughout the course of this study were the bacteria and the diatoms (fig. 3 and 4). Colonization was initiated by diatoms and rod shaped (0 5 x 0 2 bacteria. As particulate material accumulated, the rods were largely replaced by ovoid (0. 6 x 0.3 bacteria. This replacement was slower on the ridges of the scallop's shell where particulate accumulation was slower. These areas exhibited the only case of colony formation (100% cover by rods)


Figure 3. Colonization curves for microalgae (*) and bacteria (o) Average of 4 sample sets.


26 NO. INDIVIDUALS/cm2 01 ..... 01 ..... 01 X X X X X ..... ..... ..... ..... ..... 0 0 0 0 0 O't O't (!) (!) ..... 01 m X -o 1\) 0 0 (/) c :0 m -i s: m (X) 0 ::r .., (/) 0


Figure 4. Pennate diatom, bacteria, and microdetritus on surface of scallop shell immersed in seawater for 48 hrs.


J .., co (.J -... -_ .. .i) co (. J


29 even though many bacteria in other areas were seen to be in a state of division. Initial development consisted of the attachment of bacteria and diatoms. The bacteria were randomly dispersed over the surface of the shell while the diatoms showed a marked preference (85 to 95%) for the valleys. Amorphous particulate matter, which appeared to be organic in nature, was observed adhering to the peaks of the small ridges formed by the growth lines of the scallops' s shell. These particles were similar to those observed on the shells which had been exposed only to filtered seawater. Ninety eight species of diatoms, mostly Amphora, Navicula, and Nitzschia ssp., were encountered as well as 7 species of dinoflagellates. The diatoms and dinoflagellates reached a stable density at 5 hrs. at approximately x 105 individuals/cm2 The species composition of this diatom:-dinoflagellate assemblage continue to change, however, becoming more diverse as exposure time lengthened. It was these 105 species which were used to determine similarities and to construct the dendrogram. Meiofauna (table 1) were not encountered until the hr. sample and the Harpacticoids contained several undescribed species (S. Bell, Pers. com.). The majority of these were adult forms while the rest of the meiofauna were largely juveniles. Although the numbers of fauna followed an increasing trend, the similarities between replicates and samples were low and the colonization was extremely patchy. Detritus and sand grains began to accumulate in the valleys in the 24 and 48 hr. samples but did not cover the ridges until the 96 hr. samples. By the 168 hr. samples, sheets of this material had occassionally sloughed off the ridges while leaving material


30 Table 1 Meiofa unal Colonization. Sum of counts for a l l samples Exposure Time (hrs.) 5 24 48 72 96 1 68 Harpacticoida 1 0 4 15 16 54 Polychaeta 4 7 7 1 4 Nematoda 2 5 4 8 Tanai dacea 5 2 Foraminifera 5 1 0 Ostracoda 2 Miscellaneous 2 ,.... 3 Total 0 0 1 0 1 5 30 38 93


31 that had accumulated in the valleys in place. Blue-green alage, fungi, and protozoans were not quantified due to limitations in the method. The developmental process was followed only through the 72 hr. samples. Longer exposures produced samples which were so densely packed with fouling material that enumerations were invalidated. Species distribution of diatoms and dinoflagellates in the early samples of the developing assemblage exhibited a large degree of patchiness (replicate to replicate similarities <0.3) and sample to sample similarities were not significant (figs. 5 and 6). The species colonization curve shows that although each of the samples had few total species of diatoms and dinoflagellates present (compared to 105 total species), 57% (60) of the species in the final assemblage had occurred by the 5 hr. sample and 92% (97) had been encountered by the 48 hr. sample period. These low exposure-time samples have similar numbers of species per replicate (both within and between samples) and a similar density of individuals (fig. 3) but dissimilar species composition (fig. 6). The dendrogram of Bray-Curtis similarities (fig. 6) clearly shows that the 1 and 5 hr. samples were nearly all dissimilar at the 0.5 level (1C and 10 being the exception). the 0.5 level while One of the 24 hr. samples remained unclustered at the others intermediate to long exposure times. fell into clusters of samples of Conversely, all the 48 and 72 hr. samples fell into groups at levels greater than 0.5. Close examination of the groups reveals that sample series A was isolated from the other three samples sets (B, C, and D), which at long exposure times formed integrated groups. Salinity ranged from 30 to 33 ppt. and daytime


33 40 .,..,. 100 .,. .,. .,. *" 90 80 0 30 ' 0 c 3: 0 / -c / r en / 70 )> UJ / --f (.) / < w / m a. r/0 60 en 0 20 z b 50 I 40 30 1 5 24 48 72 EXPOSURE TIME (hrs)


Figure 6. Clusters formed by dendrogramming sample similar! ties where solid line (commencing at 0.5) indicates significant clustering. Sample numbers consist of exposure time in hrs (numeric) and sample series (Cyrillic).


SIMILARITY ..... 1\) (.\) 01 0) -....! (X) co Ovoo l 72C -----------------------1o o0o j 0 o oi------l o 4 8 C 0 0 o o;.... -----1 : 0o0 24C ..-------------------------; 0 0 Lf ) : 0]_7_2 -8 ___, fl L ______________________________ l : o: 7 2 0 0 0 <>I r : (!!I! 488 I L J ::: ::;:;:;: 2 4 8 I I 480 I ffl r ----tl r o(Y.'l\1\..._2_4_0---1 I : L I : : 72A I I I 4 8 A I I SA I r-; 24A I I I 1 I r 1 sc I I I ___ -: 1 : 1C I : : l 10 I &... --, I I 1 r-50 I 1 I 1 I I r58 I I I I L --, 18 I 1----I 1A L _ ------w \Jl


36 temperatures averaged 30C (diurnal range, 25-42C). Lowest temperatures were recorded in the initial sampling period (A) while the others displayed typically high summer temperatures. Discussion The initiation of the fouling process by the attachment of bacteria and diatoms seen here is in accordance with the results of others (Zobell and Allen, 1935; Wood, 1967; Corpe, 1974) who have identified these two groups as the pioneer taxa in substrate colonization. Despite the rapid settlement of large numbers of individuals of these groups, their presence represented a coverage of only 3% in the 1 hr. sample. Biotic coverage never exceeded 10% of the entire chip and bacteria were only rarely associated with diatoms. The fact that bacteria often occurred within 20 J..lm of diatoms but rarely on them, suggests the presence of a localized antibacterial agent on the surface of the diatoms (see Droop and Elson, 1966; Sieburth, 1 968, 1975). In the rare cases where diatoms had bacteria on their surfaces, the bacterial count was high suggesting that the diatom had recently died and was being used as a nutrient source by the bacteria. This is further supported by the observation that no single species of diatom was any more prone to bacterial colonization than any other species and broken diatoms frustules devoid of protoplasm were no more or less likely to have bacteria on them than any other inorganic particle. The observed affinity of both diatoms and larger particulates for attachment in the valleys of the shells is the result of mi cre-'-eddi es caused by the morphology of the scallop' s shell. These flow patterns


37 reduce the current velocity allowing the larger particles to settle Single bacteria in the observed size range, approach a state of colloidal suspension (Danials, 1979), and their initial distribution pattern is not effected by this system. The absence of diatoms on the ridges, where bacteria were common, implies that the presence of diatoms is not a requirement for bacterial fouling (Sieburth, 1975). The results of Caron and Sieburth (1981a) also support this implication but it is possible that some species of bacteria require diatomaceous fouling. These data did not allow the determination of any of these specific relationships but it remains apparent that the attachment of diatoms and of bacteria are not successionally related and it is only their abundance in the water column and their growth after settling that determines which group is numerically dominant. The occurrence of adsorbed particulate matter (less than 0.45 ).lm) was uniformly distributed on both ridges and valleys. The concentration of the particles at the peaks of the scallop's growth ridges is unexplained but probably related to electrostatic charges (Danials, 1979). It is uncertain whether or not this adsorption is required for bacterial or diatomaceous fouling, but the presence of organics certainly enhances bacterial if not diatomaceous fouling (Cviic, 1953; Renn, 1964; Young and Mitchell, 1972). Free bacteria are abundant in the estuarine water column and the microenvironment produced by the adsorbed organic material is conducive to bacterial adsorption and growth (Floodgate, 1972; Jannasch, 1978). Dissolved organic matter is also abundant in the estuarine environment and readily adsorbs at any solid'-liquid interface (Bitton and Marshall, 1979). The adsorption of these organics is a physiochemical process which, as our control shows, does not require


38 bacterial or diatomaceous mediation. Accumulation of larger (1 00 particulate matter depressed bacterial populations (see Zebell, 1946; Oppenheimer, 1960; Sieburth, 1976) and colonies were formed only on the ridges where the particulates were mechanically excluded in the initial stages. These colonies (100% cover) were no longer observed once particulates had covered the ridges. Although cell morphology is a poor criterion to use in bacterial taxonomy, the colonies on the clear shell were composed of small rods while the bacteria in the more heavily fouled areas were more commonly ovoid and it is a fair supposition that the bacteria that did well on the open flat substrate are of an entirely different type than those found among the accumulated particulate matter. The unknown amount of grazing pressure exerted by the protozoans (Zebell and Allen, 1935, found large numbers of protozoans) may have had a role in depressing bacterial populations after 48 hrs. It cannot be assumed that there is only one or two mechanisms controlling these bacterial populations and there is much work left to be done in this area. Further mechanisms in controlling bacterial and diatomaceous populations are probably concommitant with the changes composition which occurred after 24 hrs. and before 48 hrs. time the species composition of the in species During this assemblage began to stabilize. Concurrently, particulates accumulated, fungi, blue-green algae, and protozoans occurred and meiofauna settled or immigrated, initiating grazing pressures. The fungi played an active role in substrate stabilization as they were observed to be attached to both the shell and particles such as sand grains. The assemblage stabilization between 24 and 48 hrs. is reflected in


the clustering pattern of the dendrogram (fig. 6). 39 The 1 and 5 hr samples joining each other at 0.3 and 0.5 reflect variability in the initial development pattern. High similarities (>0. 5) among the long exposure time samples (48 to 72 hrs.) indicate that mechanisms such as growth, reproduction, survival, and further recruitment have been operating to bring the samples closer together. Replicate to replicate similarities, however, remained low indicating continued distributional patchiness. The distribution of groupings of 24 hr. samples about the 0.5 level mark this as the truning point in assemblage development. The integration of sample sets B, C, and D in the clusters reflects a convergence of assemblage composition while the insulation of sample set A is probably a result of lower temperatures recorded during its development (fig. 6). The change from dissimilar to similar assemblages during the 24 to 48 hr. sampling period occurs concurrently with the influx of meiofauna and the other groups. Whether the stabilization of the assemblage is a result of factors associated with these new organisms or some other factor is impossible to determine from these data. Although meiofaunal grazing undoubtedly has an effect on the species composition of the assemblage, the exact parameters remain obscure. The meiofauna (table 1) exhibited an overall progression but replicate to replicate similarities were low. The patchiness observed in the distribution of harpacticoids has been shown to be normal in epibenthic populations (Coull, 1973). The harpacticoids were first encountered as adults and represent recruitment. The other fauna, nearly all juveniles, represent settlement or recruitment of juveniles. The presence of organics, bacteria, and diatoms has been shown to enhance


further colonization ( Zobell and Allen, 1935; Corpe, 197-4; Coull 1973) Hopkins ( 1977) average) of the zooplankton in Tampa Bay -40 Crisp and Ryland, 1960; found that 19% (yearly were larvae of benthic invertebrates. The highest concentrations occurred in spring and summer and the probability that all the faunal forms were available in the water column but had not settled (nor immigrated) prior to the 24 to 48 hr. period further supports the hypothesis that microbial fouling is a required successional step for colonization by larger organisms (Caron and Sieburth, 1981b; Knight<-'Jones, 1951, 1953a). In opposition, it must be noted that few (5 to 186) of the individuals of the meiofauna collected here were filter feeders and the majority of them depended on the substrate or its inhabitants for food. The requirements for a prepared surface in substrate dependent feeders may not be as preclusive for those larvae which will develop into macrof aunal filter feeders. Some species of barnacle are not limited to settlement on a microfouled substrate even though they may prefer such a surface (Hudon et al., 1983). The latter larvae may equate microbial fouling with habitat stability, whereas, the former require it as a food source. Despite the initial distributional variability and the conclusion reached here that the bacteria and diatoms are not successionally related, it is remarkable that after only 48 to 72 hrs. the assemblages that were sampled up to two months apart had converged to show a high degree of similarity. The temporal occurrences of the physical events such as particle accumulation and organic adsorption were also equivalent. Sample sets B, C, and D reflect a typical summer development process while sample set A shows that the other seasons will exhibit similar development patterns with a varied assemblage structure.


41 Summary Fouling is initiated by the adsorption of dissolved organic matterial followed by the adhesion of bacteria and diatoms. Neither the bacteria nor the diatoms require the presence of the other. The microalgal assemblage, dissimilar among replicates at first, converges and stabilizes prior to 48 hrs. of exposure. Among sample series commencing up to 2 months apart, the initially dissimilar microalgal assemblages converge to like assemblages at between 48 and 72 hrs. of exposure. The above microbial preparation of the habitat is required for the settlement or immigration of larvae and juveniles of meiofauna and macrofauna.


THE DEVELOPMENTAL PATHWAY OF A MARINE FOULING ASSEMBLAGE AS AFFECTED BY SEASONAL AND SUCCESSIONAL EVENTS 42 Marine sessile epibenthic assemblages occurring on newly available substrate have exhibited developmental pathways which contradict the predictability and reproducibility components of successional theory. McDougall (1943), Scheer (1945), Kawahara (1965), Sutherland and Karslon (1973, 1977), Dean (1977), Osman (1977), Smedes (1979), and Keough (1983) have all reported assemblages, developing at the same physical locality, which have become less and less similar with time or which have been dissimilar from the point of substrate immersion. This d ivergence was most pronounced when comparing assemblages on substrates immersed on different dates. In these studies the assemblages followed distinctly different developmental pathways leading the investigators to conclude that the direction an assemblage takes is dependent upon random factors. Not all colonization studies have resulted in divergence. Hewatt (1935), Newcombe (1935), Aleem (1957), Haines and Maurer (1980), Sutherland (1980), Mook (1983), have all examined assemblages whi c h develop along nearly identical pathways resulting in highly similar and stable assemblages. Much of the similarity in early colonization of habitats within a specific time frame is due to the site selectivity of the larvae involved. Most larvae of sessile marine invertebrates and non-vertebrate chordates are highly selective in their choice of habitat (see Meadows and Cambell, 1972 ; Crisp, 197 4; Scheltema, 197 4 ; Thorson, 1966 for reviews).


43 Paine and Levin (1981) proposed and successfully tested a model of .secondary succession in an intertidal Mytilus assemblage which represents a modest compromise between the stochastic and deterministic views. They noted that the larval influx has a tendency to randomness but that competition eventually drives their system to a locally unique endpoint; the Mytilus dominated assemblage. Their model, termed probabilistic, allows for larval variance and multiple developmental pathways while maintaining a single stable end product. Sutherland (1974), in developing the Multiple Stable Point (MSP) hypothesis, combined the observed stochastic initiation of development with determinism stating that each of the stable points develop along predictable (once initiated) but different pathways and must be modelled separately. In the MSP model, only the initial direction (in primary or secondary succession) is subject to chance. Despite variations in larval recruitment patterns, developing assemblages which are temporally synchronized develop similarly (Sutherland and Karlson, 1973, 1977; Sutherland, 1980; Mook, 1981; Schoener and Green, 1981). The developmental pathways exhibited by assemblages colonizing identical habitats with identical temporal and spatial orientation are parallel. The similarity of the assemblages remains high even though the season (and faunal availability) changes. The unpredictability and divergence, then, can be traced directly to temporal variances in larval availability resulting in different initial points for development. The larval pool of Tampa Bay, Florida, fluctuates only mildly during summer months and most larvae are present for most of that time (Hopkins, 1977). This larval constancy, in combination with fairly constant


44 summertime physical parameters enabled an experimental scheme which .separated the effect of a change in season from that of a difference in the time of substrate immersion. The relative constancy of the habitat and fauna removes the requirement of temporal synchrony for parallel development (Mook, 1983). Because there is little or no change in physical parameters and larval availability, substrates immersed at different times are exposed to the same factors and develop along the same pathway. In this manner temporally asynchronous assemblages developing within a discrete habitat can be compareq to each other to discern the degree of reproducibility in development without seasonal influence. Assuming that there is a high degree of developmental similarity in the sequence observed, the effects of habitat alteration by thermal variation acting upon what amounts to the same assemblage in different stages of development can be seen at the onset of winter. Seasonal and successional attributes of the development of a fouling assemblage, as well as the impact of seasonal change on development, can be identified. Methods Scallop shells (Argopecten irradians) were collected and prepared as described in the previous section but were left uncut, the entire valve being used as the fouling substrate. A small hole was drilled through the center of each valve through which a piece of monofilament fishing line, knotted in the center, was reeved. The two free ends were threaded through two holes in the plexiglass strips and tied underneath. A drop


45 of silicone sealant was placed on this knot to prevent its working loose. Ten shells were attached to each plexiglass strip and constituted the ten replicates of a single sample. A total of 20 strips were prepared (200 shells) and divided into 2 subgroups of 10 strips carrying 10 valves each. These subgroups constituted the substrate used in a sampling series and were labelled sample sets A, and C to correspond to the labelling scheme used in chapter 4 for the shell chips. On May 27th, 1978, the first 10 strips (set A) were bolted onto the frame. Collection of the shells occurred at 2,3, and 4 weeks then at 4 week intervals to 32 weeks. Each replicate was labelled with a code consisting of exposure time: series: replicate (ie shell is the fourth shell of the twelve week sample of series A). Set C was emplaced eight weeks after set A. The last number is dropped to indicate the entire sample instead of the replicate. Both temperature and salinity measurements were taken at collection time. Salinity was determined in the field using a hand held refractometer. Water temperature at the time of sample collection was measured with a research grade centigrade thermometer. Daily ranges were monitored over selected periods using a 24 hr. recording thermometer. At collection, the knot below the strip was cut and the shells placed into separate glass jars to which a solution of 10% buffered formalin was added. After the assemblage had been fixed the formalin was decanted, filtered (0.067 mm) and replaced by 70% ethanol. The filtrate was washed back into the sample jar. Samples were stored in the dark until just prior to sorting. At the initition of sorting, the shell was removed to a large petri


46 dish and gently cleaned with a soft bristle brush. Under a dissecting microscope, each barnacle or bivalve was cleared of any adhering organisms and detrital material. Balanoid scuta and terga (and mollusc valves) were then removed and examined for boring organisms. The body of the barnacle was removed and discarded after specific identification. Lateral plates were removed and examined as were the opercular plates and subsequently discarded. In this fashion, the entire assemblage was broken down to the bare shell material, including the removal and examination of the bases. All the material cleaned from the surfaces was stored in ethanol during the breakdown process and subsequently filtered through a 0.067 mm screen. The fouling material was c omposed of both flora and fauna as well as organic and inorganic debris and detritus. The flora was recorded as either present or absent as quantification is difficult even in well treated samples and the brushing had fragmented many of these small algae. Fauna was sorted from the filtrate under a dissecting microscope (20 x) using a small petri dish which had been engraved with a 1 cm2 grid. Harpacticoids and nematodes were found to be extremely abundant in samples exceeding one month exposure time and, after the first few replicates had been examined, these were subsampled. Enumeration for these groups was accomplished by counting only those individuals found in selected members of the grid. Transformation to total numbers was based on experimentally determined conversion factors. Other taxa were totally censused. After the first pass in which these two groups were counted, the entire dish was examined one at a time for other fauna. This


47 process was repeated until all the material which had been cleaned from the shell had been examined. To recover small organisms from the dish, a Gilmont 2 ml microsyringe equipped with a teflon tube (interior diameter == 0.5 mm) was attached to the base of the microscope and used as a microsorting pipette. The fauna was rough sorted into major taxonomic groups and subsequently identified to the lowest possible taxon. Throughout the sorting procedure notes were kept on various aspects of the fauna (ie. size, fecundity, location). Most of these specimens were retained in the University of South Florida's Museum of Benthic Invertebrates housed at the St. Petersburg campus. These data were augmented by the invertebrate counts obtained from the shell chips used in the microfouling portion of the study. Individuals were identified to the lowest possible taxon and the data from each replicate were recorded as the replicate was sorted. Those taxa which were not at the species level were omitted and the replicate data of a sample set were averaged to produce the Sample data set. Data sets were also transformed to percent composition for examination independent of density. Progressive similarity (appendix) and SpecieshArea analysis (Preston, 1962) was used to determine sampling sufficiency. This test was applied to the replicate data both before and after the deletion of categories. Abundance profiles were generated on all categories in each data set. Diversity values (Cuba, 1981) were calculated for replicate and sample data sets. Plots of the # individuals/sample and # new species/sample, as well as colonization curves, were prepared for data


48 sets A and C Similarity values between each consecutive sample pair were calculated for sample data and transformed to Dissimilarity values for use as an indicator of the level of change between sequential samples. Dissimilarity levels for consecutive sample pairs were computer analysed and the contribution to dissimilarity due to each species, determined as a percent contribution annotated with vector magnitude values, was output for later use in determining the causative factors in the structural changes the assemblage underwent through the time course of the experiments. dendrograms using similarities (Bray and Curtis, 1957) and group averaging ordination procedures (Cody, 1974) were produced for each of the data sets as well as the sample data set which had been transformed to species presence or absence data. analysis was carried out to determine species' associations. Plots of the population curves for each species of each data set were prepared in order to verify that the species' associations produced above were not artifacts of abundance bias. Occurrence data were transformed from the observed value to the percentage o f the total abundance that the observed value represented. This was done independently for each species. This "Shape" transform removes the density bias inherent in the Bray"'Curtis method H .-mode analysis was again applied to determine the levels of similarity in the shape of the abundance profiles. To test the influence of the time of exposure as compared to the influence of the date of collection similarity values were calculated between each of the two abundance profiles generated for each species


49 using the exposure time to calibrate the curves. The curves were then calibrated by collection date and the similarity values recalculated. The same test was done on the shape transform data set. Results Data aquisition yielded 135,716 individuals belonging to 94 taxonomic groups. Seven of these categories (Oligochaeta, Nematoda, Harpacticoida, various juveniles, and unidentifiable taxa) were omitted (73,705 indivduals) resulting in a data set comprised of individuals identified to the species level. Members of the Harpacticoida and Nematoda accounted for 52.46% of all organisms and >90% of the omitted fauna. Of the remaining fauna, 14 of the species accounted for >91% of the individuals (table 2). The specific data set was 48.7% Foraminifera; 14.5% Tanaids (1 species); 12.7% Molluscs; 9.4% Cirripeds; 7.7% Podocopid Ostracods; 4.3% Polychaetes; 1% other Arthropods; 0.8% Gammarids; and 0.5% other (Ascidians, Platyhelminthes etc.). In both experiments the number of species per sample rose rapidly (3-4 wks) to about 30 (fig. 7a) then continued rising at a slower rate, stabilizing between 40 and 55. The number of new species per sample (fig. 7b), initially low, peaked at 2.:.;3 wks, then declined. A stable level was not reached but the oscillations were always lower than the peak at 2-3 wks. Plots of the numbers of individuals per sample exhibited the same general shape as the number of species per sample. There was, however, a general increase in total abundance in samples and followed by a decline to previous levels.


50 Table 2. The fourteen most abundant species. N = Total number of individuals; Freq = the number of replicates in which the species was found; % = the percent composition with respect to the data set after deletion of those individuals not indentified to the species level. Taxon N Freq Quinqueloculina costata 9471 70 15.27 Hargeria rapax 9026 63 14 Triloculina cf. rotunda 8605 69 13.87 Brachidontes recurvus 7682 59 38 Balanus eburneus 5439 64 Triloculina cf. oblonga 3466 60 5.58 Triloculina sp; 2967 64 4.78 Cyclogyra involvens 2824 46 4.55 Cyprideis sp. 2756 64 4.44 Cribroelphidium poeyanum 1893 69 3.05 Loxocythere sp. 1369 25 2.2 Polydora websteri 983 71 1.58 Brania clavata 495 60 0.09 Typosyllis aciculata 482 56 0.09 Total 57458 91.2 %


Figure 7. a. Number of species/sample. b Number of new species/sample. Set A (.), Set C (x).


NO. NEW SPECIES/SAMPLE NO. SPECIES/SAMPLE co-.. co ...... ""'""' .... .... .... .... .... 1\) """"' .... 1\) w -1>- ()) 0 1\) -1>-a> ()) 0 ::t::t::t 0 0 0 0 0 ])])]) ])])]) 1WK 1 ,._ 1W K 2WK --..._-:;.x 2WK -----X" l 3WK 3WK 4WK 4WK / / j I I / I I I 8WK -1 >$. 8WK x ' ' ' ' ' j /''x. I I 12WK 12WK \ \ \ ] \' (/) 16WK )< I 16WK )> / ., / r m 20WK ) I 20WK )( I I I J I I 24WK f/ I 24WK -l >;( I I I I I I 28WK -1 I 28WK -1 I I I I I al i I 32WK .v ::.. 32WK /;lc ll>l Vl 1\)


53 Diversity, as measured by Cuba (1981), initially was low with high levels of evenness but by the second week had progressed to near 20 while evenness declined (fig. 8). After two weeks, diversity continued to climb, finally reaching an oscillatory equilibrium between 35 and 55, centered at 41.33 while evenness fluctuated between 0.2 and 0.4. The plots of levels of dissimilarity between temporally adjacent samples (fig. 9) indicate that, for both data sets, the rate of change is great at l o w exposure times, declining after the first few weeks, and fluctuating, at low levels, throughout the development. Examining the sequential levels of dissimilarity based on species presence or absence yields similar plots but with lower amplitude oscillations in the later samples (fig. 9). Analysis of the specific contribution to dissimilarity between sequential samp les revealed that most of the changes wer e due to variations in population density rather than species composition. In all but 4 of the sample pairs (3wkwk -A; -C; 20wk-24wk -C) most of the dissimilarity was contributed by species which wer e increasing in density. Joint Q:-mode analysis of both sets of sample data produced the dendrogram presented in figure 10. All 1 wk and less samples were found to be nearer each other than to later samples but the level was low. Among the 2 wk and longer samples, three significant clusters were produced. The first consisted of samples 3wkl,' through 12wk'-A and 2wk and 3 wk-C. Group two contained only 4wk-and 8wk:-C. The last group was made up of 20wkA through and through Samples 16wkA and 20wk-C were not significantly similar to any of these groups but were closest to groups 1,1 and 3 respectively.


Figure 8. S factor of diversity (ordinate) plotted against the X factor (abcissa). Set A data are open circles, set Care *'s.


I '1\ ' 0 ' \ \ \ \ \ 10 \ ' \ \ \ \ \ ' \ \ '*-*,--"!' ..... ::::-', ........... ', ' I ' 20 30 S-F ACTOR 40 50 60 55 .9 .8 .7 .6 .5 .4 3 .2 a: 0 0 < LL I X


Figure 9. Sequential dissimilarity levels. a. Set A. b. Set C Based on species and their distribution (.) and species presence/absence alone ( *).


DISSIMILARITY ..... ....... "'""" .... i-.J G.> :,. (n C:n :.... 0> (o 0 xx:r ..... i-.J G.> :,. (n C:n :.... 0> (o 0 """ )C'-----JIC lL 1WK-C 1WK-A 2WK-C r> 2WK-A 3WK-C 3WK-A 4WK-C 4WK-A 8WK-C-t \. I I 8WK-A I I I I I 12WK-AJ I I I 12WK-C-! /I I I I (/) \! )> I !: I "'0 I r I m I I I 20WK-C-! I 20WK-Ai I I I I I I I I 24WK-CJ v 24WK-A \ \ \ \ I I 28WK-C -1 \ I I 28WK-A-l I I I I I x 32WK-CJ IJI


Figure 10. Dendrogram of sample to sample similarities where solid line (commencing at 0.5) indicates significant clustering. numbers consist of exposure time and experiment code. Sample


72HR-c ---------------------NOT CLUSTERED 48HR-A 96HR-C 72HR-A 1WK-A 1WK-C 96HR-A 2WK-A 2WK-C 3WK-C 3WK-A 4WK-A 8WK-A 12WK-A 16WK-A 4WK-C 8WK-C ---------... ----. I --------------------------------------I-1 I ----------------------------------------------I I I I I : : --I I I I I I I ---------------,. ___ _jl l---------:--_, I I l I -------------------------------------------' I I I __ ------J -----------I I I I 20WK-C 1--J 20WK-A I I 12WK-C I :_ --------J 16WK-C : 24WK-A I I 24WK-C _J r 28WK-C 28WK-A 32WK-A 32WK-C 1 .9 .a .7 6 .5 SIMILARITY .4 .3 .2 1 0 IJl 1.0


60 The pattern depicted in figure 10 reflects that seen in the dendrograms of replicates from set A and C taken separately and together. In all dendrograms based on replicate data the replicates consistently joined their sister replicates at higher levels than they joined replicates of other samples. The pattern of early replicates being isolated and late replicates forming a tight cluster while middle replicates were interlaced was common to all dendrograms (sample and replicate) with exception of those constructed on the basis of species presence or absence alone. Transformation of the data to percent composition served only to tighten the clusters. Dendrograms based on discrete sample sets produced the same pattern but the middle samples tended to be better grouped. Considering only species presence or absence in the dendrogramming procedure resulted in patterns for both data sets A and C. The samples were split into and long,.. time clusters. The break occurred between 3Wk'":: and 4wkA and 12wk,... and 16wk'-'C. In both, the 1 wk and less samples were as before. Analysis of both sets together (fig. 11) yielded a 3 cluster pattern with 1 wk and less samples again set off from the rest. Group 1 contained 2wk"",3wb,4wk'""A and 2wk,.,C. Group 2 contained 8wk-through 20wk"-A and 3wk.. through 8wk..,C. The final group contained all other samples; 24wk'through and 12wk-through analysis was carried out on replicate data for sets A and C both independantly and jointly. In all three dendrograms the clusters formed were conspicuously correlated with the total abundance of the species in the cluster. The only exception among the 14 most abundant species was Quinqueloculina costata which ranked 1st in overall abundance


Figure 11. Dendrogram of sample to sample similarities considering species presence or absence only. Solid lines indicate significant clustering. Sample numbers consist of exposure time and experiment code.


72HR-c -----------------------NOT CLUSTERED 48HR-A 1WK-C 72HR-A 96HR-A 1WK-A 2WK-A 3WK-A 2WK-C 4WK-A 8WK-A 12WK-A 16WK-A 20WK-A 3WK-C 4WK-C 8WK-C 24WK-A 12WK-C 32WK-A 24WK-C 16WK-C 20WK-C 28WK-A 28WK-C 32WK-C ================r-------I -----------------------------------_,-----------""j 1------------------------------.---------------I I I ----------------------------------------I I I I I I .....:....__ _____ J-I I :----1 I I I I I I .I I 1 I I I I I I 1--------------1 I I I I I _____ j 96HR-C -----------------------------------------------1 9 .8 .7 6 .5 .4 .3 2 .1 SIMILARITY 0 0\ 1\)


63 but did not fall into any group in any of the three treatments. Applying analysis to sample data (singly and jointly) revealed nearly the same pattern. g. costata again was unclustered. The next 4 most abundant species formed one group and the 6th through 11th ranked species formed another, exactly as in the dendrogram generated from the replicate data. Other groupings were slightly rearranged. Dendrogram construction using the shape transformed data (fig. 12) resulted in 10 clusters clearly disassociated from the abundance factor. Although no single analytical method clearly illuminated the order of events, a pattern emerged when all were taken together which is most easily seen in figure 12. Groups 1, 2, 5, and 9 contained early colonizers which did not require the presence of other species. After 2 months, groups 4 and 8 invaded and group 9 declined in abundance. Members of group 10 were not present until late in the development. Analysis of similarity between population curves yielded uselful information on 58 of the 87 species. The remaining 29 species were so rare as to completely invalidate any results. Of those 58 species, 37 ( 63. 3%) exhibited an increase in similarity when the population curves were calibrated with respect to date of collection instead of exposure time. Decreases were seen in 7 (12%) species, 14 (24%) showed no significant change <0.05 difference ). In 24 of the 37 species exhibiting increases in the level of similarity, the magnitude of that increase exceeded o. 20, the greatest being o. 78 in the case of the pycnogonid. Another 10 species had increases >0.1 and <0.2. In the 14 most abundant species all but 2 showed increases, declined, and showed no change. Of the 4 structure producing species, 2 (Balanus eburneus and Brachidontes recurvus) showed increases, and 2


Figure 12 Dendrogram produced by shape transformed interspecific affinities (occurrence similarities). Solid lines indicate significant clusters.


Cyclogyra lnvolvens Ctenodrllus serratus I I Podocopld sp. U Anthozoa sp. A ---------; Pycnogonld sp. A '--------------. Uassarius trlvl t at.t.u s : Bursatella leac h II------------------------------------------------' : Crlbroelphldl11.11 poeyanu., -------------Hargarla Branla clavata -------------' Typosyllls acl c ulata ========:J---:---l Crassostrea vlrglnlca Stylochus aooerlcanus --------------' Durvlllea Caprellld sp. A---------------, Isopod sp. A----------------' Elasmopus levis --------------------__j Tretonophalus bulloldes 1------, Cyprldels sp. I "'1 Loxocythere sp. : Anomia o Styella pllcata : Trllocullna sp. :------; Po docopld sp. L o : ' Arenlcola sp. : : Acarina sp. A 3 _j i1 '----Area : : Qulnquelocullna t>osclana : A'"Phlthoe long iCII3na 4_ ________ : : Eu..lda sangulnea Trlloculina c r oblonga : Trllocullna cr. rotund '! : Balanos aC1phl trite l"j Balanus eburneus : 8r a chldontes recurvu s : : Neaothes sueclnea : : Podarke o bscura 1 Polydora webster! -----: L ----. Crepldula r orntcat.a : : : : Podocopl d sp. E }-- Podocopld sp. D 6 _____ j : : Hlcropanope xanthlfom i s : : Podoeopld sp. X 7--, ; :-----: 1 P o docopld sp. H : 1 : Podocopld sp, r ---------------------------------------: : o Tegastld sp. A , : Crandldlerella sp. 8 --------________ j Mel! ta nl tlda : Qulnquelocullna costata : T r llocu llna llnneana var. a : Trllocullna llnnea na var. b 9----------------------. : Prlonosplo heterobranchla J Crandldlerella bonnlerolde3 -------------------------------------------------------------.I ----------------------------------------------------------_}-----------------10 . ' t t 1--; t ' ' ' Corophluno siC11le rlchthonlus braslllen31s Ctenodrllus sp. C C y madusa r llosa Ctenodrllus sp. 0 Podo copld sp. B -------------------------------------------------------------.9 8 7 6 .5 4 .3 2 1 0 CT\ \.11


66 (Balanus amphitrite rapax) showed decreases. Salinity plots indicate no change in habitat parameters over the entire course of the experiment but thermal data indicate that changes occur bet ween samples 16wk and 20wk""A ( 8wk 12wk'"'C), 24wk and 28wk'-A (16wk 20wkr>C), and 28wk and 32wk.-:C. These are a lowering of summer daytime highs, a lowering of fall nightime lows and a general spring increase respectively (fig. 2). Discussion Hopkins (1977) found a peak in benthic larval abundances in Tampa Bay during summer months. His results are supported by the observation made here during the sorting of samples that all of the dominant species were found to be reproducing throughout the project. The most notable exceptions are Micropanope xanthiformis and Gobiosoma robustum. Despite the fact that the larvae of many sessile and mobile benthic invertebrates were present in the water column during this study there is no meio"' or macrofauna! colonization prior to the 48hri-'A and the sample. Microbial analysis of the substrate during the first week of sampling (samples 1wk) determined that this time period was one of stabilization in the microalgal composition of the assemblage. If larval recruitement was truly random there would have been no lag time in larval settlement. Taken alone, this does not preclude random settlement after microbial preparation of the substrate has been accomplished. It does, however, reinforce the observations that have been made which indicate that microbial substrate preparation is required for macro and meiofaunal colonization 1951, 1953a).


67 High levels of similarity among replicates verifies the results of Schoener and Green (1981) and shows that when identical substrate is immersed in the same location at the same time the assemblages develop similarly. Levels of similarity between the short exposure time samples of the two sample sets indicate that the two assemblages commence their development along similar pathways. The very loose cluster of short exposure time samples (fig. 10 and 11) reflects a certain degree of stochastic development with respect to the initial structure of the assemblage. During this time, however, the species curves (fig. 7a and 7b), serial dissimilarity curve (fig. 9) and the diversity plots (fig. 8) all indicate a very rapid rate of change and by 2 ""'3 wks (fig. 10) convergence with respect to exposure time has occurred. It is unlikely that the competitive factors modeled by Paine and Levin (1981) could have had any affect in this short time especially when considering the very low abundances seen in the 1 wk samples. Macrofauna! and meiofaunal species composition among early samples exhibits no pattern with regard to which of the colonizers arrives first. There is no apparent required order to the initial colonization sequence other than the marked tendency for the earliest arrivals to be larval or juvenile forms rather than adults. The Harpacticoida are conspicuous exceptions, nearly all being adult forms. This is consistent with the patchy and micro-migratory nature of harpacticoids (Bell et al., 1978; Bell, 1979). Successional events therefore can be summarized by the sequence taken from the dendrogram. Specifically, early colonization is undertaken by groups 1 2, 5, and 9; Species in group 9 decline and species of groups 4 and 8 invade after 2 months. The assemblage is


68 successionally stable until 28 to 32 weeks when members of group 10 occur. Development is similar with respect to exposure time even in these asynchronous samples due to the constancy of physical parameters and species availability. When the assemblages encounter the shift in temperatures between 16 Sep. and 14 Oct. they are at and exposure time and samples of set C are developing along the same pathway that set A had. The thermal changes are reflected in the plots of # species/sample and new species/sample (fig. 7) and in diversity (fig. 8) but are most pronounced in the sample to sample dendrogram (fig 10). Sample 16wk!..A represents a discontinuity between the middle and long exposure time samples. All of the samples taken after this thermal event fall into the last major grouping of the dendrogram and the samples are now more similar with respect to their date of collection rather than exposure time. Analysis of the contributions to dissimilarity by each species showed that the major result of the thermal impact was an increase in abundances of species already present while having a minimal, but detectable, effect on species presence or absence. Analysis of similarity between population curves, showing that most populations were more similar with respect to date of collection than with respect to time of development corroborates the conclusion that control shifts from temporal to thermal as development precedes. Aspection (Shelford et al., 1935) in assemblages throughout the seasons is brought about by the seasonal thermal shift, including new species and extincting others, while not greatly affecting the overall structure and composition of the assemblage. This is possible due to the nature of the significance of the magnitude of change in the


69 habitat. Changes in the physical parameters (T0 S0loo etc., Jones, 1950) and biotic parameters affect each species in a unique manner and, in subtropical waters, the magnitude of change in the habitat is smaller than in temperate zones. The degree of change caused mild aspection in this study, while the larger changes experienced in temperate zones caused disruption (Sutherland and Karlson, 1973, 1977; Dean, 1977). Even in a subtropical estuary, as in this study, the development of two asynchronous epifaunal assmeblages, which were initially controlled by successional factors, became seasonally controlled as the physical bounds of the habitat shape were altered. Often it is difficult to know what constitutes a physical difference between habitats. The assemblage encountered here bears little similarity to the one described by Mook (1981, 1983) from the same zoogeographical province. Mook, however, utilized vertically oriented fouling plates and the scallop shells in this study were arranged horizontally. Although the structurally dominant forms that Mook found (Styella plicata and various Bryozoa) were poorly (or not at all) represented on the scallop shells, they were the dominant forms on the legs of the frame and the undersides of the plexiglass panels carrying the shells. It is apparent that the horizontal scallop shell (and upper surfaces of the panels and frame) are not homologous to vertical surfaces or the undersides of horizontal surfaces even when fully submerged. That vertical and horizontal surfaces constitute separate habitats and that different habitats are often more easily recognized by the fauna than by the investigator is supported by the observation that the plexiglass panels were heavily colonized by Bryozoa and Ascidians on the undersurface except for the portion directly beneath the scallop shells


70 (shaded). The latter area was heavily colonized by the photonegative barnacles completely eliminating the former. Upper surfaces of the plexiglass betwen the shells were not colonized by either the cirripeds or the group but by infaunal and mobile epifaunal species inhabiting the shallow layer of silt and detritus which had accumulated there. The data presented here are in general agreement with Shelford et al. (1935) and Wells (1961) showing aspection overlying a base, year round resident assemblage. The new occurrences coincident with the thermal alterations are all low in density and are quickly deselected by the short duration of the winter temperature regime. Still, the number of species/sample fluctuate in the same manner as those presented by Schoener (1974) for fouling Her objection that studies of sessile epifauna are characterized by incomplete data sets cannot be applied to these data. The fluctuation seen in the colonization curves is not successional but an artifact of the occurrence of rare species combined with seasonal changes in the habitat and concommittant compositional changes. Variability in larval settlement then is more highly correlated with habitat inequities than with simple differences in time and space with regard to substrate emplacement. It is the physical boundaries of the habitat hypershape, which often do change within time and space that results in the observed variability in colonization of substrate emplaced asynchronously. The magnitude of the temporal shift in substrate emplacement may actually be artifactual. Tropical areas exhibit lower amplitude thermal fluctuations and a more constant supply of larvae. In temperate zones, the temporal offset carries with it the notion of great


71 variability in thermal regime and composition of the larval pool. It is these factors, not simply time itself, which serves to modify asynchronous assemblage development. A distinction must also be made between succession and development. Where successional change connotes the shift from one sere to the next, it is not instantaneous and, especially in primary succession, time is needed to develop the sere. The macrofaunal developmental stage of this study commenced at hrs and lasted wks at which time convergence occurred. Prior to convergence, the two assemblages were different (fig. 10) but rapidly and deterministically changing (fig. 7,8,9, and 10). The data indicates that stochastically determined intitial points lead to the same stable point and a reevealuation of the selection forces involved in initial colonization is required. The reviews of Meadows and Campbell, (1972) Crisp, (1974) Scheltema, (1974) and Thorson (1966) describe works that, when taken as a whole, totally refute any hypothesis of random larval distributions and site selections between habitats. On an expanse of equitable (homogeneous), microbially prepared habitat, however, the spatial distribution of settling larvae must be random in its initial stage. Facilitation, Inhibition, Tolerance (Connell and Slatyer, 1977), gregariousness (Knight,..Jones, 1953a,b,c; Knight'"Jones and Moyse, 1961; DeSilva, 1962) and competetive exclusion (Connell, 1961, 1972, 1976) interactions cannot function when there are no current occupants to facilitate, inhibit, etc, the arriving larvae. Coupled with the observation that the available habitat at this stage (2(""3 da) is very nearly homogeneous, there is no biological force acting to deflect settling larvae from a random choice of site provided that it remains within the subject habitat. Once a


. 72 single larva metamorphoses and occupies a particular site, the forces listed above can become active. As with most physical force functions, the effect diminishes with increasing distance from the source. Thus a single juvenile may preclude random settling, but only within his sphere of influence. Typically this sphere (with respect to juveniles) is only a few mm in size and the remainder of the substrate is still open to stochastic distributions. The overall effect of this colonization scheme is directional. Prior to colonization, only larvae finding the site suitable will attempt settlement. Once some juveniles are established, the site suitability in their vicinity changes but arriving larvae presented with an inhibitory, settled juvenile, can avoid that inhibition as long as ample uninhibited, space remains. Facilitation (self-detrimental or sel f"iperpetuati ng) and gregariousness are positive forces (attract! ve rather than repulsive) increasing the likelihood of further settlement and expanding the sphere of influence. Facilitation and inhibition (as well as the other forces) are not exerted indiscriminantly. There must be a specific object of facilitation, inhibition, etc. As an example of the dichotomy of specific interelationships, Obelia dichotoma inhibits Balanus crenatus but facilitates Ascidia ceratodes (Standing, 1976). Thus the positive pressures of facilitation and gregariousness, already shown to be more of a shaping force than inhibition during early colonization, serves to actively select subsequent and specific) colonizers thereby shaping the structure of the early assemblage It is in this fashion that the stochastic distribution of juveniles becomes damped and convergence occurs within 2 weeks time (fig. 10 an d 11).


73 Successional synchrony continues until the physical parameters are changed. In many areas the magnitude of the change is great enough to cause a shift in the successional track that an assemblage is pursuing. In others, where the change is smaller, yearly aspect ion occurs. In both, the impact of a seasonal change cannot be considered a truly successional change but rather a potential instigator of change. Summary Larvae or juveniles of meiofauna or macrofauna are not present in the developing assemblages until after 24 to 48 hrs. of exposure. All those occurring in the 24 hr. samples are harpacticoids and these were largely adult forms representing immigration. Colonization during this time corresponds to the stabilization of the microalgal assemblage but no cause and effect can be shown Fouling assemblages developing on like substrate but having been immersed 8 weeks apart initially are similar with respect to time of exposure. As development continues, the assemblages are impacted by changes in the physical parameters and convergence occurs with respect to the date of collection. The impact of the first thermal event was an increase in abundance of nearly all species present. The second thermal event was correlated with minimal changes in species composition. The developing assemblages exhibited succession, influenced by changes in physical parameters and seasonal aspecti on.


THE SUCCESSIONAL AND REPRODUCTIVE STATUS OF HARGERIA RAPAX IN A DEVELOPING POPULATION 74 Hargeria rapax (Tanaidacea) was found to be an important member of the assemblage developing on hard substrate emplaced in a cove of Tampa Bay, Florida. Juveniles were among the first immigrants in the develomental sequence and the species quickly became one of the dominant members of the assemblage both numerically and functionally. More than 9,000 individuals were collected over 10 months time in the twin sampling program, providing a set of unique data for analysis of population growth and reproduction. The species was originally described by Harger (1879) as Leptochelia rapax. Lang (1973) erected the genus Hargeria dubia (in part) has since been synonomized (Heard, 1982). L. dubia from the Pacific coast remains distinct. Females of L. savignyii (North Atlantic and Mediterranean Sea) are indistinguishable from those of rapax (Boesch and Diaz, 1974) and the slight differences in the male may be phenotypic although they have not been synonomized. Wells (1961) reports both from the Newport River Estuary, No. Carolina. Pursuant to taxonomic refinements, the species can be considered cosmopolitan or limited to Bermuda and the east and Gulf coasts of North America (Heard, 1982). Common in nearshore, low energy environments, H. rapax tolerates salinities ranging from 2 ppt. to oceanic (Boesch and Diaz, 1974). rapax and L. dubia build tubes of mucous and detritus forming dense mats in the upper layers of sediments (Brenchley, 1981) or on the


75 surface of hard substrate. No specific work is known concerning the tube structure of H. rapax but several other tanaids have mucous lined tubes with detritus or sand affixed to the outside (Greve, 1967). Methods Two arrays of scallop shells, (Argopecten irradians), were placed in a shallow lagoon of Tampa Bay, Florida. The first, labelled set A, was emplaced on May 27th, 1978, the second, 8 wks later, on July 22nd and labelled set C. Samples of each set were recovered at weekly intervals for 4 weeks and at 4 week intervals from 4 to 32 weeks yielding data on two similar but separate developing populations. The samples were fixed in 10% buffered formalin and transferred to 70% ethanol. Individuals of H. rapax were sorted from the rest of the fauna into vials (See previous section for a complete description of the collection methods). During sorting, numbers of eggs/brood, manoa/brood, juveniles/brood, and non-brooding juveniles sharing the same tube as an adult female were counted. Total lengths (TL) of the young in adult tubes and in their own tubes were taken as they were encountered. Carapace lengths ( CL) were measured after all the samples had been sorted. Females representing the three readily apparent reproductive states (visible, internal, developing eggs; brooding; brooding and internal developing eggs) were prepared for histological sectioning. These were first transferred from the 70% ethanol to S29 dehydrant (Technicon corp.) for several days. They were again transferred to fresh S29 24 hrs prior to embedding. Just prior to embedding these were treated with 2, 30


76 minute washes in fresh S29 followed by 3, 20 minute baths of UC676 (Technicon corp.). Rather than attempting to transfer these small organisms and risk mutilation and the dislodging of eggs from the marsupium, a Gilmont micro'-pipette was used to transfer the liquids during the above preparation. Pre-"mel ted paraffin (mp=56i-157C) was kept liquid in small watch glasses on a laboratory grade hot plate. The animals were subjected to 3, 20 minute baths of paraffin before placement in the block. Non:..destructi ve transfer was accomplished with a fine tipped pair of flexible forceps. After hardening, the blocks were cut into 6 thick sections with a microtome, mounted, and stained in accordance with the Harris techniques (Luna, 1968). Number and size of eggs/individual were recorded and notes made on the position and orientation of the ovaries and reproductive structures. Gut contents were examined and notes kept on the apparent physiological condition of the animals in different reproductive states. Results In each of the developing populations, initial immigration occurred as early as 48 hr after emplacement (table 1). Immigrants belonged to size classes indicating that they were either juveniles which had just emigrated from the parental tube (juvenile-2) or had passed through only one post'-emigration molt (preparatory Populations densities remained low until the 20 wk sample in assemblage A and the 12 wk sample in assemblage c, after which their densities increased dramatically (fig. 1 3).


Figure 13. Population curves for H. rapax. Mean number of individuals/sample with range bars. Top, Set A; Bottom, Set C.


NO. INDIVIDUALS ............ ... -1\) (,t) (.11 (7) ...... Q) c.o 0 ... 1\). (,t) ...... Q)C.00 ooooooooooooo oooooooooo ooooooooooooo oooooooooo 1WK-AJ I I I I I I I I I I 1WK-C 0" C) 2WKC 2WK-A-3WK-C 3WK-A-4WK-C 4WK-A-8WK-C I awK-AI 12WK-C+ 12WK-A I en 16WK-C \ \ > 16WK-A s:: ""0 r 20WK-A m 20WK-C t----; 24WK-C-l 1-( 24WK-A 28WK-C -t 28WK-A 32WK-C-f I 32WK-Ai I --4 <


79 Carapace lengths from the 8 wk through 28 w k samples (set C) are plotted in the histograms in figures 14 and 15. Prior to 8 wks only a very few indi victuals are present and they are all in the smallest size class. At 8 wks (fig. 14a) a few larger individuals were encountered. By 12 wks (fig. 14b) the full range of sizes have been encountered and the distribution among the size classes is typical of a developing population (Jeffers, 1978). At 16 wks (fig. 14c ) there is a large increase in the number of small individuals which is continued in the 20 wk samples (fig. 14d). Between 20 and 24 wks a decrease in the production or immigration of juveniles occurs and the histogram displays a more even distribution among the size classes (fig. 15a). By 28 wks (fig. 15b) the number of young is increasing and the histogram takes on a pattern more typical of a mature population (Jeffers, 1978). Females of sample (fig. 15c) were sorted first into categories based upon the visible reproductive state of the individuals. These are juveniles with no apparent ovarian development, individuals with internal eggs visible through the body wall but without any development of oostegites, individuals which are brooding eggs or young but in which a second set of internal developing eggs are not present, and individuals which are brooding one set of young and have a second set of eggs developing within the body cavity (fig. 16). Juvenile-1 class individuals and marsupial juveniles are not included in figure 16. The latter, which molt into a juvenile stage upon leaving the brood chamber, had lengths centered about a CL of 0.125 mm. CL' s of juveniles showing no ovarian development (j 2) are centered about 0.20 mm (fig. 16a). Figure 16b was constructed from measurements of females with eggs or


Figure 1 4 Frequency of occurrence (%) of females of rapax with respect to carapace length. a.) Sample 8wkc; n=19. b.) Sample 12wkc ; n=68. c.) Sample 16wkc ; n=276. d.) Sample 20wkc; n=617.


60-50-40-30-20->-10-I 30# 20 10 30 20 10 n n a n :1 c d .15.20 .25 .30 .35 . .60.65 .70 .75 CARAPACE LENGTH (mm) 81


Figure 15. Frequency of occurrence (%) of rapax with respect to carapace length. a.) Females, sample b.) Females, sample 28wkc; n=521. c.) Females, sample 32wkc; d ) Males, pooled froo all samples n=75.


(.) z w ::::> 0 w a: u. cP. I al 20 10 309 20 10 I; J I; rrn 1-r-, J 9. d .15 .20 .25 .30 .35 .40 .45 .50 .55 .60 .65. 70 75 CARAPACE LENGTH (mm) 83


Figure 16. Frequency of occurrence (%) of females of rapax from sample 32wkc with respect to carapace length and separated by reproductive status. a.) class; no apparent ovarian activity. n=189. b.) Preparatory and 2 and Copulatory female<"1; ovaries active, eggs mature or immature, but always internal. n=230. Copulatory female'-2; females with eggs in the marsupium. n=33. c.) d.) Copulatory females with a set of eggs developing internally and brooding juveniles in the marsupium. n=6.


50 40 30 20 10 ....a r-1-r-,_ i 0 40ij w c a: 30 u. '(/}. 20 10 30 20 10 d .15 .20.25 .30.35 . .60 .65.70 .75 CARAPACE LENGTH (mm) 85


86 ovary visible through the body wall and tends to trimodali ty. Upon closer examination of the specimens, it was noted that the modes are made up of: 1.) Preparatory Individuals with distinct ovarian development but no apparent ova (MCL 0.30 mm); 2.) Preparatory female.-2: Individuals with developing ova (MCL 0.375 mm); and 3.) Copulatory Individuals in which eggs with well developed yolk were visible (MCL = 0.45 mm). Figure 16c is composed of those females of sample which were brooding eggs or juveniles and exhibited no evidence of maturing internal eggs. Those females which had marsupial juveniles as well as maturing ova visible within the body cavity are presented in figure 16d. The population level in the first experiment remained at approximately 2 ind/ em 2 through the 16 wk sample whence it climbed to 17/cm2 by 20 wk. In the second experiment the same initial levels were seen but the population was depressed only through the 8 wk sample and increased between 8 and 16 wks. This period of population increase, although different between experiments with respect to exposure time, corresponds with respect to the calendar date (16 Sep. 14 Oct.). Prior to that date ( 16 Sep.) the daily temperatures reached highs which were often >40C. After that date the daily highs were near 25C. The 16 wk sample (fig. 14c), taken after this drop in temperature exhibits a large increase in the number of juveniles present. At 20 wk (fig. 14d) the number has further increased as has the number of preparatory and brooding females. The temperature regime was again altered prior to the 24 wk sample (corresponds to 28wkr.;A to 32wk-A drop in abundance). At that time the overnight lows fell to near 13C. These winter lows were of short

PAGE 100

87 duration and had to previous levels by the 28 wk sample Even though these animals are tubiculous, they were often disrupted during the sorting procedure and the number of eggs or juveniles in the brood sac could not be considered to be the total that had been present in vivo. Only the juveniles and eggs of those females which remained in their tubes were counted. The number of eggs/female being brooded ranged from 10 to 32 (mean=21. 4 ; mean-15 in histological specimens). The number of marsupial young (usually in the manca stages) ranged from 6 to 35 (mean=17.6). Measurements taken of eggs within the marsupium during sorting as well as during the histological examination fell between 0 1 and 0 2 mm in diameter. Even those eggs which contained developed manca stages apparently on the verge of hatching fell within the stated range. Newly hatched manca were consistantly measured at 0 1 mm CL and 0 .45 mm TL. Juveniles in their first molt were found both in the marsupium and sharing a parental tube. Stage 2 juveniles were found in adult tubes and in their own tubes. Those in the marsupium were consistently o 75 mm long (TL). Juveniles sharing adult tubes were measured from 0 75 mm (TL) to 1 5 mm (TL) (mode = 1 0) Of those juveniles {j-2) found within their own tubes (recognizable by the cross sectional cir cumference relative to body size), only one was encountered which measured less than 1.0 mm (TL) Paired ovaries (fig. 17a) extend from posterior half of carapace into the first and sometimes second pereonal segments and are located to the digestive tract. Immature ova are produced by the ovarian germinal cells at the anterior end of the ovary (fig. 17a) and move toward the telson while maintaining a positi on dorsally adjacent to

PAGE 101

Figure 17. of histological sections of Hargeria rapax a. Longitudianl section centered about suture between the carapace and the first pereonal segment. mid-pereonal segment. Ov; ovary: C; Caeca: epithelium: Do; Developing oocyte. b. Cross section of Oo; Oocyte: Ge; Germinal

PAGE 102

Oo c \.

PAGE 103

90 the caecum (fig. 1 7b) Yolk development takes place within the pereonal segments, the maturing eggs never extending into the pleonal segments. During yolk development the nuclei of the eggs remain in close proximity to the digestive tract while the yolk extends dorsolaterally (fig. 17b) Adults with developing eggs continue to feed and the digestive tract is always fully packed with food material which was typically but fragments of polychaete setae, foraminiferal tests, and dinoflagellates were recognized. Adults with marsupial eggs or young often had empty or nearly empty digestive tracts and feeding is reduced in this stage. During brooding, much of the body structure loses its robustness. Gut epithelial cells are reduced in size, muscu lature is less robust, and the body cavity is much depressed dorso-ventrally. Despite this generally degenerative condition the ovaries often remained active. No brooding females were seen which had eggs developing yolk but the ovaries were producing immature ova Very few males (n = 75) were encountered during the study and they were never among the first immigrants. The males first occurred after 12 wks exposure in both experiments at which time a single male was found in each of the samples (F/M ratio = 78:1 and 75: 1 respectively). All males were pooled and measured resulting in the CL histogram (fig. 15d) which exhibits a slight degree of skewness Discussion Much of the reproductive scheme of !:! rapax is apparent when the size class data are combined with the histological information. Ovarian

PAGE 104

91 germinal cells located in the first pereonal segment (fig. 17a) produce immature ova which migrate posteriorly prior to maturation. Yolk development occurs in paired chambers lying alongside the digestive tract and dorsal to the caecum (fig. 17b). Oocyte nuclei reamin in close proximity to the gut wall suggesting an active role for the egg cell in nutrient uptake, conversion and storage (fig. 17b). Mature eggs are extruded into the marsupium where development through the manoa stages (discussed by Lang, 1973) to the j'-'1 stage occurs. The molt from j to occurs at or just after the time when the juvenile leaves the marsupium. The stage animal may remain for a time within the parental tube but emigrates prior to the next molt. The j-2 animal may either emigrate and form a completely new tube or cut a hole in the wall of the adult tube and construct a branch tube. Roubalt ( 1937) has determined that differentiation of gonads occurs during the jP2 stage. During the next stage, preparatory femaleF1, the ovaries become active with developing oocytes but never with mature ova. Preparatory female-2 class individuals develop oocytes but it is not until the molt from preparatory to copulatory that the folded oostegites at the base of the legs are visible (see Gardiner, 1975). Prior to egg extrusion, the oostegites unfold forming the marsupium, but there is no molt or growth. Notes on stomach contents during brooding show that the female in this stage feeds very little and the chances for survival beyond breeding may be poor. When gut material was noted, it never filled more than 10% of the tract. That material was found in the intestine of some females brooding manoa and j-1 offspring indicates that while feeding is much

PAGE 105

92 reduced at this stage, it does not cease altogether. Brooding females examined histologically exhibited an atrophied musculature and general degeneration. (fig. 16d). Some, however, do survive to produce a second brood Tanaid males have been reported to develop both directly and by protogony (Gardiner, 1975; Highsmith, 1982, 1983). Adult males were of the same size class as brooding females but there was a slight extension of the tail of the distribution into the smaller size classes (fig. 15d). It is probable that the males close to, and longer than, the MCL for brood females were produced by protogony and the smaller ones by direct development. This supposition also serves to explain the larger juveniles which had no ovarian development (fig. 16a) as preparatory males. Both Smith (1906) and Roubalt (1937) found adolescent and adult males dubia (these specimens may rapax). The establishement and development of a population of H. rapax, from immigration to population maturity, can now be presented. Prior to immigration, newly available substrate must accumulate a layer of bacteria, and detritus (see previous two sections) in order to make the habitat acceptable to the tanaid. This preconditioning provides food (Kneib et al., 1980) as well as material from which the tanaid constructs his tube. All immigrants observed were either j uvenil eF-2 hydr opho b i c or preparatory femalei-'1 class animals. The exoskeleton, gregariousness and tubicolous habit species' seem to preclude active migratory behaviour and a mechanism has not been established which would induce emigration from nearby parent populations. It can, however, be hypothesized that unsuccessful predation by fishes (Odum and Heald, 1972; Kneib et al., 1980) would disrupt various members

PAGE 106

93 of the parent population from their tubes. The smaller individuals thus liberated would be less likely to be subsequently captured and injested than the larger whiter adults. They would also be more susceptible to lateral transport by water motion bringing small individuals to the newly available habitat. Physical perturbances may induce the same result. Brooding females were not observed until the 8 wk sample and these are more likely to be the result of growth rather than immigration. The members of the juvenile size class observed in the 12 wk sample are composed of immigrating individuals as well as those produced within the developing population. The number of new individuals produced by local reproduction, however, are not sufficient to e xplain the total increase in density, especially the increase in preparatory females, in the 12 week sample. Highsmith (1982) dubia to be a gregarious species and this same behaviour in .!! rapax plays a role in increasing the immigration rate over that of the original unpopulated substrate. In comparing these two experiments, both successional and seasonal attributes can be identified. Although initial colonization occurred within the first week in both experiments, the population growth in experiment A exhibited a prolonged lag phase in comparison with that of experiment C. When the populations did begin to increase, they were temporally synchronous. The high temperatures prior to 16 Sep. may have depressed population growth by reducing reproduction and survivability. The low overnight temperatures encountered prior to the 24wk sample are correlated with both the depressions in the population curves (fig. 13) and the reduced number of juveniles in sample 24wk-:C (fig. 15a). By the 28 wk sample which was again subjected to summertime overnight lows, a better defined

PAGE 107

94 mode for preparatory females (fig. 15b) is seen. The occurrence of the tanaid with respect to exposure time is controlled by successional factors while the densities and intrinsic growth is controlled by thermal events (previous section). Despite the large size of the late stage population, males were still only infrequently found. Highsmith (1982) reports a female/male ratio of 10: 1 for L. dubia. Smith (1906) collected several thousand L. dubia (probably g. rapax) and only 80 males. Lang (1953b) found no males among 300 individuals. Ratios in these experiments had a mean value of 132:1. These males, as in L. dubia, (Brenchley, 1981; Highsmith, 1982, 1983) lack mouthparts and do not feed. There was also no trace of a caecum but the alimentary canal was visible through the body wall although much reduced. The extreme paucity of males in these populations combined with the lack of mouthparts and short life span of the male suggests the possibilty of alternate modes of reproduction. Hermaphroditism is known among tan aids (Lang, 1953a) but self fertilization appears unlikely in this instance. Parthenogetic reproduction, another possibility, is not uncommon in Arthropods and has been reported in members of the Amphipoda and Isopoda 1960). The Amphipod and Isopod discussed by were geographic parthenogenetic species, ie. the mode of reproduction was not present throughout the entire population but only in geographically distinct subpopulations.

PAGE 108

95 Summary .!:! rapax was found to be an early successional species requiring only small amounts of detritus and possibly microbial colonization. Immigrants are mostly in the juvenile:-2 and preparatory femal&-1 size classes. Fluctuations in population density were correlated with temperature changes. Ovarian differentiation occurs in the size class but they do not become active until the preparatory class. Maturing ova are not seen until the preparatory stage is reached. After the molt from preparatory to copulatory female'-1 occurs the animal is ready to reproduce. During this stage, without further molts, the oostegites unfold and the ova are extruded. Ovarian activity continues in preparation for a possible second brood. Reproduction occurs despite the absence (early samples) or rarity of males raising the question of asexual reproduction. Some evidence for both protogynous and direct development of males is presented.

PAGE 109

A MODEL FOR THE INITIAL DEVELOPMENT OF A SESSILE ASSEMBLAGE 96 Initial colonization of hard benthic substrate has often been interpreted as a random series of events (Drury and Nisbet, 1973; Sutherland, 1974; Sutherland and Karlson, 1977). Just as often the events appear to lead to an ordered assemblage (Hewatt, 1935; Newcombe, 1935; Sheer, 1945; Northcraft, 1948; Aleem, 1957; Thorson, 1957; Wells, 1961; Haines and Maurer, 1980; Sutherland, 1980; Paine and Levin, 1981; Glasser, 1982; Mook, 1983; ) or one which reflects the composition of the larvae (Coe and Allen, 1937; Kitching, 1937; McDougall, 1943; Kawahara, 1965; Sutherland and Karlson, 1973; Sutherland, 1974; Dean, 1977; Osman, 1977; Smedes, 1979; Keough, 1983). At times, both convergence and divergence occurs (Sutherland and Karlson, 1973; Mook, 1983). Whether or not convergence or divergence was seen as the end result, the initial colonization appeared to be stochastically controlled. The models presented here are based on stochastic events ameliorated by defined intra and interspecific interactions. Connell and Slatyer (1977) have presented 3 models of successional pathways and species interactions. 1.) Facilitation: "Early occupants modify the environment so that it becomes less suitable for subsequent recruitment of early succession species but more suitable for recruitment of late succession species."

PAGE 110

97 2.) Inhibition: "Early occupants modify the environment so that it becomes less suitable for subsequent recruitment of both early and late succesion species." 3.) Tolerance: "Early occupants modify the environment so that it becomes less sui table for subsequent recruitment of early succession species but this modification has little or no effect on subsequent recruitment of late succesion species." Facilitation occurs when the presence of one species actually enhances the settling of another species. Self facilitation (gregariousness) can occur as well 1951, 1953a, 1953b; DeSilva, 1962; Crisp and Meadows, 1963; Meadows and Williams, 1963; Gallagher et al 1983). There is still a question as to whether this must lead to self destruction or whether it may lead to self perpetuation as well. For the purposes of the model, the recognition is made that both pathways are possible and they will both be lumped into the facilitation category. One aspect of succession which is often overlooked is the difference between changes in the assemblage due to succession and those due to development. This refinement recognizes the fact that, although one sere leads to the subsequent sere, there is a time factor involved in producing that second sere. This developmental stage is most pronounced in the initial colonization of newly available habitat space. In examining the initial development of assemblages, it is apparent that the models of Connell and Slatyer (1977) are functional but are slightly

PAGE 111

98 modified. 1.) Facilitation: The presence of a member of a facilitating species enhances the nearby settling of more individuals of the same or of different species. 2.) Inhibition: The presence of a member of an inhibitory species will preclude the nearby settling of more individuals of the same or of different species. 3.) Tolerance: The presence of an individual of a particular species has no effect on the nearby settlement of another individual of the same or of a different species. Using these definitions, it is apparent that a species can be a interspecific facilitator, or both. The same may be said for the other two models. It is also apparent that a single species can be or functional as well. Facilitation, inhibition, and tolerance requires an interaction between members of two species and these interactions are species specific. That is, species A may facilitate one species and inhibit another, while being tolerant of the third. Gallagher et al ( 1983) described many such dual roles in mud flat infauna. Standing ( 1976) found that Obelia dichotoma inhibited Balanus crenatus while facilitating the settlement of Ascidia ceratodes. A theoretical model was developed which combines the nature of larval availability and the deterministic nature of larval

PAGE 112

99 behaviour and species interactions. The model, incorporating these functions, is presented (fig. 18) as a flow diagram of discrete events in the colonization procedure. These events were then incorporated into a computer simulation of the process in order to test the outcome of the combined random and deterministic functions. Theoretical Model A hypothetical model of larval site selection was formulated (fig. 18). This model incorporates the concepts of Connell and Slatyer (1977) in the following manner. At any given point in time a single larva will encounter the substrate. This larvae is a semi:-stochastic product of larval availability and active behavioural site selection traits (Thorson, 1966; Meadows and Campbell, 1972; Crisp, 1974; Scheltema, 1 97 4) At the inti tiation of colonization there are no residents and the habitat is totally homogeneous. Therefore there are no forces acting to deflect site selection within the habitat from random. Once a single larva has settled, that portion of the habitat in the vicinity of the settled larva is no longer equitable with the rest of the substrate. In the near vicinity of the resident, the forces inherent in Connell and Slatyer's (1977) models, as interpreted here, are active. Throughout the rest of the substrate, homogeneity and random selection remain the dominant factors shaping the events. The process of larval arrival and site selection is continued until the habitat is saturated.

PAGE 113

Figure 18. Theoretical model flow diagram. Initially the habitat is equivalent to the substrate. When a single larva encounters the habitat, it may attempt to settle or depart pursuant to the fit of the habitat and niche hypershape. Facilitation and inhibition may induce settlement or rejection of the new arrival. The presence of thfe new species alters the habitat hypershape and such an alteration may or may not affect the fit with respect to the niche hypershape of the subsequent larval arrival.

PAGE 114


PAGE 115

102 The Computer Simulation The computer simulation is based on several premises. 1.) The composition of the larval pool is random within certain limitations. Specifically, it is recognized that the actual composition of the larvae at any point in time is basically a semi-stochastic function pursuant to larval behaviour and seasonal reproductive cycles. While abundances for large areas may be catagorized along a spectrum from rare to e xtremely abundant, in a particular, highly localized, water mass, the composition at any point in time may vary a few percentage points. 2.) From the larval pool, only certain larvae will attempt settling on a given substrate. This segment of the larvae is composed of those individuals which are physiologically prepared to settle and whose niche hypershape fits the habitat hypershape of the available substrate. The larval pool represents only those larvae which would find the habitat acceptable if it were not for these three biotic interactions. 3.) The substrate represents a newly available isolated (Keough, 1984) habitat which is homogeneous. From these beginnings, the simulation works through a series of steps in constructing the assemblage. 1.) The larval pool is created by randomly selecting species (coded If the abundance limitation for that category has been exceeded, the selection is rejected and a new one made until the array of available larvae (larval pool composition) has been filled.

PAGE 116

103 2.) The habitat is prepared as a matrix of site loci. The habitat is assumed to be homogenous and all sites are available for colonization. 3.) A single larva is randomly chosen from the larval pool. Since the pool is composed of larvae with differing levels of abundance, this does not mean that each species has an equal chance of selection but that each individual larvae of the pool has an equal chance of selection. 4.) A site is randomly selected. 5.) At this point in the program a branching occurs. The selected larva is checked for facilitation, inhibition, and tolerance (table 3). Inhibition is checked first. If the site selected has adjacent sites with larvae present, the new larva is rejected if the species code of the larva at the adjacent site is 1 of the new larva. That is species f/3 would be inhibited if an adjacent site is already occupied by species f/2 or f/4. It is apparent that species #10 is only inhibited by species 119 and therefore has less inhibition than the others. It is also apparent that species 111 is inhibited by open space and cannot colonize until late in the development (late succession species). Facilitation is then checked. Species #5 and 118 were selected as If the new larva is one of these two then sites up to 5 sites away from the initial site are checked for concodified species. If a site is occupied by a larva of the same code, the new larva will select the nearest open site, discarding the original site. If none is found, the larva will settle at the original site, barring inhibition. Facilitation takes precedence over inhibition if a conflict arises. Even if the initial site is occupied, the new larva, if it is a #5 or #8, will check nearby sites for facilitation. There are only two instances which result in rejection of a self-facilitative species. The new larva will

PAGE 117

104 Tabel 3 Interactions between species. + = facilitation, = inhibition, and o = tolerence. Species 111 was also totally preempted by open space. Species code 2 3 4 5 6 7 8 9 10 1 + + + .-" + + + + + + 2 0 :... 0 0 0 0 0 0 0 3 0 0 0 0 0 0 0 4 0 -;-0 0 0 0 0 5 + 0 0 0 0 6 0 !-0 0 0 7 0 0 0 8 + 0 9 0 10 0

PAGE 118

105 be rejected only if: 1.) the original site is occupied and there is no facilitative action 2.) the original site is unoccupied and inhibited and there is no facilitation. Tolerance occurs if both the checks of facilitation and inhibition return null functions. The computer program was run several times to examine the results of modifications of the original parameters. For all runs, larvae and sites were selected at random from the larval pool. In the first run (C1), all behavioral functions (facilitation, inhibition, and tolerance) were deleted. The result was based upon random larval and site selection, the only criterion for settling being that the site chosen was unoccupied. In the second set of runs facilitation, inhibition and tolerance were active. Species /15 and 118 were selected as Species /11 was inhibited by open space and species /110 had reduced inhibition. All others were inhibited by species whose codes were 1 of their own code. In the third set (C3-1 the larval pool was altered with respect to relative abundances when the substrate was 20% saturated emulating a change in larval abundance due to seasonal changes. Finally, in runs the larval pool was altered when the site matrix had reached 50% saturation. In each run, the initial larval pool was approximately equivalent with respect to relative abundances of species. The actual abundances were randomly chosen, but the limitations resulted in similar pools The resulting assemblages and those occurring when the site matrix was partially saturated (20 levels of saturation were chosen for

PAGE 119

106 Table 4. Initial composition of the larval pool (abundance 1) and the abundance of each species in the pool after alteration (abundance 2). notation indicates range. code abundance abundance 2 340 23 50 10 2 20 0 375 75 3 40 0 375 75 4 60 0 300 0 5 340 1 4 40 0 6 100 0 300 0 7 120 0 120 0 8 312 20 20 0 9 328 15 60 0 10 325 30 200 0

PAGE 120

107 analysis) were subjected to Bray;...curtis Similarity analysis (Bray and Curtis, 1957) and Group Averaging Ordination (Cody, 1974). Results In all runs of the computer simulation, species II 1 ,5,8,9, and 10 were nearly equal in abundance within the available pool. The abundance level limitations placed on the random selection sequence were chosen such that the species which were to be subjected to biotic interactions would initially have equitable abundances making the results more easily interpretable. Run C1 (no biotic interactions) resulted in an assemblage which had no discernable pattern with respect to the distribution of individuals within the site matrix. The composition of the final assemblage bore a strong resemblance to the available pool (similarity= 0.96). Taken as a control, this lack of pattern and high similar! ty to the larval pool indicated that the program functioned correctly with respect to the random forces built into the selection of larvae and sites. Runs and produced assemblages which steadily deviated in composition from the larval pool as the biotic functions became dominant over the stochastic selection functions. These two runs were treated as replicate developments and subjected to similarity and ordination producing the dendrogram shown in figure 19. In the final assemblages produced by these runs the species compositions were strongly influenced by the interactive biotic functions. Species /11 was severely depressed (76% decrease from the larval pool) as it was only able to settle in sites with no adjacent sites which

PAGE 121

Figure 19. Dendrogram of similari tes produced by run C2:-1 of the computer simulation. The sample code is composed of a letter corresponding to the replicate and a number corresponding to the level of saturation in percent sites occupied.

PAGE 122

1 9 3-A 3-8 2-8 2-A 1-8 1-A 5-A 6-A 4-A 4-8 5-8 7-A -8-A '#. 9-A _J 10-A w 6-8 > 7-8 w _J 9-8 z 10-8 0 8-8 .... 20-A < a: 20-8 ::::> 30-A .... 30-8 < C/) 40-A 40-8 50-A 50-8 60-A 60-8 70-A 80-A 70-8 80-8 90-A 100-A 90-8 100-8 8 SIMILARITY 7 6 .5 .4 .3 2 -------., I I I I I I :--------, I I I I I t--------------109

PAGE 123

110 were uninhabited. Low abundance species (codes 2,3, and 4) had final densities close to those of the larval pool. Mid abundance species (codes 6 and 7) were moderately influenced, usually negatively, with respect to initial abundances. species exhibited significant increases over the initial levels. Species #9 and #10 had lowered abundances but #10 was depressed less than #9 was. During runs C}-1, C4:-1, and C4:-2 the composition of the initial pool was altered (table 4) in order to examine differences in development brought about by changes in larval availability. In the C3 runs the availability was altered after the habitat was 20% saturated and in the C4 runs, after 50% saturation. The C3 runs produced assemblages which were much nearer to the C4 runs (mean similarity = 0.83) than to the C2 runs (mean similarity C2-C3 = 0.53) (table 5). In all runs, the level of similarity after 10% saturation (table 5) indicated that development was parallel prior to alterations in larval availability. Despite the severe reduction in availability of species #5 and #8 during the C3 and C4 runs, these self-facilitators increased more rapidly than would be expected in stochastic models, quadrupling (after reduction of availability) in runs C3 and doubling in C4. There are three species (II 5,8,and 10) whic h consistently formed large patches (fig. 20). Species 115 formed 26 patches leaving 17 individuals isolated (n = 229) Species #8 formed 20 patches leaving 13 isolated (n=189). Species #10, a non""facilitator, formed 29 patches leaving 28 isolated (n=140). Mixtures of species other than these three (II 5 8, and 1 0) also formed patches but individuals of the other 7 species, when taken alone, did not.

PAGE 124

111 Table 5 Similarity levels among six developing simulated assemblages. C2 models were allowed to develop normally. C3 and C4 models underwent changes in larval availability at 20% and 50% site saturation levels respectively. The upper right sector contains similarity values at 10% site saturation and the lower left, at 100% saturation. C2:-1 C2:-2 C3;.:.1 C3F 2 C4r-1 C4:-2 C2!0:1 '- 7 4 81 .73 78 1.0 C2-2 95 85 .90 79 74 C3,..;1 .54 .51 -.81 79 .80 C3.-2 .55 .52 .93 .. .76 73 .70 .69 .81 .80 "" .78 C4,...2 .70 67 83 .88 .93 ;...

PAGE 125

Figure 20. The site matrix as filled by the computer simulation run C2-1. The shaded area are sites occupied by species 115 (self-facilitating). Other species of particular interest {1/8 and 111 O) are annotated with their species codes. The blank areas annotated with an M are single individuals or mixtures of the remaining 7 species.

PAGE 126

11 3

PAGE 127

11 4 Discussion All runs of the computer simulation, with the exception of C1 (random) produced assemblages which exhibited developmental characteristics and aggregation patterns similar to those found in nature. The developmental pattern (fig. 19) is remarkably similar to the patterns produced by the fouling assemblage (fig. 10) and the microalgal assemblage (fig. 6). Similarity between replicates at equivalent stages of development is initially low and partially isolated from the successively tighter mid and late stage assemblages. Initially all experimental r uns developed along the same pathway as indicated by the matrix of similarities at the 1 O% saturation level (table 5). Replicate runs remain convergent (mean similarity between final replicates at termination= 0.937. Table 5). When larval abundances are changed in the middle of a run (C3 and C4), the only function which is altered is the probability of selection of each species of larva. The effect on the development is dramatic. If the C2 runs abundances) are considered to exhibit the normal sequence of events, then the development of the other, altered assemblages are divergent. That is, the direction of the development changes. Synergistic with with this divergence is the level of site saturation at which the larval availability was altered. C4 runs (altered at 50% saturation) produced final assemblages which have a higher degree of similarity to the C3 runs (altered at 20% saturation) than to the normal runs (C2). This occurred despite the fact that the C4 runs had spent more developmental time under the influence of the initial

PAGE 128

115 larval pool. This tendency of the C4 runs to converge with the C3 runs rather than the C2 set is the result of complex interactions between stochastic selection and behavioral selection functions. As the first larva occupies a site, that site and its 8 adjacent sites are no longer homogeneous with the rest of the habitat. They are controlled by inhibition. That single site and its 120 nearest sites are controlled by self-facilitation only if the first species was a self-facili tater and the new arrival is concodified. The control then may be as little as 0. 9 % and as much as 12% biologically controlled after only a single individual has settled (88 to 99.1 % remain stochastic). The latter is highly dependent upon the identity of the new arrival and represents an overestimate of total biotic control. As the degree of site saturation increases, more and more of the control is passed from the realm of randomness to that of ordered, biological interactive control. As the degree of site saturation increases a greater and greater percentage shifts from random control to biotic control and the assemblage shifts from one reflecting the larval availability to one reflecting the active interactive forces of the species present. Examination of the final assemblage compositions indicated that the degree of biological control with respect to individual species is also dependant to a degree upon the abundance of the species in question. Low density species (II 2,3,4) exhibited abundances in the final assemblages which were similar to those in the larval pool. Species #4 was only slightly depressed if at all. All three of these species are subject to inhibition alone. Four of the five inhibitory species are low density species as well. Species #1 is low because of its own inhibition

PAGE 129

11 6 by open space, # 2,3, and 4 are low because of the low abundance in the larval pool. The last inhibitor (#5) inhibits only species #4. Since #5 is drawn into aggregates, its distribution throughout the sites is not as widespread in early stages as it would be if it were not gregarious. These factors all serve to reduce the chance of an inhibitory encounter, leaving the abundance of species #4 similar to what it is in the larval pool. Higher, late stage abundances of species 115 effects the small reduction seen in species #4. Moderately abundant species (# 6 and 7) were both slightly depressed with respect to availability. This was due to a combination of their inhibition of each other and the inhibition of the very abundant species #5 and 118. The last two species (II 9 and 10) were heavily and moderately depressed respectively. Species 119 is inhibited by species #8 (very abundant) and by species #10. Species #10 is highly abundant in the larval pool and has half the inhibition factor of 119 as it is only inhibited by 1 species where all others (except #1) are inhibited by 2 others. Increases in abundance over that of the larval pool was seen in both self.:..facil i tating species (II 5 and 8). In all runs, facilitation significantly increased final levels over those in the pool as well as over non facilitators (# 9 and 10) which had the same larval levels in the pool. Facilitation also produced aggregations in the distribution of the species among the sites (fig. 20). When the availability of these self-facilitators was reduced in runs C3 and C4, patches still formed and settlement still exceeded that of non facilitators. The facilitation

PAGE 130

11 7 function enhanced settling, overriding random functions and concentrated the sites chosen into aggregates. There are two mechanisms operating to produce aggregations: Abundance and gregariousness (self-facilitation). Considering an aggregation to be a group of concodified species, it is apparent (fig. 20) that three species exhibit this trait. Species #10 does so because a large number of individuals have settled and they cannot avoid adjacency simply due to their abundance relative to the availability of sites. The other two aggregate forming species (# 5 and 8) are gregarious and form groupings due to the behavioral factors involved in selecting a site. Even when the program was run with low levels of these latter 2 species, they consistently formed clusters. That these aggregates are formed by seperate mechanisms is supported by the fact that in species # 5 and 8, significantly more clusters were seen than were isolated individuals (26 patches: 17 isolated and 20 patches: 13 isolated respect! vely). Species #10 produced 29 patches and 28 isolated individuals. Patch size was also much larger in species # 5 and 8 than in 1110. The % of all individuals within a patch was 92 and 93% for species # 5 and 8 and 80% for #10. Mean patch size was 8.0 and 8.75 for species #5 and 8 and only 3.8 for species #10. The aggregates formed which consisted of mixtures of species other than # 5,8, and 10 are an artifact of the self.-facilitation and are also found in nature. The gregarious mechanism operating in this instance draws gregarious arrivals towards existing groupings of the species leaving open space at the edges which are then available for the non -facilitative species. It is in this manner that an assemblage develops a congregation of a single species in one area while leaving

PAGE 131

118 another nearby area filled with a diverse mixture of species with lowered attributes. As a test of the flexibility and applicability of the simulation, initial abundances of the self::.facilitating species were lowered to 10% of the overal composition to test the effect of gregariousness on low density larvae. The result was that these species were still more abundant in the final assemblage than in the larval pool. Increases were generally double to quadruple the random values. Aggregates still formed but were smaller in size and number. Cross-facilitation or interspecific facilitation was added with the result that both species were enhanced in their final abundances and the patches which formed were equitable mixtures of species. Patch sizes were larger due to the selecting a facilitating species when the facilitators. the two interacting increased chance of two are reciprocal The last modification was to completely change the specific composition of the larval pool. In the first model, only the relative abundances were altered, in this last modification, all species of the original pool were deleted and 10 totally new species were selected. This resulted in a much more radical divergence in development. It is apparent that intermediate alterations (retaining some species and adding other new species) would result in intermediate levels of divergence. This model contains no function for individual death, population senescence or obliteration due to seasonal or habitat changes, intrinsic reproduction, represent a development. competetive death, good first attempt or at predation. It does, however, computer modeling o f assemblage

PAGE 132

119 The results of the simulation closely corresponds to the events seen in tropical and subtropical environments (fig. 6, fig. 10, Mook, 1983; Sutherland, 1980) when the modifications to the larval pool are minimal or kept within the same specific composition. It also reproduces events observed in systems where a strong structurally dominant species directs assemblage development (Hewatt, 1935; Newcombe, 1935; Haines and Maurer, 1980; Paine and Levin, 1981). When new species are added to the availability list and the pool is altered at different stages of development, the patterns and divergences seen more closely resemble those events documented in temperate zone studies (McDougall, 1943; Kawahara, 1965; Sutherland, 1973; Sutherland and Karlson, 197 4, 1977; Dean, 1977; Osman, 1977; Smedes, 1979) It may be, therefore, that the non-directional and divergent nature of developmental pathways observed in temperate zones is due to the disruption of larval availability. Such a disruption may be best interpreted as a shift from one developmental pathway to another rather than as an indication of stochastic control of single pathways. The effects of varying the area of available habitat (disturbance patch size) were examined only briefly with the computer program but the initial results indicate that the effects noted by Paine and Levin (1981) and Keough ( 1984) would be reproduced as well. The results of the simulation predict that small patches would become quickly filled with the most abundant larval form which corresponds to the Bryozoan dominance reported by Keough (1984). This is because, the stochastic larval selection forces dominate the developmental pathway until site saturation is approximately 20%. By that time, in small patches originally colonized by bryozoa (generally inhibitors of ascidia), the force of

PAGE 133

120 inhibition becomes dominant and the ascidia are rejected. Because the distance at which one species is inhibited by the presence of another remains constant, the percent area inhibited with respect to the percent colonized by the bryozoa increases as the size of the patch decreases. In larger patches, subjected to the same initial concentrations of larvae, there are enough sites which remain uninhibited for the ascidia to establish themselves. Summary Drawing on both the stochastic and deterministic schools of thought, a computer simulation program was devised which reproduces ordination and patch patterns seen in many primary and secondary successional sequences. Modifications of model input parameters, emulating seasonal alterations of larval abundances, produced ordinations which closely resemble those from successional studies undertaken in temperate zones. The ability to reproduce patterns in nature verifies, to a degree, the processes programmed into the simulation and l eads to the following conclusions. 1 ) Successional events are predictable providing a sufficient amount of initial information is available. 2.) Stochastic control of development is limited to the very early stages of colonization. 3 ) Biological interactio ns are the true driving forces in middle and late stage development. 4.) Seasonal changes are best interpreted as the results of a shifting from one successional chain of events to another.

PAGE 134

1 21 GENERAL DISCUSSION Predictability, and therefore, reproducibility are essential elements of successional theory (Odum, 1971). And yet it has often been noted that these are apparently lacking in the development of, and progression through, seral assemblages of the epibenthos on hard substrata. The foregoing sections of this work have revealed the sequence of events in the development of an epibenthic assemblage colonizing natural sustrate in a subtropical estuary. Within hours of submergence dissolved organic material (DOM) adsorbs to the surface of the substrate. The rate of adsorption as well as the exact chemical character of the film is dependant upon ambient concentrations of the DOM and the charge distribution inherent in the chemical matrix of the substrate involved (Danials, 1979). This event, while it may not be a successional requirement for subsequent colonization, certainly enhances the habitat with respect to microbial species. These microbial species, dominated by bacteria and diatoms, constitute the second event in the observed sequence. It is apparent from the data that neither of these microbial types are requisites for the other. Studies using vertically oriented substrate (Zobel and Allen, 1935; Corpe, 1973, 1974; Caron and Sieburth, 1981a) have often found bacteria to precede diatoms but that is a result of the orientation of the substrate rather than of successional interelationships. Vertical

PAGE 135

122 surfaces will not be as susceptible to diatomaceous colonization simply as a result of particle settling physics and differences in the inherent properties of adsorption. Vertical and horizontal surfaces, even within the same water mass, constitute different habitats which will direct assemblage development into different successional pathways. Two notable events occurred in the microbial colonization sequence during this study. First, there was an apparent transition from early, colony forming, rod shaped bacteria colonizing fresh surfaces to non-colony forming ovoid and spiral bacteria seen in areas with large amounts of detritus. Corpe (1974) noted a similar transition from Gram negative rods to Caulobacter and Saprospira on glass slides. Secondly, the four microbial assemblages produced were not strictly similar at termination of the SEM phase of the study. Samples from experiment A did not converge with the other three (fig. 6) but that difference did not influence the nature of the meio and macrofauna! development. There are two possible explanations which are not necessarily exclusive of each other. The character of the microbial assemblage in experiment A could have converged after monitoring ceased or the exact nature of the assemblage may not be as important as the fact of its presence. While larvae are sensitive to chemicals and microbial films (Meadows and Campbell, 1972), most of the initial colonizers seen here are generalistic in their choice of food items or are filter feeders not dependant on the microbes. The latter, however, may well equate microbial development with substrate stability, a critical factor for sessile forms. Only after the substrate had been microbially prepared did meiofaunal and macrofauna! forms appear. The fact that these forms were

PAGE 136

123 available prior to microbial preparation verifies that preparations status as a successional requisite of macrofauna! assemblage development. The meiofaunal and macrofauna! species exhibited no set order in their arrival. Similarity between early replicates of the same experiment and among early replicates between experiments was lower than what would be expected if colonization was predictable. Even so, they converged rapidly. This lack of order in the extremely early arrivals cannot be interpreted as disorder in the successional pathways but must instead be viewed as a series of semi stochastic events in the development of the second sere in the sequence. It is a transition between the microbial sere and the first macrofauna! sere resolved by the short time interval between the taking of the samples. Replicates within experiments and samples between the two experiments converged with respect to exposure time as both experiments followed the same successional track even though they had been emplaced on different dates. Just before October 14 the temperature regime changed. This change impacted the two assemblages in like fashion even though they were at different levels of development with respect to exposure time. The overall effect of this thermal variance was to synchronize the development with respect to time of year instead of duration of immersion. The habitat had changed, but not so drastically as to comprise a disturbance. Instead it promoted synchrony. The synchrony with respect to time of year displayed in these samples can be traced directly to species abundances. H. rapax was chosen for detailed examination but most of the other species of the assemblage exhibited the same basic population parameters. Specifically, the presence or absence of these species was seen to be controlled by

PAGE 137

124 habitat suitability (succession) but the levels of abundance were more closely associated with the time of year (To). A smaller component of the convergence was attributable to winter immigrants and late successional species xanthiformis and G. robustum). In other areas where the alteration approaches the catastrophic, the assemblages would have experienced a successional track'"" shift. The compos! tion of the available colonizers would undergo a large scale alteration and they would be presented with substrate in different phases of development (different habitat hypershapes). Divergence in assemblage structure results (Sutherland and Karslon, 1973, 1977). Studies carried out in temperate zones, then, don't exhibit stochastic succession as much as periodic, large scale, disturbances which alter the entire course of succession. A single successional sequence must, by definition, occur within the confines of physical processes. In many areas, however, the physical parameters are not constant enough for successional events to be fully exhibited. A patchwork of successional segments is seen instead. In subtropical areas, as in this study, the pattern is allowed to run to completion, or at least to intraseral stabilization. Along the continuum of control proposed by Sanders (1968), the tropics and subtropics are within the range of biological dominance allowing successional events to be more clearly seen. In other areas, physical changes are great and successional developments are given little chance at unperturbed completion. The computer simulation of developing assemblages verifies facilitation, inhibition, and tolerance (Connell and Slatyer, 1977) as strong forces shaping the outcome of colonization but more importantly, perhaps, 1 t displays a level of flexibility which can account for the

PAGE 138

myriad of discordant results produced by fouling studies. 125 It is significant that the simulation commences under stochastic control but changes, internally and solely on the basis of Connell and Slatyer' s (1977) interactive models, to biological control. It is also educational in that it can produce many of the same results as have been noted in the literature, from deterministic to stochastic, simply by emulating a variation either in the intitial or configuration. Successional theories founding premises of predictability and reproducibility are strongly substantiated by these results. The criterion for expecting predictions to materialize are, however, tightened. "The total physical, biological, and chemical factors must be viewed as a single entity in order to understand the total concept of colonization on a newly submerged substrate. (Alfieri, 1975)" It is imperative that we become more adept at recognizing what parameters comprise differences in habitat. We must also recognize the difference between disturbances within a habitat and those which cause a complete shift from one habitat to another. Lastly, it must be noted that, in order to predict successional pathways, an intimate knowledge of the available species, their interelationships, and their requirements in a habitat is imperative. Drury and Nisbet (1973) wrote that "A comprehensive theory of succession should be sought at the organismic level and not in the emergent properties of the community." The succession that we see, then,

PAGE 139

126 is not an all encompassing order to biotic development and redevelopment but is a facade composed of the results of an intricate network of synergistic, somewhat probabilistic, biotic responses to physical and biotic pressures within large scale physical bounds. As such there is a certain degree of latitude inherent in the exact structures of seral assemblages and the time spent in each sere but little in the order of assemblage occurrence in undisurbed and unshifting habitats. Once the intricacies at the organismic level are more clearly understood, a more comprehensible theory addressing these emergent community properties can be fabricated.

PAGE 140

127 LITERATURE CITED Aleem, A.S. 1957. Succession of Marine Fouling organisms in test panels immersed in deep water at LaJolla, Cal. Hydrobiologica 11 Alfieri, D.J. 1975. Organismal development on an artificial substrate I. July 1972' ;..; 6 June 1974. Estuarine and Coastal Marine Science 3:4651-i472. Baier, R.E., E.G. Shafrin and W.A. Zisman. 1968. Adhesion: Mechanisms assist or impede it. Science 172 : 1360 E1368. Bell, S.S., M.C. Watzin and B.C. Coull. 1978. Biogenic structure and its effect on the spatial heterogeneity of meiofauna in a salt marsh. Journal of Experimental Marine Biology and Ecology Bell, s.s. 1979. Short and long term variation in a high marsh meiofauna community. Estuarine and Coastal Marine Science Bitton, G. and K.C. Marshall (ed). 1979. Adsorption of microorganisms to surfaces. J; Wiley and Sons, N.Y. 439pp. Boesch, D.F. and R.J. Diaz. 1974. New records of peracarid crustaceans from oligohal ine waters of the Chesapeake Bay. Chesapeake Science 15(1):56'-58. Bray, J.R. and J.T. Curtis. 1957. An ordination of the upland forest communi ties of southern Wisconsin. Ecological Monographs 27(4):325F349. Brenchley, G.A. 1981. Disturbance and community structure. experimental study of bioturbation in marine soft environments. Journal of Marine Research An bottom Brewer, R.H. 1978. Larval settlement behavior in the jellyfish aurelia aurita (Linnaeus) (Scyphozoa; Semaeostomeae). Estuaries 1 ( 2) : 1 1 22. Brown, T.J. and J.R. Sibert. 1977. Food of some benthic harpacticoid cope-pods. Journal of the Fisheries Research Board of Canada 34:10281-11031. Cairns, J. and M.S. Henebry. 1982. Interactive and noninteractive protozoan colonization processes; In: J. Cairns (ed). Artificial substrates. p. Ann Arbor Sci:-Ann Arbor, Mich. 279pp.

PAGE 141

128 Calder, D.R. and M.L. Brehmer. 1967. Seasonal occurrence of epifauna on test panels in Hampton Roads, Virginia. International Journal of Oceanology and Limnology Caron, D.A. and J.McN. Sieburth. 1981a. Disruption of the primary fouling sequence on fiber:-:glass reinforced plastic submerged in the marine environment. Applied and Environmental Microbiology 41 Caron, D.A. and J.McN. Sieburth. 1981b. Responses of peritrochous ciliates in fouling communities to seawaterr-accommodated hydrocarbons. Transactions of the American Microscopical Society H. 1960. Sex determination. In: T.H. Waterman (ed). The physiology of crustacea. p. Academic press, N.Y. Chi a, F .s. 1977. Perspectives: Settlement and metamorphosis of marine invertebrate larvae. lE= F.S. Chia and M.E. Rice (eds). Settlement and metamorphosis of marine invertebrate larvae. p. Elsevier, N.Y. 290pp. Chia, F.S. and L.R. Bickell. 1977. Mechanisms of larval attachment and the induction of settlement and metamorphosis in coelenterates: A review. In: F.S. Chia and M.E. Rice (eds). Settlement and metamorphosis of marine invertebrate larvae. p. Elsevier, N.Y. 290pp. Cody, M.L. 1974. Competition and the structure of bird communities. Princton Univ. Press. Princton, N.J. Coe, W.R. and W.E. Allen. 1937. Growth of sedentary marine organisms on experimental blocks and plates for nine successive years at the pier of the Scripps institution of Oceanography. Bulletin of Scripps Institute of Oceanography Technical Series Connell, J.H. 1961. The influence of interspecific competition and other factors on the distribution of the barnacle, Chthamalus stellatus. Ecology Connell, J .H. 1972. Community interactions on marine rocky intertidal shores Annual Review of Ecology and Systematics 3:169e:192. Connell, J.H. 1976. Competetive interactions and the species diversity of corals. In: G.o. Mackie (ed). Coelenterate ecology and behavior. p. N.Y. Connell, J.H. and R.O. Slatyer. 1977. Mechanisms of succession in natural communities and their role in community stability and organization. American Naturalist Connell, J.H. 1978. Diversity in tropical rain forests and coral reefs. Science

PAGE 142

129 Conner, E.F. and E.D. McCoy. 1979. The statistics and biology of the spectesHarea relationship. American Naturalist 11 3( 6) :791 Corpe, W.A. 1973. Microfouling: The role of primary film forming marine bacteria. Tn : R.F. Acker, B.F. Brown, J.R. DePalma, and W.P. Iverson (eds)-. Proc. 3rd Int. Cong. Mar. Corr. Foul. p. Northwestern Univ. Press. Evanston, Ill. 1974 Corpe, W.A. 1974. Periphytic marine bacteria and the formation of microbial films on solid surfaces. In: R.R. Colwell and R.Y. Morita (eds). Effect of the ocean environment on microbial activities. p. 397 M417. Univ. Park Press, Baltimore. Coull, B.C. 1973. Estuarine meiofauna: A review: Trophic relationships and microbial interactions. In: L.H. Stevenson and R.R. Colwell (eds). Estuarine microbial ecology. pp 499-"512. Univ. So. Carol. Press; Crisp, D.J. and H. Barnes. 1954. The orientation and distribution of barnacles at settlement with particular referenc e to surface contour. Journal of Animal Ecology 23:142F162. Crisp, D.J. and J.S. Ryland. 1960. Influence of filming and of surface texture on the settlement of marine organisms. Nature 180:119. Crisp, D.J. and P.S. Meadows. 1963. Adsorbed layers: The stimulus to settle in barnacles. Procedings of the Royal Society London Series B. Crisp, D.J. 1973. Mechanisms of adhesion of fouling organisms. In: R.F. Acker, B.F. J.R. DePalma, and W.P. Iverson (eds); Proc. 3rd International Congress on Marine Corrosion and Fouling p. 691:-!709. Northwestern Univ. Press. Evanston, Ill. 1974 Crisp, D.J. 1974. Factors influencing the settlement of marine invertebrate larvae. In: P.T. Grant and A.M. Mackie (ed). Chemoreception in marineorganisms. p. 177
PAGE 143

130 Dempsey, M.J. 1981b. Colonisation of antifouling paints by marine bacteria. Botanica Marina XXIV:185A191. DeSilva, P.H.D.H. 1962. Experiments on choice of substrate by Spirorbis larvae (Serpulidae). Journal of Experimental Biology 39 Dexter, S.C., J.D. Sullivan, J. Williams and s.w. Watson. 1975. Influence of Substrate wettability on the attachment of Marine Bacteria to various surfaces. Applied Microbiology Droop, M.K. and K.G.R. Elson. 1966. Are pelagic diatoms free from bacteria? Nature 2l1 :1096:<-:1097. Drury, W.H. and I.C.T. Nisbet. 1973. Succession. Journal of the Arnold Arboretum Harvard Univ. Eckelbarger, K.J. 1977. Larval development of Sabellaria floridensis from Florida and phragmatopoma californica from southern California (Polychaeta, Sabellariidae), with a key to the Sabellariid larvae of Florida and a review of the development of the family. Bulletin of Marine Science Fager, E.W. 1971. Pattern in the development of a marine community. and Oceanography 16(2):241M253. Fenchel, T.M. 1970. Studies on the decomposition of organic detritus derived from the turtle grass Thalassia testudinum. Limnology and Oceanography 15:14M20. Fischer'"'Piette, E. 1934. Sur l'equilibre des faunes: Interactions des moules, des et des cirripedes. Ccomptes Rendus de la Societe des Biogeographie No. 92. p. E. 1935. Histoire d'une mouliere. Bulletin Biologique de la France etla Belgique Fletcher, M. and G. Loeb. 1979. The influence of substratum characteristics on the attachment of a marine pseudomonad to solid surfaces. Journal of Applied and Environmental Microbiology 37:67ii.72; Fletcher, M. 1980. The attachment of bacteria to surfaces in aquatic environments. In: E.H. Beachey (ed). Bacterial adherence. p. Chapman and Hall, London. Floodgate, G.D. 1972. The mechanism of bacterial attachment to detritus in aquatic systems. Memoirie dell'Istituto Italiano di Idrobiologia 29 Suppl. Gallagher, E. D., P.A. Jumars and D.O. Trueblood. 1983. Facilitaton of benthic succession by tube builders. Ecology

PAGE 144

1 31 Gardiner, L.F. 1975. The systematics, postmar supial development and ecology of the family Neotanaidae (Crustacea: Tanaidacea) Smithsonian Contributions to Zoology #170 Smith Inst. Press. Glasser, J W 1982 On the causes modifi c-ation of the biotic 119(3):375..,390 of temporal change in communities: environment American Naturalist Goulder, R., E .J. Bent and A.C. Soak. 1981. Attachment to suspended solids as a strategy of estuarine bacteria. In: N .V. Jones and W.J Wolff (ed). Feeding a n d survival strategies Of estuarine organism s p. P lenum Press N .Y. Grant, P T. and A M Mackie (ed). 1974 Chemoreception in marine organisms. Academic Press, N.Y. Gray, J S 1966 The attractive factor of intertidal sands to Protodrilus symbioticus Giard. Journal of the Marine Biological Association of the United Kingdom Greve, L 1967 On the tube building of some Tanaidacea. Sarsia 29: Hadfield, M.G. 1 977 Metamorphosis in analysis o f s timulus and response. (eds). Settlement and metamorphosis p Elsevier, N .Y. 290pp. marine molluscan larvae: An I n : F.S. Chia and M.E. Rice or-marine invertebrate larvae. Haines, J.L. and D. Maurer. 1980. Quantitative faunal associates of the serpulid polychaete Hydroides dianthus, Marine Biology Hairston, N .G. 1959 Species abundance and community organization. Harger, 0. 1879 Notes o n New England Isopoda. Procedings of the United StatesNational Museum Heard, R .W. 1982 Guide to northeastern G ulf of Consortium common tidal marsh invertebrates of the Mexico. Mississippi Alabama Seagrant Hewatt W .G. 1935 Ecological succession in the Mytilus californianus habitat as observed in M onterey Bay, Ca. Ecology Highsmith, R.C. 1982. Induc e d settlement and metamorphosis of sand dollar (Dendraster excentricus) larvae in sites: Adult sand dollar beds Ecology Highsmith, R C 1983 Sex reversal and fighting behavior: Coevolved phenomena in a tanaid crustacean. Ecology Hopkins, T.L. 1977. Zooplankton distribution in surface waters of Tampa Bay, Fl: Bulletin of Marine Science 27(3):467;.;478

PAGE 145

132 Horn, H.S. 1974 The ecology of secondary succession. Annual Review of Ecology and Systematics Hudon, C., E Bourgetand P. Legendre. 1983. An integrated study of the factors influencing the choice of the settling site of Balanus crenatus cyprid larvae. Canadian Journal of Fisheries and Aquatic Sciences Jannasch, H.W. and P.H. Pritchard. 1972. The role of inert particulate matter in the activity of aquatic microorganisms. Memoirie dell'Istituto Italiano di Idrobiologia 29 Suppl. Jannasch, H.W. 1978. Microorganisms and their aquatic environment. In: W.E. Krumbein (ed). Environmental biogeochemistry and geomicrobiology. Vol. 1': The aquatic environment. p 17,:.;24. Ann Arbor Sci., AnnArbor Mi. 394pp. Jeffers, J.N. R 1978 An ecological applications. introduction to systems analysis: Edward Arnold press, London. 198pp. with Jones, G.W. and R E. Isaacson. 1983 Proteinaceous bacterial adhesins and their receptors. CRC Critical Reviews in Microbiology 10(3):229h 260 Jones, N.S 1950 Marine bottom communities. Biological Reviews of the Cambridge Philosophical Society Kawahara T 1965 Studies on the marine fouling communities. III. seasonal changes in the initial development of test block communities. Report of Faculty of Fisheries Prefectural University of Mie Keough, M J 1983. Patterns of recruitment of sessile invertebrates in two subtidal habitats. Journal of Experimental Marine Biology and Ecology Keough, M J 1984 Effects of patch size on the abundance of sessile marin e invertebrates. Ecology 65 ( 2): 423,.:,437. Kirchman, D., s Graham, D. Reish and R. Mitchell. 1982. Bacteria induce settlement and metamorphosis of Janua (Dexospira) brasiliensis Grube (Polychaeta: Spirorbidae). Journal of Experimental Marine Biology and Ecology 56:153F.163. Kitching, J.A. 1937 Studies in sublittoral ecology II. Recolonization of the upper margin of the sublittoral region. Journal of Ecology Knatz, G. 1978 Succession of copepod species in a middle atlantic Estuary. Estuaries 1(1):68f'71. Kneib, R.T A.E Stiver and E.B. Haines. 1980 Stable carbon isotope rati"os in Fundulus heterocli tus muscle tissue and gut contents from a North Carolina U .S.A. Spartina alterniflora marsh. Journal of Experimental Marine Biology and Ecology 46( 1) :89.-98.

PAGE 146

133 Knighti-iJones, E.W. 1951. Gregariousness and some other aspects of the settling behavi"our of Spirorbis. Journal of the Marine Biological Association of the United Kingdom E.W. 1953a. Laboratory experiments on gregariousness during settling tn Balanus balanoides and other barnacles. Journal of Experimental Biology E.W. 1953b. Decreased discrimination during settling after prolonged -planktonic life in larvae of Spirorbis borealis (serpulidae). Journal of the Marine Biological Association of the United Kingdom E.W. 1953c. Some further observations on gregariousness in marine larvae. Brittsh Journal of Animal Behaviour 1 :81f.:>82. E.W. and J. Moyse, 1961, Intraspecific competition in sedentary marine animals. Symposia of the Society for Experimental Biology Kriss, A. Y. 1959. Marine microbiology (deep water) Izvestiya Akademii Nauk USSR, Moscow. In Russian: Translated by J .M. Shewan and Z. Kabata, Oliver and Boyd Edinburgh. 529pp. Lang, K. 1953a. Apseudes hermaphroditicus n. sp. A hermaphroditic tanaid fromthe antarctic. Arkiv for zoologi, series 2, Lang, K. 1953b. The postmarsupial development of the Tanaidacea. Arkiv for zoologi, series 2, Lang, K. 1973. Taxonomische und phylogentische untersuchungen uber die Tanaidace en (Crustacea). Zoolica Scripta 2: Lee, W.Y. and B.J. McAlice. 1979. Seasonal succession and breeding cycles of three species of 'Acartia (copepoda: Calanoida) in a Maine estuary. Estuaries Lewis, C.A. 1977. A review of substratum selection in free., and symb'iotic cirripeds. In: F.S. Chia and M.E. Rice (eds). Settlement and metamorphosis of-marine invertebrate larvae. Elsevier, N.Y. 290pp. Luna, L.G. (ed). 1968. Manual of Histological Staining Methods of the Armed Forces Institute of Pathology. 3rd ed. McGraw Hill, New York. MacArthur, R.H. and E.O. Wilson. 1963. An equilibrium theory of insular zoogeography. Evolution Margalef, R. 1968. Perspectives in ecological theory. University of Chicago press, Chicago. 111pp. Marshall, K.C. and Cruickshank. 1973. Cell surface hydrophobicity and the orientation of certain bacteria at interfaces. Archives of Mikrobiology 91

PAGE 147

134 May, R.M. 1976. Patterns in communities. IN: R.M. May (ed). Theoretical Ecology. Principles and applicattons. p. Saunders, Philadelphia. 317pp. McDougall, K.D. 1943. Sessile marine invertebrates at Beaufort, North Carolina Ecological Monographs 13(3) :321kl374. Mcintosh, R P. 1967. An index of diversity and the relation of certain concepts to diversity. Ecology Mcintosh, R.P. 1980. The relationship between succession and the recovery process and ecosystems. In: J. Cairns (ed). The recovery process in damaged ecosystems. p. Ann Arbor Sci., Ann Arbor, Mi. 279pp. Meadows, P.S. and G .B. Williams. 1963. Settlement of Spirorbis borealis Daudin larvae on surfaces bearing films of microorganisms. Nature 1 98 : 61 Meadows, P.S. and J.J. Campbell. 1972. Habitat selection by aquatic invertebrates. Advances in Marine Microbiology Mihm, J.W., W.C. Banta, and G.I. Loeb. 1981. Effects of adsorbed organic and primary fouling fi1rrrs on bryo zoan settlement. Journal of Experimental Marine Biology and Ecology Mook, D .H. 1981. Effects of disturbance and initial settlement on fouling community structure. Ecology Mook, D.H. 1983. Responses of common fouling organisms in the Indian River Florida, to various predation and disturbance intensities. Estuaries Morse, D.E., N. Hooker, H. Duncan and L. Jensen. 1979. acid, a neurotransmitter, induces planktonic abalone larvae to settle and begin metamorphosis. Science Neihoff, R. and G. Loeb. 1973. Molecular fouling of surfaces in seawater. In: R.F. Ack"er, B.F. Brown, J.R. DePalma, and W.P. Iverson (eds); Proc. 3rd Int. Cong. Mar. Corr. Foul. p. 710l-7718. Northwestern Univ. Press. Evanston, Ill. 1974. Neihoff, R. and G. Loeb. 1974. Dissolved organic matter in seawater and the electric charge of immersed surfaces. Journal of Marine Research 32:5'"'12. Nelson, W.G. 1979 Additions to the amphipod crustaceans of North Carolina. Estuaries Neumann, R. 1979. Bacterial induction of settlement and metamorphosis in the planula larvae of Cassiopea andromeda (Cnidaria: Scyphozoa, Rhizostomeae). Marine Ecology Progress Series 1(1):21 :28. Newcombe, C.L. 1935. A study of the community relationships of the sea mussel; Mytilus edulis L. Ecology 16:234-243.

PAGE 148

135 Northcraft, R.D. 1948. Marine algal colonization on the Monterey Peninsula; California. In: D.J. Reish (ed). Biology of the oceans. p. 29>44. Dickerson Pub.Co., Belmont. 236pp Odum, E.P. 1971. Fundamentals of ecology. 3rd ed. W.B. Saunders Co., Phnadelphia. 574pp. Odum, W.E. and E.J. Heald. 1972. Trophic analysis of an estuarine mangrove communi'ty Bulletin of Marine Science Oppenheimer, C.H. 1960. Bacterial activity in sediments of shallow marine bays. Geochemica et Cosmochimica Acta. 19:244 :260. Osman, R.W. 1977. The establishment and development of a marine epifaunal commonity. Ecological Monographs Paine, R.T. and S .A. Levin. 1981. Intertidal landscapes: disturbance and the dynamics of pattern. Ecological Monographs 51(2):145F-178. Patten, B.C. 1971. A primer for ecological modeling and simulation with analog and di'gi tal computers. In: B.C. Patten (ed). Systems analysis and simulation in ecology. Academic press, N.Y. Preston, F.W. 1962. The canonical distribution of commonness and rarity: Part I ; Ecology Part II. Ecology Renn, C.E. 1964. The bacteriology of interfaces. In: H. Heukelekian and N.c. Dondero (eds). Principles and applications in aquatic microbiology. p. J. Wiley and sons, N.Y. Roubault, A. 1937. Dimorphisme et croissance chez un Tanaidace. Travaux de la Station biologique de Roscoff. Sanders, H.L. 1968. Marine benthic diversity: A comparative study. Ameri c an Naturalist 102(925):243H 282. Santos, S.L. and J .L. Simon. 1980. Response of soft-7-
PAGE 149

136 Shelford, V.E., A.O. Weese, L.A. Rice, D.I. Rassmussen and A. MacLean. 1935. Some marinE! biotic communi ties of the Pacific coast of North America. Ecological Monographs Sieburth, J.McN. 1968. The influence of algal antibiosis on the ecology of marihe mtcroorganisms. In: M.R. Droop and E.J.F. Wood (eds). Advances in microbiology of-rhe sea; Acad. Press, N.Y., p. 63F;94. Sieburth, J. MeN. 1975. Microbial seascapes. Univ. Park Press. Sieburth, J. MeN. 1976. Bacterial substrates and productivity in marine ecosystems. Annual Reviews of Ecology and Systematics 7:2598 285. Smedes, G.W. 1979. Succession and community. 220pp. U. Delaware Jnter.nat. Ann Arbor, Mich. stability in dissertation. a marine fouling Univ. Microfilms Smith, G.W. 1906. High and low dimorphism, with an account of certain Tanatdae of the bay of Naples. Mi ttleilungen aus der Zoologischen SLation zu Neapel .-' Standing, J.D. 1976. Fouling community structure: Effects of the Obelia dichotoma, on larval recruitment. In: G.O. Mackie (ed). Coelenterate ecology and behavior. p. 155t-:l64-.-Plenum Press, N y Sutherland, J.P. and R.H. Karlson. 1973. Succession and seasonal progression in the fouling community at Beaufort, North Carolina. In: R.F. Acker, B.F. Brown, J.R. DePalma, and W.P. Iverson (eds); Procedings of the 3rd International Congress on Marine Corrosion and Fouling p. 906H926. Northwestern Univ. Press. Evanston, Ill. 1974. Sutherland, J.P. 1974. Multiple stable points in natural communi ties. American Naturalist Sutherland, J.P. and R.H. Karlson. 1977. Develompent and stability of the fouling community at Beaufort, N.C. Ecological Monographs 47(4):425'-'446. Sutherland, J.P. 1980. Dynamic of the epibenthic community on roots of the mangrove Rhizophora mangle, at Bahia de Buche, Venezuela. Marine Biology Thorson, G. 1957. Bottom communities. In: J. Hedgepeth (ed). Treatise on marine ecology and paleoecology. Vol. 1. p. 461r-534 Geological Society of America Memoir 67. Thorson, G. 1966. Some factors influencing the recruitment and establishment of marine benthic communi ties. Netherlands Journal of Sea Research Visscher, J.P. 1927. Nature and extent of fouling ships' bottoms. Bulletin of the United States Bureau of Fisheries

PAGE 150

137 Wells, H.W. 1961. The fauna of oyster beds, with special reference to the salinity factor. Ecological Monographs Wells, H.W., M.J. Wells and I.E. Gray. 1964. The calico scallop community in North carolina. Bulletin if Marine Science Gulf and Caribbean 14(4):561<-:593. Wisely, B. 1960. Observations on the settling behaviour of larvae of the tubeworm, Spirorbis borealis Daudin (Polychaeta). Australian Journal of Marine and Freshwater Research Wood, E.J.F. 1967. Microbiology of oceans and esutuaries. Ch. VIII. Some economical aspects of water microbiology. P. Elsevier, New York. 319 pp. Young, L.Y. and R. Mitchell. 1974. The role of chemotactic response in primary film formation. In: R.F. Acker, B.F. Brown, J.R. DePalma, and W.P. Iverson (eds). Procedings of the 3rd International Congress on Marine Corrosion and Fouling 1972 p. Northwestern Univ. Press. 1974. Zebell, C.E., and E.C. Allen. 1935. The significance of marine bacteria in the fouling of submerged surfaces. Journal of Bacteriology 29: 2391":251 Zebell, C.E. and D. Anderson. 1936. Observations on multiplication of bacteria in different volumes of stored seawater and the influence of oxygen tension and solid surfaces. Biological Bulletin of Woods Hole 71 Zebell, C.E. and C.B. Feltham. 1937. Bacteria as food for certain marine .Invertebrates; Journal of Marine Research 8 : Zebell, C.E. 1943 The effect of solid surfaces upon bacterial activity. Journal of Bacteriology Zebell, C.E. 1946. Marine microbiology. 240 pp. Chronica Botanica, Waltham Mass.

PAGE 151


PAGE 152

139 PROGRESSIVE SIMILARITY TESTING: AN EVALUATION OF SAMPLING SUFFICIENCY. In an ecologically uniform area, each new sample taken returns fewer and fewer new species to the data set and the slope of the speciesharea (SHA) curve approaches zero. Preston (1962) and McArthur and Wilson (1963) have defined this power function as where S = the cumulative number of species encountered after a given area (A) has been sampled. The units of the variable A have been expressed in terms of area (ranging from m2 to Km2 ) or the total number of individuals sampled (N). The observed reduction in the rate of occurrence of new species in new samples has often been viewed as an indication that the number of samples taken provides a reasonable reflection of the actual assemblage under investigation (Hairston, 1959 ; Preston, 1962; Mcintosh, 1967). Typically, either the point of greatest curvature or the point at which the slope of the curve falls below 0.75 has been interpreted as the point at which further sampling is unwarranted. The point of maximum curvature is, mathematically, the point of the most rapid decrease in the rate of return of new species. The point at which the slope o. 75 usually falls to the right of the point of maximum curvature and therefore indicates that a more complete

PAGE 153

140 data set has been obtained. The use of the curve as a means of determining sampling sufficiency represents a gross oversimplification of the biotic relationships involved (Conner and McCoy, 1979) but remains a functional application. Depending on the assemblage under consideration, a variety of curves are produced. If the actual assemblage contains very few species, the curve becomes saturated (reaches a plateau) after only a few samples have been taken and sampling may be terminated before distributional parameters have been adequately defined in the data set. If the actual assemblage contains only a few dominant species and a very large number of very rare species, the Si-iA relationship (non;.;plateau producing) will indicate that a very large number of samples must be taken to ensure adequacy of sampling. The proposed method of testing for sampling sufficiency is designed to be more sensitive to the distributional facets of the assemblage. In any sampling scheme, with the exception of total censusing, samples or accumulated data which exactly duplicate the actual assemblage are never obtained. Instead a working model of the assemblage (an approximation) i.s constructed based on the sample data. This model, ideally, should relay enough information about the actual assemblage for its ecologic and population parameters to be estimated. The model can only do so if sampling has been sufficient. In analysing the sampling procedure, it is apparent that if only one replicate were taken, the data would be less reliable than if a second were taken, and so on. The Progressive Similarity test is based on the supposition that the researcher can evaluate his effort by examining

PAGE 154

1 41 the parameters of the data set as the data is gathered. This is the same assumption used in justifying the use of the S'".A curve in determining sufficiency. In the method, replicate 111 (R1) is deemed equivalent to model assemblage 111 (M1). M2 is the average of the data contained in R1 and R2. The third model assemblage (M3) is the average of the data from R1, R2, and R3, and so on until a series of models are constructed which represent sequential stages in the accumulation of information. Each successive model in the sequence is based on a sample containing one more replicate than the previous one. Often the models will contain portions of individuals, a situation found objectionable by some biologists (Sanders, 1968). It is acceptable in this test as the analysis is not a prediction of an expected assemblage but remains simply a mathematical test for sufficiency. Using the full data set, the Brayi'\Curtis (Bray and Curtis, 1957) similarities are then calculated for each successive pair of model assemblages. The results are interpreted as a series of of similarities reflecting the change in total information (or lack of it) which resulted from the addition of each new replicate. In this manner, the investigator is able to ascertain the potential value of continued sampling. If any given model shows a high degree of similarity with the previous model, the last replicate has not significantly contributed to the information returned in the new model. The level of similarity commonly deemed significant (by convention) in assemblage is 0.5. In analysing various data sets produced by faculty and students at the University of South Florida, I have ascertained that similarities above 0.9 between successive samples

PAGE 155

1 42 is more appropriate as a level of significance. To verify the utility of this test, an assemblage of known composition (S = 34 ; N = 504) was constructed and sampled. Both S -'-A and p,-.s testing was carried out on the data returned. Replicates were taken from the artificial assemblage 50 individuals at a time (constituting one replicate) and these prior to taking the next replicate. The sr::A and P""S curves produced by analysis of these replicate samplings (table 6) are presented in figure 21. The third line in figure 21 is the level of similarity that the model assemblages (average of replicates) had with the known assemblage. In applying the P H S test to various data sets two points became clear. First, the p;.,s test is much more sensitive to distributional parameters than the test and as such, the curve will oscillate when sampling assemblages with internal aggregations. Second, the p:-;s test has a poor sensitivity to the addition of new species after the first few replicates, especially if they are relatively rare (low abundance). As a result the p:-:s test is not recommended as a replacement for testing but as a companion.

PAGE 156

Figure 21. Species:Area curve (triangles), Progressi ve .'"'Similari ty curve (middle curve; circles), and model to actual assemblage similarity curve (lower curve; boxes). r2 = 0.96, 0.80, 0.93 respectively.

PAGE 157

0 ,... en w I-c( () ...I Q. w c: u.. 0 c: lt) w m :::> z c > c z u.. 0 0 c: 0 w N m :::> z

PAGE 158

145 Table 6. Similarity values between successive pairs of model assemblages (M series); values between each model assemblage and the actual assemblage (A); and the accumulation of species/sample (S-A). N s Progressive Similarity M1xM2 M2xM3 M3xM4 M4xM5 M5xM6 M6xM7 M7xM8 M8xM9 M9xM10 0.78 0.90 0.92 0.94 0.95 0.97 0.96 0.97 0.97 Model to Actual Similarities M1xA M2xA M3xA M4xA M5xA M6xA M7xA M8xA M9xA M10xA 0.71 0.80 0.83 0.87 0.89 0.89 0.91 0.92 0.92 0.90 50 17 Species occurrence I Number of Individuals Sampled 100 23 150 23 200 26 250 27 300 27 350 30 400 30 450 31 500 31


Download Options

No images are available for this item.
Cite this item close


Cras ut cursus ante, a fringilla nunc. Mauris lorem nunc, cursus sit amet enim ac, vehicula vestibulum mi. Mauris viverra nisl vel enim faucibus porta. Praesent sit amet ornare diam, non finibus nulla.


Cras efficitur magna et sapien varius, luctus ullamcorper dolor convallis. Orci varius natoque penatibus et magnis dis parturient montes, nascetur ridiculus mus. Fusce sit amet justo ut erat laoreet congue sed a ante.


Phasellus ornare in augue eu imperdiet. Donec malesuada sapien ante, at vehicula orci tempor molestie. Proin vitae urna elit. Pellentesque vitae nisi et diam euismod malesuada aliquet non erat.


Nunc fringilla dolor ut dictum placerat. Proin ac neque rutrum, consectetur ligula id, laoreet ligula. Nulla lorem massa, consectetur vitae consequat in, lobortis at dolor. Nunc sed leo odio.