USF Libraries

Morphological variability in stressed Amphistegina gibbosa (Foraminiferida) in the Florida Keys

MISSING IMAGE

Material Information

Title:
Morphological variability in stressed Amphistegina gibbosa (Foraminiferida) in the Florida Keys
Physical Description:
vii, 60 leaves : ill. ; 29 cm.
Language:
English
Creator:
Harney, Jodi N.
Publisher:
University of South Florida
Place of Publication:
Tampa, Florida
Publication Date:

Subjects

Subjects / Keywords:
Amphistegina gibbosa -- Morphology   ( lcsh )
Dissertation, Academic -- Marine science -- Masters -- USF   ( fts )

Notes

General Note:
Thesis (M.S.)--University of South Florida, 1996. Includes bibliographical references (leaves 50-52).

Record Information

Source Institution:
University of South Florida
Holding Location:
Universtity of South Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 022467124
oclc - 35783574
usfldc doi - F51-00125
usfldc handle - f51.125
System ID:
SFS0044176:00001


This item is only available as the following downloads:


Full Text

PAGE 1

MORPHOLOGICAL VARIABILITY IN STRESSED AMPHISTEGINA GIBBOSA (FORAMINIFERIDA) IN THE FLORIDA KEYS by JODI N HARNEY A thesis submitted in partial fulfillment of the requirements for the degree of Master of Science Department of Marine Science University of South Florida May 1996 Major Professor : Pamela Hallock Muller Ph .D.

PAGE 2

Graduate School University of South Florida Tampa Florida CERTIFICATE OF APPROVAL Master s Thesis This is to certify that the Master's Thesis of JODI N. HARNEY with a major in Marine Science has been approved by the Examining Committee on March 25 1996 as satisfactory for the thesis requirement for the Master of Science degree Examining Committee : Major Pamela Hallock-Muller Ph .D.

PAGE 3

ACKNOWLEDGMENTS Field support Associates SEM and darkroom Graphics Statistics USF Marine Science Department Graduate coordinator Committee members Major professor NOAA-NURC/UNCW Key Largo Subcontract Nos. 9120 9204.4 9322 9515 Geological Society of America Cole Award (to P Hallock-Muller) Dana Williams Helen Talge Rob Walker Sandy Nettles Dawn Olson Strawn Toler Tony Greco Chad Edmisten Dr Robert Muller Dr. Ted Van Vleet Dr Franco S Medioli Dr. Lisa L. Robbins Dr. Pamela Hallock-Muller

PAGE 4

LIST OF TABLES LIST OF FIGURES ABSTRACT INTRODUCTION Purpose TABLE OF CONTENTS Life cycles in foraminifera METHODS Sample collection External morphology Internal morphology Additional samples Limitations of methods RESULTS External morphology Protoconch diameter Additional samples DISCUSSION External variability Sphericity and elongation Variability in the spiral radius Sources of morphologic variability summarized Protoconch size variability Evidence for trimodality in A. gibbosa Advantages of schizogony Evolutionary significance of trimodality Ill iv v 1 4 5 10 10 12 13 16 17 19 19 28 32 35 35 35 38 40 41 42 45 47

PAGE 5

CONCLUSIONS 49 REFERENCES 50 APPENDICES 53 APPENDIX A. COMPLETE RESULTS OF ANALYSES OF VARIANCE PERFORMED ON THE PARAMETER RS/RU 54 APPENDIX B COMPLETE RESULTS OF ANALYSES OF VARIANCE PERFORMED ON PROTOCONCH DIAMETER 56 APPENDIX C COMPLETE RESULTS OF STATISTICAL TESTS PERFORMED ON KNOWN (TL-1991) AND SUSPECTED (CR-TYPE S) SCHIZONTS WITH RESPECT TO RS/RU AND PROTOCONCH DIAMETER 58 APPENDIX D COMPLETE RESULTS OF LINEAR REGRESSIONS PERFORMED ON DMAX DINT AND DMIN FROM M018-584 59 APPENDIX E. COMPLETE RS/DMIN CORRELATION RESULTS FOR EACH SAMPLE 60 II

PAGE 6

LIST OF TABLES Table 1 Summary of sample sites depths and dates of collection 11 Table 2 Descriptive statistics summary of external morphologic parameters for each sample and for pooled Conch Reef samples 21 Table 3 Summary of variance analyses employed to test for significant variability in the Rs/Ru parameter 26 Tab le 4 Descriptive statistics summary of protoconch diameter measurements for each sample and for pooled Conch Reef samples 30 Table 5 Summary of variance analyses employed to test for significant variability in protoconch diameters 30 Table 6 Descriptive statistics summary for known megalospheric specimens (PR-1982) known schizonts (TL-1991 ) and Type "S" field specimens (CRType S) 33 Table 7 Summary of correlation analyses between Dmin (x) and Rs (y) 39 iii

PAGE 7

LIST OF FIGURES Figure 1 Normal dimorphic life cycle of Heterostegina Amphistegina 9 and many other foraminiferal genera F i gure 2 Atypica l trimorphic life cycle of Heterostegina depressa collected from marginal environmental conditions 9 Figure 3 Map of sample collection sites in the Flor i da Keys 13 Figure 4 Diagram of external and internal test measu r ements collected from each specimen 14 Figure 5 Scanning electron micrographs plate 24 Figure 6 Frequency h i stograms of the pa r ameter Rs/Ru 27 Figure 7 Frequency histograms of protoconch d i ameter 31 Figure 8 Frequency histograms for known microspheric specimens (a) known megalosphe r ic specimens (b), known schizonts (c) and suspected schizonts (d) 34 Figure 9 Illustration of linear relationship between Dmax and Dint (a) and between Dmax and Dmin (b) 37 iv

PAGE 8

MORPHOLOGICAL VARIABILITY IN STRESSED AMPHISTEGINA GIBBOSA (FORAMINIFERIDA) IN THE FLORIDA KEYS by JODI N. HARNEY An Abstract Of a thesis submitted in partial fulfillment of the requirements for the degree of Master of Science Department of Marine Science University of South Florida May 1996 Major Professor : Pamela Hallock-Muller, Ph.D v

PAGE 9

Since the summer of 1991, populations of Amphistegina gibbosa on coral reefs in the Florida Keys have exhibited anomalous symptoms of stress including diatom endosymbiont loss, progressive cytoplasmic deterioration impaired calcification, anomalous reproductive events, and reduced population densities Field and laboratory observations prior to 1991 indicated that a metagenic life cycle (alternation of sexual and asexual generations) was typical for Amphistegina spp. In 1991-92 four successive asexual generations were observed within a single cultured lineage, demonstrating departure from the normal dimorphic life cycle and the discovery of a third mode of reproduction in this spec i es (schizogony). Relatively rapid recovery of population densities beginning in late 1992, indicated that Amphistegina gibbosa may have altered its life history strategy from alternation of generations to successive asexual generations to obtain maximum reproductive benefit. In samples collected from Conch Reef in 1994-95 the high variability in shell morphology was attributed to a high frequency of deformed individuals the existence of a schizont-like morphotype and previously undescribed variations in the spiral side of some tests. Protoconch diameters were also highly variable and reflected the presence of schizonts in the natural population. Conspicuously few microspheric specimens were found at 18m sites (<1 %); reproduction at this depth appeared to be predominantly schizogonous At 30m depth, more microspheric specimens were present (-10%), along with megalospheric schizonts indicating that both sexual and schizogonous reproduction were successful at this depth Although biometric analyses of external and internal test vi

PAGE 10

features were unable to verify physical trimorphism in these foraminifera a triphasic life cycle is evident. The presence of probable schizonts in samples collected from Molasses Reef in 1984 indicate that trimodality was probably common in A. gibbosa even before it was recognized in stressed Conch Reef populations Abstract Approved: -"--------'-'---.....:....::'-------------Major Professor : Pamela Hallock-Muller, Ph.D Professor Department of Marine Science Date Approved : vii

PAGE 11

INTRODUCTION Shelled protozoans of the class Foraminiferida with their easily-fossilized agglutinated or calcium carbonate skeletons have been studied by micropa l eontologists for more than two centuries The i r great abundance and ease of collection are features wh i ch facilitate their use in an array of scientific r ealms includ i ng biostrat i graphy evolutionary studies and paleoenvironmental reconstruction efforts Nineteenth-century research on these microfossils initiated the description of hundreds of genera both modern and ancient (e.g. Lamarck 1801; Montfort 1808; D Orbigny 1839) Studies on living foraminifera began with Lister s (1895) work on life cycles and research on the biology of these protists continued into the 20th century (e g., Le Galvez 1935) Cushman (1927) developed the modern taxonomic classification of foraminifera and was the first to investigate their economic use in petroleum exploration In the last 60 years, investigative methods for larger' calcareous foraminifera have been developed and used in stratigraphic correlations evolutionary studies and biological investigations (Drooger, 1993). More recently particular attention has been paid to the benthic foraminifera inhabiting coral reef env i ronments of trop i cal and subtropical lati tudes c i rcling the globe Many extant larger species host photosynthetic algae (e.g., diatoms and 1

PAGE 12

chlorophytes) in a symbiotic relationship that is exceedingly beneficial in the calcification and growth of these marine protists (Lee and Anderson, 1991 ) In their ecological similarities to zooxanthellate corals there lies considerable value for use as sensitive environmental indicators and as proxies for studies of reef health (Hallock 1995). Although Amphistegina is relatively small and morphologically simple compared with most other larger taxa it is one of the most important genera Not only are they the most widespread and abundant extant genus with algal symbionts (Hallock 1988) but they have a fossil record extending back more than 50 million years providing the ancestral stock for at least one lineage of orbitoid larger foraminifera the Discocyclinidae (Giaessner, 1945). Through the i r history they have been prolific calcium carbonate sediment producers, constituting as much as 90% of the sand-sized sediments in some Pacific atolls (McKee eta/. 1959) Benthic habitat depths of these foraminifera range from very shallow waters to over 1OOm, with density distributions on the order of 104 -107 individuals per square meter on suitable hard or phytal substratum (Hallock et a/., 1986a) Members of this genus typically grow to 1-3mm in diameter, harbor diatom endosymbionts, and have a life span of three to six months depending on environmental conditions (Hallock 1981; Hallock eta/., 1986b) Populations of Amphistegina spp possibly worldwide, are currently subject to an unknown stress that was init i ally discovered in the Florida Keys in the summer of 1991 (Hallock et at., 1993) Color loss commonly called bleaching," was found in more than 80% of post-juvenile (>0.6mm in diameter) 2

PAGE 13

Amphistegina gibbosa specimens collected at 1 0-30m depth on Conch Reef and other sites in the Florida Keys (Hallock eta/., 1995). Cytological investigation revealed symbiont digestion empty pore vaults and irreversible cytoplasmic damage (Talge and Hallock 1995). Monthly monitoring of the Conch Reef population since the onset of this disease revealed that the proportion of individuals exhibiting symbiont loss increases each spring peaks near the summer solstice and declines in late summer and fall (Hallock eta/., 1995) Population densities in September 1992 were 5% of those in September 1991, indicating elevated rates of mortality reduced reproductive success or both Partial recovery of densities has occurred as the incidence of bleaching has declined (Hallock eta/., 1995). In addition to symbiont loss and reduced population densities other major symptoms observed include : an apparent increase in morphologic variability in both juvenile and adult size classes ; shell damage (breakage epiphytization parasitization) ; progressive irreversible cytoplasmic damage ; and reproductive anomalies, including successive asexual generations previously unknown in this genus Detailed study of this event is providing invaluable information on how foraminifera respond to stress However symbiont loss itself does not leave a permanent record in the sediments. Understanding morphologic responses will be the key to recognizing similar stress events in the fossil record 3

PAGE 14

Purpose The purpose of this study was to assess morphological responses using biometric analysis of selected internal and external parameters to prolonged stress in Amphistegina gibbosa populations Specific questions that I addressed include : 1 Has variability in external morphology {i e test shape) increased in stressed field populations of A. gibbosa as compared with populations sampled before 1991? 2 Does test shape in stressed populations show temporal or bathymetric trends? 3 Do embryonic chamber dimensions of specimens collected from stressed field populations e x hibit significant anomaly or increased variability when compared to specimens collected before 1991? 4 Does protoconch size-range in individuals from stressed field populations demonstrate a departure from the size-range demonstrated in a normal dimorphic life cycle? The significance of understanding these stress effects is two-fold It provides insight concerning the adaptability of foraminifera to ongoing global environmental change and it will be helpful to investigators studying episodes of environmental change in the fossil record. 4

PAGE 15

Life cycles in foraminifera To investigate the hypothesis that environmental stresses have had an impact on the morphology and reproduct i ve strategy of this foraminifer it is necessary to consider the embryonic elements of the test. Thus it is useful to present a selective review of foraminiferal life cycles as they relate to morphology A more complete review can be found i n Lee eta/. (1991 ) Heterostegina depressa a large orbito i dal, symbiont-bearing benthic foraminifer is similar to Amphistegina gibbosa in i ts circumtropical occurrence habitat, life cycle and general biology Both its normal and environmentally modified life-history strategies have been studied extensively by Rottger et a/. (1986, 1990a) and R6ttger (1990) Rottger s research is invaluable to this study, as it provides a possible explanation of the reproductive anomalies observed in Amphistegina gibbosa specimens collected from stressed field populations from the Florida Keys since 1991. Rottger eta/. (1986) reported that reproduct i on i n Heterostegina and many other foraminiferal genera is metagenic ; that is dimorphic generations of sexually and asexually-reproducing individuals alternate as shown in Figure 1. The agamont (b) is diploid multinucleate and reproduces only asexually Its reproductive nuclei undergo meiosis (unlike other protozoa) and the parental cytoplasm is evacuated from the test ; asexual multiple fission then commences outside the parent's empty shell (R6ttger et a/., 1986) An unknown mechanism efficiently separates the diatom endosymbionts from the parental cytoplasm 5

PAGE 16

during these events, then redistributes them to the daughter cells when fission is complete Each asexual reproduction normally produces several hundred haploid young of uniform size shape and color in both Heterostegina and Amphistegina (Rottger eta/. 1986 ; Hallock 1985). By this mechanism, the uninucleate offspring already harbor the diatom endosymbionts necessary for their survival, growth and calcification When they reach adult size these gamonts (a) undergo mitotic gametogenesis in wh i ch the entire cytoplasm in converted to gametes and shed (Rottger et a/ 1986) Successful gametic union produces a zygote which matures into the diploid agamont and begins another cycle Reproduction whether se x ual or asexual normally occurs in adult size classes Gamonts and agamonts differ in reproductive strategy nuclear state and morphology producing clear gene r ational dimorphism. The pr i mary mechanism to distinguish between these products of sexual and ase x ual cycles lies in differences in their embryonic chamber dimensions (Rottger eta/. 1990a) The first and second chambers of the embryon are called the protoconch and deuteroconch respectively The dimensions of the protoconch are establ i shed during initial calcification and are determined by the volume of protoplasm around which it is built. Thus the size of the embryon is recorded during ontogeny preserved through adulthood and can be extracted from equator i a l sections The agamont is microspheric as its protoconch is noticeably smaller than that of the megalospheric gamont. In a study of Heterostegina depressa (Rottger eta/. 1990a) microspheric spec i mens had protoconch d i ameters on the average 6

PAGE 17

of 281-Jm; protoconchs of megalospheric gamonts averaged 1251-Jm. A clear understanding of this distinction is reached when one considers that daughter cells produced by asexual multiple fission (immature gamonts) begin life much larger than do products of zygosis (primordial agamonts) Through this distinction, a researcher can discern the ontogenetic history of specimens by analyzing their chambers in cross-section Murray (1991) expressed these morphological distinctions by designating Type A individuals as those with large proloculi (first chambers) and smaller tests and Type B individuals as those with small proloculi and larger tests. In addition he states that Type A morphologies (megalospheric gamonts) are far more numerous in natural populations than Type B morphologies (microspheric agamonts) Microspheric specimens are generally rare in field collected samples (Drooger 1993) because asexual reproductions contribute far greater numbers of offspring to the population than do gamete-broadcasting events Figure 2 illustrates the atypical trimorphic life cycle that has been observed in H depressa specimens collected from shallow environments with very low population densities (Rottger eta/., 1986) The broken arrows indicate how the megalospheric, haploid products of asexual reproduction (expected to be gamonts) may deviate from classical dimorphic alternation of generations by undergoing asexual multiple fission instead of gametogenesis once they reach maturity These "schizonts differ from gamonts in more than just their reproductive activity Observations in culture revealed that they reproduce at 7

PAGE 18

smaller s i zes have slower growth rates and possess protoconchs with smaller more variable diameters (R6ttger 1990 ; R6ttger eta/., 1990b) Because of the adaptive complexity and difficulty in studying complete life cycles there was some confusion initially in R6ttger s work. The suppression of the sexual cycle seemed so persistent that he published (R6ttger et a/., 1986) and later retracted (R6ttger et a/., 1990a) the hypothesis that what he was seeing was the emergence of a new apogamic species. This mistake was beneficial however, as it led to the realization that several generations of agametic schizogony may take place before sexual reproduction occurs It is clear that this cycle of multiple schizont formation exists, although the mechanism(s) controlling its initiation duration and closure are still largely unknown It may be an adaptive life history strategy during stressed conditions when population densities are low and the probabilities for gametic union and zygotic success are limited The existence of trimorphic life cycles in foraminifera has been postulated since the early part of this century by researchers who found some specimens didn t fit into classical life cycles (Leutenegger 1977; Drooger, 1993). The occurrence of three biologically-different forms in a single life cycle has been reported in several foraminiferal genera but complete generations suites must be observed from birth through reproduction over several cycles in order to prove trimorphism This has been documented only in Heterostegina depressa (R6ttger eta/., 1990a) 8

PAGE 19

AGAMONT microspheric d i p loi d Figure 1 Normal dimorphic life cycle of Heterostegina Amphistegina and many other foraminiferal genera In metagenesis generations of sexual reproduction (via gamete broadcasting a) and asexual reproduction (via multiple fission b) alternate (Adapted from Rottger et a/ .. 1986, 1990a .) ZVGOTE ;-----, . .. ::_ '-._ haplo i d SCHIZONT mega l ospheri c haploid AGAMONT megalospheri c diploid Figure 2 Trimorphic life cycle of Heterostegina depressa collected from margina l environmental cond i tions (Adapted from Rottger et a/.. 1986 1990a ) 9

PAGE 20

METHODS Sample collection Field populations of Amphistegina gibbosa were sampled in June and September 1994 and in January March and June 1995 from sites at 18m and 30m depth on Conch Reef an extensive spur and groove formation lying 8km offshore of Key Largo in the Florida Keys (Figure 3). Field samples were also collected at 18m depth in May 1995 from nearby Molasses Reef for comparison with archive specimens from the same site collected in May 1984 Table 1 summarizes the sites depths and dates of the field collections Additional samples used in this study are also listed in the table, and the significance of each is discussed separately in a later sect i on Sample collection was carr i ed out following procedures established by Hallock eta/. (1993). For each sample SCUBA divers collected pieces of reef rubble into labe led plastic bags and transported them to the surface Sediment filamentous algae and attached foraminifera were later removed by scrubbing the rubble with a small brush and washing the pieces in seawater at the shore lab in Key Largo 10

PAGE 21

SOURCE/SITE DEPTH DATE R18-694 Conch Reef 18m 6/94 R18-994 Conch Reef 18m 9/94 R18-195 Conch Reef 18m 1/95 R18-395 Conch Reef 18m 3/95 R18-695 Conch Reef 18m 6/95 R30-694 Conch Reef 30m 6/94 R30-994 Conch Reef 30m 9/94 R30-195 Conch Reef 30m 1/95 R30-395 Conch Reef 30m 3/95 R30-695 Conch Reef 30m 6/95 R-lYPE S Conch Reef 30m 1994-95 Mo l asses Reef" 18m 5/84 Molasses Reef 18m 5/95 Puerto Rico* NA 11/82 Laboratory culture NA 199 1 -92 Micros herics Laborato culture NA 1994 Table 1 Summary of sample sites, depths and dates of collection An asterisk (*) indicates specimens were taken from archive samples -M olasses Reef Conch Reef Figure 3 Map of sample collection sites in the Florida Keys 11

PAGE 22

The resultant sediment slurry was transported to our laboratory in St. Petersburg decanted into a 150 x 20mm petri dish and placed in an environmental chamber ( cf. Hallock et a/., 1986b for discussion of light and temperature conditions) Living A. gibbosa specimens move to the top of the sediment when left for several hours and these were picked from sediment samples with the aid of a stereomicroscope From each sample examined a minimum of 30 individuals with test 0 8mm were randomly selected rinsed in deionized water dr i ed marked on their umbilical sides and glued to numbered squares on a micropaleontological slide labeled with the date depth and site of collection These specimens were then used for morphometric analysis. External morphology Each individual was placed under a stereomicroscope fitted with a M i croComp Image Analysis System which projected the specimen's image onto a computer monitor The equipment was calibrated using a Zeiss stage micrometer and repeated measurements had a precision of 0.001 mm A digitizing pad and mouse were used to first trace the outline of each specimen s equatorial view Maximum (Dmax) and intermediate (Dint) diameter lines were then drawn perpendicular to one another through the protoconch region as shown 12

PAGE 23

in Figure 4b The whole specimen v iew of each i nd i vidual was described and any u n usual features were noted ( e.g breakage and general morphotype ) Each spec i men was removed one at a t i me from the m i cropaleontological slide and set on its apertural edge beneath the imaging camera While several techniques were employed to p r event the specimens from falling over glue stick smeared onto a glass slide worked best to secure them The outline of each individual s radial view was traced and the measurements shown i n Figure 4a were collected The endpoint of the umb i lical rad i us line (Ru) was used as the start i ng point for the spiral rad i us l ine (Rs) such that these two measurements could s i mply be added together to obtain the specimen s minimum diameter (Dmin) The umbilical faces of each specimen had been marked for identificat i on and this prevented confusing the two radii. Each spec i men was glued back to its numbered square awaiting sectioning The equatorial and rad ial measurements were entered i n Quattro 4 0 spreadsheets and analyzed graph i cally and stat isti cally The morphometr i c parameters were later subjected to corre lati on and variance ana l yses to determine their relat i ve significance Internal morphology Following external test measurements spec i mens were sectioned and prepared for embryonic chamber analys i s To obtain med i an sect i ons each i nd i v i dual was or i ented umb ili ca l -side up i n a drop of warm Lakes i de 13

PAGE 24

I< )al Ru Rs I< )al Dmin (a) Radial view (c) Internal v i ew AB=Proto c on c h diamet e r CD=Deut erc on c h diamet e r EF=Embry o n he i ght Dint I< )al Dmax (b) Equatorial view (umblical side) B D Figure 4 Diagram of external and internal test measurements collected from each specimen 14

PAGE 25

thermoplastic cement atop a labeled glass slide. After the embedded specimen had cooled the umbilical face of the foraminifer was gently rotated by hand over wet fine-grained (400 grit) sanding paper The section was then rinsed blotted dry, dabbed with blue food coloring and examined under both a binocular scope and a transmitted light microscope to view the embryonic chambers. Several stepwise repetitions of sanding and examination were necessary to obtain satisfactory sections through the protoconch Unusual features were also noted for each specimen during sectioning Specimens successfully sectioned were removed from the thermoplastic by rinsing with methanol. They were then rinsed in deionized water, dried, mounted on an aluminum stub and sputter-coated with gold-palladium. A DS-130 scanning electron microscope (SEM) equipped with a 35mm camera was used to examine and photograph each spec i men at magnifications of approximately 270x and/or 360x Automatic numbering of photographs and the use of a keypad ensured identification of numbered specimens Images on the resultant negative strips were viewed one at a time over a small light source beneath the MicroComp video system previously described such that the photographic image appeared on the computer monitor. The equipment was calibrated using the scale bar visible on each micrograph, and repeated measurements had a precision within 11Jm. The internal biometric measurements (cf Drooger 1993 for discussion) illustrated in Figure 4c were collected using the mouse and digitizing pad from those specimens with clearly visible embryonic 15

PAGE 26

chambers Diameter lines were drawn through the protoconch and across the k i dney shaped deuteroconch ; when the entire embryon was visible a line perpendicular to these diameters was drawn to measure total he i ght. It was often difficult to measure the deuteroconch diameter and embryon height with certainty, thus protoconch diameters were collected for a greater number of ind i viduals Additional samples Three of the samples listed i n Table 1 (denoted PR 1982 TL-1991 and CR-Type S) have not yet been mentioned PR-1982 is a sample of 10 A. gibbosa specimens produced in culture by the multiple fission of a microspheric parent that was collected from Puerto Rico in November 1982 The biometric data on these specimens represents a known megalospheric generation. The TL-1991 sample is composed of 19 A. gibbosa specimens known to be schizonts These indiv i duals were part of a lineage cultured in 1991-92 by Talge (pers comm 1995) in which four successive asexual reproductions occurred The sample denoted CRType S consists of specimens selected from various field collections in 1994-95 based on their schizont-like morphology that was very similar to individuals from Talge s cultured lineage In addition four clone parents from a culture in March 1991 that were found to be microspheric upon sectioning are used as a reference for comparing relative protoconch s i zes 16

PAGE 27

Limitations of Methods A complete description of the morphotypic range of the Western Atlantic Caribbean species Amphistegina gibbosa is needed to investigate the degree of polymorphism present. The objectives of this study did not initially include the investigation of test morphology In retrospect more detailed examination of external parameters by Eigenshape analysis of both equatorial and radial views would have been more useful. Unfortunately, destruction of the shell by sectioning prevents reanalysis of features Problems are inevitably encountered during the delicate process of sectioning and cleaning specimens for examination of internal parameters. Brittle broken and deformed individuals often did not survive the sectioning process or displayed poor internal chamber preservation Occasionally micrographs of sections were masked by debris or were otherwise unusable Typically 50% of the specimens in each sample yielded data on protoconch size For statistical analysis these measurements were pooled by depth (18m and 30m) and by year (all Conch Reef collections from both depths in 1994-95) to provide satisfactory sample sizes Preservation of the entire embryon (nucleoconch) structure was even more difficult to obtain Data collected from the deuteroconch and the nucleoconch were not analyzed because even pooled sample sizes were i nsufficient. 17

PAGE 28

Conventionally protoconch diameters are optically measured us i ng a stereomicroscope (Drooger 1993) as it is a cheaper less time-consuming method than SEM analysis The use of scanning electron microscopy prov i des greater accuracy and more detailed examination but it does so at the expense of the number of specimens that yield internal data 18

PAGE 29

RESULTS External morphology Maximum diameters are consistently used by researchers to indicate size in Amphistegina and other larger foraminifera (e g., Larsen 1976; Hallock 1985; Drooger, 1993) Other measurements were statistically analyzed as the following morphologic ratios: Dmax/Dmin (elongation), Dmax/Dint (sphericity), Rs/Ru, Rs/Dmin and Ru/Dmin. These "secondary data" parameters are frequently used in biometric studies of foraminifera (Drooger 1993) Descriptive statistics of these parameters were compiled and are summarized in Table 2 The samples examined included archived specimens from 18m depth on Key Largo s Molasses Reef in 1984 (M018-584) specimens collected live from the same site in 1995 (M018-595), and five samples each from the 18m and 30m sites on Conch Reef collected quarterly in 1994-95 (e g CR18-694 was collected in June 1994 from 18m on Conch Reef) Data from the five Conch Reef 18m samples were pooled and analyzed (denoted CR-18m), as were those from the 30m site (denoted CR-30m), and from all ten samples collected from Conch Reef (denoted CR-all) 19

PAGE 30

Table 2 (facing page) Descriptive statistics summary of external morphologic parameters for each sample and for pooled Conch Reef samples Shown are : number of observations (n) mean, standard error (SE), minimum value (min) and maximum value (max) Dmax=maximum diameter ; Dmax/Dmin=elongation ; Dmax/Dint=sphericity ; Rs/Ru=spiral radius/umbilical radius ; Rs/Dmin=spiral radius/minimum diameter ; Ru/Dmin=umbilical radius/minimum diameter

PAGE 31

Dmax Dmax/Dm in Sa mole n Mean SE Min Max n Mean SE Min M018-584 48 0 6917 0 0268 0 3677 1 1251 M018-584 48 2 5450 0 0330 2. 0411 3 1093 M018-595 30 0 8984 0 0234 0 5871 1 2155 M018-595 30 2 4778 0 0427 1 9601 2 9703 30 1 1212 0 0391 0 6882 1 3971 CR18-094 30 2 5596 0 0540 2 0340 3 3271 30 1 1363 0 0440 0 7331 1 5830 CR18-994 30 2 6604 0 0551 1 9866 3.4166 30 1 .0899 0 0386 0 7418 1 5986 CR18-195 30 2 3995 0 0505 1 8344 3 1071 37 1 1745 0 0339 0 7384 1 5273 CR18-395 37 2 5926 0 0514 2. 1119 3 3956 33 1 0699 0 0322 0 7025 1 6093 CR18-095 33 2 7073 0 0523 2. 1922 3 5332 30 1.1098 0 0416 0 6941 1 5753 CR30-094 30 2 3208 0 0638 1 .7217 3 2920 CR30-994 30 1 1938 0 0413 0 7543 1 5691 CR30-994 30 2 2994 0 0579 1 8749 3 2548 30 1 0417 0 0248 0 8334 1 3380 CR30-195 30 2.4475 0 0475 1 8942 2 9414 30 1.3203 0 0318 0 9150 1 6349 CR30-395 30 2 5144 0 0734 1 6530 3 3746 "R30-095 30 1 0974 0.0373 0 6793 1 4639 CR30-095 30 2 3643 0 0814 1 5379 3 5742 160 1 1199 0.0168 0 6882 1 6093 160 2 5866 0 0247 1 8344 3 5332 lcR-30m 150 1 1526 0 0177 0 6793 1 6349 150 2 3892 0 0298 1 5379 3 5742 lcR-all 310 1 1357 0 0122 0 6793 1 6349 310 2.4911 0.0200 1 5379 3 5742 Dmax/D int Rs/Ru !sample n Mean SE Min n Mean SE M in M018-584 48 1 1135 0 0080 1 0155 1 2953 M018-584 48 0 7451 0.0194 0.4140 1 1068 M018-595 30 1 0837 0 0087 1 0033 1 1841 M018-595 30 0 .6822 0 0428 0 1042 1 1015 icR18-094 30 1 .0749 0 0141 0 8934 1 2874 30 0 7827 0 0542 0 2056 1 5803 icR18-994 30 1 0967 0 0128 0 9835 1 2134 30 0 7370 0 0502 0 3407 1 3791 CR18-195 30 1 0951 0 0106 1.0000 1 2233 icR18-195 30 0 6841 0 0332 0.2 183 0 9767 37 1 0887 0 0095 0 9534 1 1810 37 0 7642 0 0347 0.1776 1.2729 jcR18-095 33 1 1059 0 0118 1 0080 1 2871 33 0 7913 0 0354 0 3777 1 2619 icR30-094 30 1 1192 0 0103 1 0175 1 2596 30 0 5433 0 0528 0 1705 1 2726 icR30-994 30 1 1021 0.0102 1 0043 1 2122 30 0 3838 0 0192 0 1461 0 .6867 icR30-195 30 1 1137 0 0087 1 0065 1 1876 30 0 6209 0 0420 0 2236 1 0369 icR30-395 3o 1 0850 0 0084 0 9698 1 1681 30 0 6774 0 0660 0 1419 1 8287 lc;R30-$95 30 1 0758 0 0091 1 0000 1 1635 lc;R30-095 30 0 6615 0 0464 0 2519 1 2193 lcR-18m 160 1 0923 0 0052 0 8934 1 2873 160 0 7531 0 0187 0 1776 1 5803 lcR-30m 150 1 0992 0 0043 0 .9698 1 .2596 jcR-30m 150 0 6364 0.0249 0 1419 2 1916 lcR-all 310 1 0956 0 0034 0 8934 1 2873 lrR-all 310 0 6967 0 0157 0 1419 2 1916 Rs/Dmin Ru/Dmin n Mean SE Min Ma) !sample n Mean SE Min Ma M018-584 48 0 4236 0 0065 0 2928 0 5254 M018-584 48 0 5764 0 0065 0.4746 0 7072 M018-595 30 0 3926 0 0178 0 0944 0 5241 M018-595 30 0 6074 0 0178 0.4759 0 9056 lcR18-094 30 0.4232 0 0186 0.1705 0 6124 icR18-094 30 0 5768 0 0186 0 3876 0 8294 CR18-994 30 0.4113 0.0157 0 2541 0 5797 icR18-994 30 0 5887 0 0157 0.4203 0 7459 CR18-195 30 0 3990 0 0128 0.1792 0.4941 icR18-195 30 0 6010 0 0128 0 .5059 0 8208 CR18-395 37 0.4242 0 0129 0 1508 0 5600 icR18-395 37 0 5758 0 0129 0.4400 0 8492 CR18-$95 33 0.4346 0 0114 0 2741 0 .5579 33 0 5654 0 .011 4 0.4421 0 7259 CR30-694 30 0 3312 0 0212 0 1457 0.5600 30 0 6688 0 0212 0.4400 0 8543 CR30-994 30 0 3838 0 0192 0 1461 0 6867 30 0 6162 0 0192 0 3133 0 8539 CR30-195 30 0 3701 0 0174 0 1827 0 5090 30 0 6299 0 0174 0.4910 0 8173 icR30-395 30 0 3781 0 0234 0 1243 0 6465 30 0 6219 0 0234 0 3535 0 8757 lc;R30-095 30 0 3845 0 0170 0 2012 0 5494 lc;R30-095 30 0 6155 0 0170 0.4506 0 7988 lc;R18m 160 0 4190 0 0064 0 1508 0 6124 160 0 5810 0 0064 0 3876 0 8492 lc;R-30m 150 0 3696 0 0089 0 1243 0 6867 150 0 6304 0 0089 0 3133 0 8757 lcR-all 310 0 3951 0 0056 0 1243 0 6867 rR-all 310 0 6049 0 0056 0 3133 0 8757 21

PAGE 32

Normal A. gibbosa specimens generally possess biconvex or planoconvex morphotypes (Hofker 1964). Biconvex tests (Figure 5a) are characterized by nearly equal radii such that the ratio of the spiral side to the umbilical side (Rs/Ru) range from 0 8 to 1 0 Planoconvex tests (Figure 5b) are usually characterized by very low Rs/Ru values (down to 0.1) due to the distended nature of the umbilical side. While performing measurements on the radial views of specimens I noted that specimens in which the spiral radius was greater than the umbilical radius were about twice as common in the Conch Reef samples as in the archive Molasses Reef samples Analysis of variance (ANOVA) techniques were employed to test the significance of variability between several samples with respect to Rs/Ru (cf. Zar, 1984 for discussion) Table 3 summarizes the six analyses performed and lists the experimental F-values critical F-values and significance at the 95% confidence level of each test. Complete ANOVA results are provided in Appendix A. To ensure the validity of sample comparison between Molasses and Conch Reefs ANOVAs 1 and 2 were performed ANOVA 1 demonstrated significant variability in Rs/Ru existed between MO 18-584 and MO 18-595 ANOVA 2 found no significant difference in Rs/Ru values between M018595 and CR 18-694 thus concluding that the Molasses and Conch samples are constituents of the same geographic population and are statistically comparable ANOVA 3 revealed significant variability between the archive sample collected from 18m depth on Molasses Reef in 1984 (M018584) and those 22

PAGE 33

Figure 5 (facing page) Scanning electron micrographs plate Amphistegina gibbosa specimens shown : A: normal biconve x morphotype scale bar=500j..Jm; B : normal planoconvex morphotype scale bar=5001Jm ; C: schizont-like morphotype (Type "S"), scale bar=500j..Jm; D-E : deformed specimens scale bar=5001Jm ; F-L: variability of morphotypes found in 1994-95 samples from Conch Reef, scale bar=5001Jm ; M : equatorial section of test scale bar-1 OOj..Jm; N : megalospheric protoconch scale bar=100j..Jm; 0 : microspheric protoconch scale bar=1001Jm

PAGE 34

24

PAGE 35

collected from 18m depth on Conch Reef in 1994-95 (pooled as CR-18m) Histograms of Rs/Ru frequencies were constructed to illustrate those differences The shape of the histogram for M018-584 (Figure 6a) is similar to that for CR18m (Figure 6c} but highly planoconvex specimens 0 3) and specimens with spiral radii greater than umbilical radii (Rs/Ru > 1.0) were more common in the 1994-95 CR-18m samples Similarly ANOVA 4 tested for significant variability between samples collected in 1994-95 from 18m and 30m depth on Conch Reef Comparison ofF values indicated that the bathymetric variability in Rs/Ru is also significant. Comparison of histograms of Rs/Ru frequencies (Figures 6c and 6d) demonstrated that planoconvex specimens (Rs/Ru < 0 5) were far more common at the deeper site ANOVA 5 was performed to test for significant variability between the five quarterly samples collected from 18m depth on Conch Reef in 1994-95 Comparison of F values revealed that no significant seasonal variability exists with respect to the Rs/Ru parameter. A s i milar conclusion was gathered for the five quarterly Conch Reef 30m samples from the results of ANOVA 6 2 5

PAGE 36

AN OVA Scale of variability Samples included F-value F-crit Signif? 1 Long term M018-584 and M018-595 29 .83 3 .84 yes 2 Geographic M018-595 and CR18-694 2 .11 4 .01 no 3 Long term M018-5 84 and CR-18m 251. 38 3 87 yes 4 Bathymetri c CR-18m and CR-30m 23 53 3 87 yes 5 Seasonal (18m) CR18-$94 CR18-994 CR181 95, 2.34 2 42 no CR18-395 and CR18-$95 6 Seasonal (30m) CR30-$94, CR30-994, CR30-195, 1 05 2.43 no CR30-395, an d CR3 0-$95 Table 3 Summary of variance analyses employed to test for significant variability in the Rs/Ru parameter. Significance level 95% Complete ANOVA results are provided in Appendix A 26

PAGE 37

a b M018-584 d CR-18m 40 "' u ; 20 :I r:r I!! u. 10 0 30 25 C2o "' u i 1 5 :I r:r .t 10 5 0 M018-595 CR-30m Figure 6 Frequency histograms of the parameter Rs/Ru Shown : samples collected from 18m depth on Molasses Reef in 1984 (a) and 1995 (b); samples collected from Conch Reef during 1994-95 at depths of 18m (c) and 30m ( d ) 27

PAGE 38

Protoconch diameter Measurement of protoconch s i ze is the primary method used by researchers to distinguish between products of asexual and sexual reproduction (Rottger eta/., 1986 ; Drooger 1993) Individuals produced by asexual multiple fission are called megalospheric," while those resulting from gametic union are termed microspheric" (Rottger eta/. 1986 ; Lee eta/., 1991) The internal morphology of equatorial sections was examined as illustrated in Figure 4c. Only protoconch diameters are reported here as t he number of useable sections revealing deuteroconch diameter and embryon height were insufficient for analysis Scanning electron micrographs of megalospheric and microspheric sect i ons of A. gibbosa specimens are shown in Figure 5 Plates Nand 0 respectively Descr i ptive statistics compiled for the protoconch diameter data collected from each of the samples are summarized in Table 4 The samples examined and the statistics shown are the same as for the external data, though sample sizes are smaller because only -50% of the sect i ons yielded useable protoconch diameters Analysis of variance techniques were again employed to test the sign i ficance of variability between samples w i th respect to protoconch size. Table 5 summarizes the six analyses performed and lists the experimental F values, critical F-values, and significance of each test at the 95 % level. The ANOVA tests performed on protoconch data (Appendi x B) had the same purpose and simi lar results as those discussed i n the prev i ous section on 28

PAGE 39

external morphologic variability The results of ANOVA 1 established significant protoconch diameter variability between Molasses Reef samples from 1984 and 1995 ANOVA 2 established that Molasses and Conch Reefs are geographic subpopulations that can be reliably compared. ANOVA 3 revealed that significant variability existed between the archive Molasses Reef 18m sample ( 1984) and the pooled Conch Reef 18m samples (1994-95). The frequency histograms of protoconch diameters for these two samples are difficult to interpret. The M018-584 sample (Figure 7a) is vaguely trimodal with nearly 8% of the protoconchs < in diameter. The CR-18m sample (Figure 7c) is unimodal with< 1% of the protoconchs < in diameter and with a greater proportion of protoconchs > in diameter. Strong peaks over the protoconch diameter size-range are present in both samples. ANOVA 4 showed that the depth-related variability in protoconch size that existed between pooled 18m and pooled 30m Conch Reef samples collected during 1994-95 was also significant. The protoconch diameter histogram for CR30m (Figure 7d) approaches a normal distribution The deeper site contains more specimens with very small protoconchs and the peak is stronger in the 30m samples. As with the Rs/Ru parameter no significant seasonal variability (between the five quarterly samples at each depth) could be detected in ANOVA 5 (18m) or ANOVA 6 (30m). 29

PAGE 40

Protoconc h diameter Sample n Mean SE Min Max M018-584 20 36. 9 1 5 15 4 60. 0 M018595 12 39. 1 1 9 26. 1 55 0 16 37 0 1 3 28. 6 55 9 13 33. 5 1 0 27 0 44 4 7 40. 0 2 1 25.4 54 3 13 41. 1 1 8 22 5 60.4 10 45 9 1 5 37 0 57.4 17 37 9 1 6 22 0 52 3 pR30-994 19 34 1 1 3 21. 0 49.7 11 35. 7 1 3 22. 2 46. 8 20 36. 8 1 6 23 3 54 3 17 40. 7 1 3 25. 6 53 6 59 39. 0 1 2 22 5 60.4 84 37 0 0 9 2 1 0 54 3 143 37 9 0 7 2 1 0 60. 4 Table 4 Descriptive statistics summary of protoconch diameter measurements for each sample and for pooled Conch Reef samples Shown are : number of observations (n), mean, standard error (SE), minimum value (min), and maximum value (max) AN OVA Scale of variability Samples i ncluded F-value F-crit Signif! 1 Long term M018584 and M018-595 6 36 4 .10 yes 2 Geographic M018-595 and CR18-694 2 15 4 .17 no 3 Long term M018-584 and CR-18m 251. 38 3 .87 yes 4 Bathymetric CR-18m and CR-30m 23 53 3 87 yes 5 Seasonal ( 18m ) CR18-694, CR18 -99 4 CR18-195 2 34 2.42 no CR18-395, and CR18-695 6 Seasonal (30 m ) CR30-694, CR30-994 CR30-195 1 05 2.43 no CR30 -3 95. and CR30-695 Table 5. Summary of variance analyses employed to test for significant variability in protoconch diameters Significance level 95% Complete ANOVA results are provided i n Appendix B 30

PAGE 41

25 r: .. & 10 :!! ... 5 M018-584 CR-18m {b) ------M018-595 {d),_ ________ _, CR-30m Figure 7 Frequency histograms of protoconch diameter. Shown : samples collected from 18m depth on Molasses Reef in 1984 (a) and 1995 (b); samples collected from Conch Reef during 1994 95 at depths of 18m (c) and 30m (d) 31

PAGE 42

Additional samples Descriptive statistics on the external and internal parameters (Dmax, Dmax/Dmin Dmax/Dint Rs/Ru Rs/Dmin Ru/Dmin, and protoconch diameter) were compiled for samples PR-1982 TL-1991, and CR-Type S ; they are shown in Table 6. To i nvestigate the presence of the schizont morphology" in field populations I compared the Rs/Ru parameter of known schizonts (TL -1991) to that of suspected schizonts (CRType S) At-test of means detected no significant difference in mean Rs/Ru values between the two samples I also performed an AN OVA to compare sample variances of the Rs/Ru parameter and no significant difference was detected between known and suspected sch i zonts. The complete experimental results of these tests are prov i ded in Appendix C Protoconch size-frequency distributions were also constructed for known microspher ics (Figure 8a), for known megalospheric specimens from the PR1982 sample (Figure 8b), for known schizonts of Talge s lineage (Figure 8c) and for suspected schizonts from field collections (Figure 8d) The microspheric and megalospheric histograms illustrate the normal dimorphic variability present in Amphistegina gibbosa protoconch diameters. Known schizonts can possess a wide range of protoconch diamete r s but a conspicuous peak is present over the 35-451-Jm range. The protoconchs of specimens possessing a schizont-like morphotype all fall within the 35-551-Jm range 1 compared the protoconch diameters from TL-1991 and CR-Type S using the same statistical methods : a t-test of means and an analysis of variance No 32

PAGE 43

significant difference in protoconch size could be detected between known and suspected schizonts (Appendix C). These results demonstrate that the field morphotype we believe to be "schizont-like" is not significantly different from that of known schizonts from Talge's lineage (with respect to the external parameter Rs/Ru and protoconch size). Parameter Sample n Mean SE Min Max Dmax PR 1982 10 1 1708 0.0532 0 9361 1 3927 TL-1991 19 0 9906 0 0312 0 8177 1 4840 CR Type S 16 0 8999 0.0405 0 6401 1.1877 Dmax/Dmin PR-1982 10 2.4347 0.0554 2 0923 2 6752 (Elongation) TL-1991 19 2 9330 0 1000 2 2808 3 9619 CR Type S 16 3 1650 0 1045 2 5732 4 0874 Dmax/Dint PR -1982 10 1 1294 0 0173 1 0579 1 2227 (Sphericity) TL-1991 19 1 1444 0 0307 0 9882 1 6100 CR Type S 16 1 0974 0 0141 1 0100 1 1969 Rs/Ru PR-1982 10 0 6079 0 0661 0 3945 0 9547 TL-1991 19 0 9179 0 0668 0.4638 1 7298 CR-Type S 16 0 7709 0 0465 0 3944 1 0717 Rs/Dmin PR-1982 10 0 3689 0 0249 0 2829 0.4884 TL1991 19 0 4645 0 0181 0 3168 0 6337 CR -Type S 16 0 4575 0 0159 0 3364 0 5672 Ru/Dmin PR-1982 10 0 6311 0 0249 0 5116 0 7171 TL1991 19 0 5355 0 0181 0 3663 0 6832 CR Type S 16 0.6180 0 0298 0 4324 0 8528 Protoconch diam. PR-1982 4 77.8 5.4 65. 5 103 0 TL-1991 13 53 8 3 1 37 1 85 2 CR -Type S 8 47 6 2 0 37.4 58 6 Table 6 Descriptive statistics summary for known specimens (PR-1982), known schizonts (TL-1991), and "TypeS" field spec1mens S) Shown are: number of observations (n) mean standard error (SE), m1mmum value (min), and maximum value (max) 33

PAGE 44

a c 60 50 !!40 ,., u :; 30 20 u.. 10 0 25 20 e15 .. 6-10 u.. 5 Known microspherics Known schlzonts [11_) J b d 25 20 c .. 610 u.. 5 Suspected s chlzonts (CRType S) Figure 8 Frequency histograms for known microspher i c spec i mens ( a) known megalospheric specimens (b) known schizonts (c) and suspected schizonts (d). 34 l

PAGE 45

DISCUSSION External variability Shell biometry is a widely-used tool in the study of foraminifera Morphotypes have previously been analyzed with digitizing equipment for use in bathymetric (e.g., Spencer, 1992) stratigraphic (e.g., Saraswati 1995), and evolutionary (e.g., Drooger, 1993) studies Because the developmental stages of foraminiferal life histories are encapsulated in the test biometry of internal chambers is useful in life cyc l e investigations of both modern (e g., Rottger eta/. 1986) and fossil (e.g. Biekart eta/., 1985) foraminifera. Drooger (1993) summarized his lifelong research on radial foraminifera and prov i ded guidelines for biometric studies of external and internal morphology. Sphericity and elongation Of the morphometric ratios (secondary data) the parameters of sphericity ( Dmax/Dint) and elongation (Dmax/Dmin) are of common use in foraminiferal test biometry (Drooger 1993) Both parameters are functions of test size which 35

PAGE 46

reflects age in these foraminifera (Hallock 1985) The relationship between size and sphericity is demonstrated by the linear plot of Dmax vs Dint (Figure 9a) Similarly, the relat i onship between size and elongation is demonstrated by the plot of Dmax vs Dmin (Figure 9b) Both graphs were constructed using data from the Molasses Reef 18m sample from 1984 and are correlated 94% and 98% respectively (see Appendix 0 for complete linear regression results) The sphericity and elongation of Amphistegina gibbosa shells are thus relatively constant functions of test diameter Other researchers (e.g., Larsen 1976; Hallock 1979) have reported that relationships between thickness and diameter for Indo-Pacific species of Amphistegina (and other rotaliine foraminifera) are approximately linear but tend towards a maximum thickness. This trend is especially evident in soritacean genera (e g., Archaias and Sorites) whose thickness to diameter plots are actually curvilinear (Hallock 1979) In A gibbosa spec i mens collected in 1984, the relationship between Dmax and Dmin remained linear even i n very large specimens (Figure 9b). Elongation (Dmax/Dmin) and its inverse (thickness-to-diameter ratio) are frequently used in depth-related morphocline studies (e.g., Spencer, 1992). Orooger ( 1993) Hallock et a/ ( 1986) and Pecheux ( 1995) are among the researchers who have reported that trochospira l and plan i spiral tests of larger benthic foraminifera become flatter and elongation values increase (T/0 values decrease) with increasing depth of habitat. This change is interpreted as a shade adaptation attributed to lower light levels and the corresponding 36

PAGE 47

decreased act i vity of photosynthetic symbionts (Drooger 1993) Hallock eta/. ( 1986b) showed fo r A. lessoni that it is also related to decreased water motion These studies show that biconvex (Ru"'Rs) and planoconvex (Ru>>Rs or Rs>>Ru) forms were generally more abundant in shallower habitats, while compressed forms dominated in deeper waters Contrary to these studies of Amphistegina in the Indo-Pacific the Caribbean species A. gibbosa demonstrates an opposite trend I found that planoconvex specimens consistently occur in greater numbers at the 30m site on Conch Reef and this is supported by what Hallock has observed since the Keys r esearch began in 1991 (pers comm., 1995) Thus the depth related morphoc l ine of A. gibbosa is not influenced by the shade adaptation observed fo r other species a 1.6 .---------, c: e. 1 4 (j; Qi 1 2 E "' i5 1 08 Q) E .2l 06 "" ,. . . ""':. : ..!: 0.4 0.6 0.8 1 1.2 1.4 1.6 1 8 Maximum diameter {Om ax) b 0.8 ,.-----------------, c E 0 7 e. (j; 0 6 Qi 05 i5 E 0.4 :::1 E c 0 3 . . .... ..:. : -. .. 0 2 0.4 0.6 0 8 1 1 2 1.4 1 6 1 .8 Maxim u m diame ter { Om ax ) Figure 9 Illustration of the linear relationship between Dma x and Dint (a) and between Dmax and Dmin (b) Data are from M018-584 37

PAGE 48

In field samples from Conch Reef biconvex tests were typically lenticular in shape when viewed along the test axis, but a continuous gradient existed between several sub-morphotypes (Figure 5, Plates F-L) Some specimens with nearly equal radii (Ru Rs) were lobate, pointed, or curved (Figure 5, Plates F-J), while others were twisted (Plate K) or globular (Plate L). One of the shape variations observed for biconvex tests was that of the schizont-like morphotype (Plate C) The "S" morphotype possesses a very wavy perimeter (axis or "keel") and an extremely thin, discoid test that is often pointed at both ends and abnormally curved Many of the schizonts from Talge's lineage of successive asexual reproductions demonstrated similar features. The irregular presence of Type S individuals ranged from 1 0-50% abundance in the samples with no perceptible trends. Elongation values for these individuals range from 2 6 to 4.1 (Table 6) which are much higher than the normal range of 2.0-2.7 observed for this species (Hallock eta/., 1986b). Known schizonts from Tatge s lineage were also highly elongate, possessing values of 2.3-4.0 (Table 6). These morphologic similarities indicate the previously undescribed "S" morphotype from the natural population possesses features of known schizonts. Variability in the spiral radius Normal A. gibbosa specimens are often unequally biconvex," and increased thickness (Dmin) is usually attributed to greater umbilical radii, as the spiral side often remains flat (Hofker 1964). The descriptive statistics in Table 2 38

PAGE 49

however, show that the field samples from Conch Reef in 1994-95 demonstrate a greater range in Rs/Dmin ratios (0. 1243-0.6867) compared to the archive sample from Molasses in 1984 (0. 2928-0 5254) Correlation coefficients between minimum diameter and spiral radius are listed for each field sample in Table 7. Specimens analyzed from the Molasses Reef 1984 sample exhibited 93% correlation between Rs and Dmin indicating the spiral side of the test was generally a "stable" feature dependent on minimum diameter. In contrast the spiral radius and minimum diameter measurements from 1994-95 field samples demonstrated reduced and highly variable correlations of 13-71% (Appendix E). Some of the variation observed in the elongation and Rs/Ru parameters therefore, must be attributed to the distended spiral radii of many specimens in the Conch Reef samples. This effect is occasionally pronounced to the point of deformity, as in very globular morphologies (Figure 5, Plate L). SAMPLE r M018-584 0 93 M018-595 0 46 CR18-694 0 34 CR18-994 0 .64 CR18-195 0 .71 CR18 395 0 .63 CR18-695 0 77 CR30-694 0.42 CR30-994 0.61 CR30-195 0.35 CR30-395 0.13 CR30-695 0 59 Table 7. Summary of correlation analyses between Dmin (x) and Rs (y) Complete results are provided in Appendix E. 39

PAGE 50

Sources of morphologic variability summarized Populations of some species of foraminifera tend to adapt to the environmental conditions of their local habitat, and distinct morphologies (ecophenotypes) often arise as a result (Drooger, 1993). The previously undescribed "flexibility of the spiral side of the test could be an ecophenotypic adaptation found in stressed populations as could be the novel Type S morphology thought to be indicative of schizonts. Both observations whatever their cause are significant sources of test shape variability in the samples collected from Conch Reef in 1994-95 The high degree of variability in external morphology present within and between 1994-95 Conch Reef samples is also due to the breakage calcification damage, and shell deformities that are among the observed stress symptoms (e. g., Figure 5 Plates D and E). Bathymetric variability is due to these factors plus the observed shift in dominant morphology with depth (i. e., planoconvex forms dominate in deeper habitats) Depth-related morpho l ogic trends in test shape of the Western Atlantic-Caribbean species Amphistegina gibbosa should be stressed as they are vastly different from those documented in Indo-Pacific species and other rotaliine genera Increased levels of UV-b (280-320nm) radiation are suspected of inducing the symbiont loss observed at Conch Reef as well as at other sites in the Flor i da Keys Australia Jamaica Hawaii and Palau (Hallock et. a/. 1995) If high levels of UV are to blame for the bleaching observed in these foraminifera since 1991 40

PAGE 51

perhaps they are causing increased mutation rates in the population If so the range of morphologic variability observed may be a partly genotypic response Protoconch size variability Reproductive anomalies and their ensuing effects have been noted in both field and culture studies since the onset of the stress in 1991 They persist throughout the year at all depths sampled but increase in frequency during the summer months Laboratory observations of reproductive activity show: reproduction in pre-adult size classes (0 5-0 9mm diameter) ; reduced numbers of offspring produced by asexual events ; highly variable sizes and shapes of offspring produced by ase x ual events including high incidence of juvenile deformity ; high juvenile mortality ; and successive asexual reproduction (previously undescribed i n this genus). Perhaps the most significant finding of the reproductive studies carried out since the onset of the degenerative disease involves Talge s lineage of successive asexual reproduction (Talge, pers comm. 1995) Specimens collected in November 1991 were cultured and monitored for reproduct ive activity as usual. When asexual reproductions occurred offspring which were maintained in culture and monitored for growth began to reproduce asexually at unusually small sizes (diameters 0.5-0 8mm) In the following year four 41

PAGE 52

successive asexual generations were observed within this single lineage Schizont production in the Amphistigenidae is undocumented, and th i s was the first observation of such activity in Amphistegina by Hallock and her students in 20 years of research on this genus (Hallock, pers comm., 1995) Megalospheric schizonts have been postulated in other genera of modern larger'' foraminifera (Leutenegger 1977) but until Talge s 1992 A. gibbosa lineage their actual formation by agamonts had been observed only in Heterostegina depressa (Rottger eta/., 1986 ; Rottger eta/., 1990a) Having positively established the existence of a third reproductive mode (the schizont) i n A. gibbosa questions arose as to whether or not this "biological trimodality could be detected in environmentally stressed populations as seen in H. depressa Evidence for trimodality in Amphistegina gibbosa Talge (pers. comm., 1995) observed four successive asexual generations in an A. gibbosa lineage whose original stock was collected from the profoundly impacted Conch Reef population in fall 1991. The external morphology of these known schizonts was similar enough to the field-collected S morphotype that the test measurements of the two samples could not be statistically distinguished Rottger et a/ ( 1986 1 990a) recognized that schizonts of Heterostegina depressa reproduced at smaller sizes and had slower growth rates than megalospheric gamonts Over the last four years of study on stressed A. gibbosa from the Florida Keys Hallock and T alge have observed that most of the indiv i duals that 42

PAGE 53

reproduced or attempted reproduction in culture were small specimens (0 50.9mm) and many morphologically appeared to be Type S individuals (i.e. suspected schizonts). Although slower growth rates were not detected in either schizonts or suspected schizonts recognition of a "schizont" morphology suggests the existence of trimodality in natural populations of A. gibbosa. Beyond the morphologic similarity between known schizonts and field collected Type "S" individuals further evidence that schizonts are present in the natural population can be inferred from the protoconch size distributions given in Figures 7 and 8. From Figure 8a we see that known microspheric A. gibbosa specimens have minute first chambers falling into the size range 15-251-.Jm. Protoconchs of known megalospheric specimens are much larger ranging from 651-.Jm to over 1001-.Jm (Figure 8b) Rottger eta/. (1986 1990a) demonstrated that protoconch diameters of Heterostegina depressa schizonts span a wide range overlapping and falling in between microspherics and normal megalospherics A similar result is illustrated in Figure 8c for Amphistegina gibbosa : the distribution of protoconch sizes observed for the known schizonts occupies a range from 351-.Jm to as large as 801-.Jm. Individuals possessing the Type S morphology have a similar distribution as shown in Figure 8d occupying the size range 35-551-.Jm. The conspicuous peaks that appear in the distributions of known and suspected schizonts at 35-451-.Jm were thus interpreted to be indicative of products of schizogony Surprisingly the telltale schizont peak was not only observed in the protoconch size distributions of the 1994-95 stressed field 43

PAGE 54

samples (i e., M018-595 CR-18m and CR-30m) but was present in the archived sample from 1984 as well (i. e M018-584) This ind i cates that schizogony was occurring commonly long before it was detected The M018-584 histogram indicates that schizonts were accompanied by both m i crospher i c products of sexual reproduction (agamonts) and megalospheric products of normal asexual reproduction (gamonts) in the natural population Thus from protoconch diameter ranges I concluded that the life cycle of this species is typically triphasic (although not obligatory) Variances of protoconch d i ameters compared between depths on Conch Reef were significant although the means for pooled samp l es were very similar (from Table 2) The normality of the 18m curve indicates that schizogony is the dominant mode of reproduction at this depth whereas at 30m both schizogonous and sexual reproduction take place Many researchers have noted that microspheric specimens are conspicuously absent in shallow habitats (e g., Drooger 1993 ; Lee eta/. 1991 ) and even in deeper waters (where se x ual reproduct i on is generally more common) they are uncommon Microspheric specimens have m i nute protoconchs and were easily distinguishable during sect i oning by the very small size of the first second and subsequent chambers Therefore I was able to tabulate the number of microspheric specimens encountered in each sample even if no embryonic data was later collected during SEM analys is. Of the 48 i ndividuals p r esent in the Molasses Reef 18m sample from 1984 nine were lost or ground th r ough before relative protoconch s i ze was noted Of 44

PAGE 55

the remaining 39 individuals three were classified visually as microspheric upon sectioning (-8%). The field sample from Molasses Reef 18m collected in 1995 contained no microspheric specimens in 30 sections Similarly out of the 137 sections from Conch Reef 18m only one microspheric specimen was found If microspheric specimens were occurring in 1984 proportions 13-14 specimens should have been found in modern 18m samples Microspheric individuals were found in "normal proportions in the 30m samples from Conch Reef ; 118 sections yielded 12 microspheric individuals (-10% ) This indicates that if se x ual reproduction is occurring at all at 18m the offspring survived to adulthood less frequently in modern samples than either in 1984 or at present at 30m depth In summary I conclude that the megalospheric schizont morphology is more common in the 1994-95 Conch Reef population than in 1984 particularly at 18m In addition the relative infrequency of microspheric specimens at 18m depth on Conch Reef during the sampling year indicates a shift to predominantly schizogonous reproduction possibly in response to stress The adaptive significance of this stress response i s discussed in the next section. Advantages of schizogony The hypothes i s that A. gibbosa has altered its life history strategy in response to environmental stress is supported intuitively when one considers the advantages of asexual reproduction in stressed conditions. Offspring produced by multiple fission begin life larger and receive their alga l symb i onts d i rectly from 45

PAGE 56

the parent. In contrast products of gametic union must find and incorporate a sufficient number of symbionts Drooger (1993) suggested that high light levels or other environmental stresses may inhibit the photosynthesis and free life" of algal cells, thus making it even more difficult for microspheric juveniles to acquire algal symbionts. Juvenile mortality rates are very high but decrease exponentially with increasing size of young (Hallock 1985) thus megalospheric progeny of asexual reproductions have energetic advantages and far greater chances of survival than do their microspheric counterparts. Reproduction by multiple fission is an effective method for maintaining high population densities. Population densities of A. gibbosa on Conch Reef experienced a relatively rapid recovery beginning in the fall of 1992 indicating a switch to successive asexual generations. Such a rapid expansion of numbers by asexual activity may provide maximum reproductive benefit for stressed populations. This calls for reinterpretation of the premise that sexual reproduction is advantageous under stressed conditions. Although sexual reproduction has the advantage of contributing genetic variability necessary for adaptation (e .g., Cushman, 1948; Sagan and Margulis, 1987) it is energetically expensive and has a high probability of failure in species that are gamete broadcasters. The odds for gametic union and survival to adulthood are limited in stressed environments with low population densities This may explain the reduced frequency of microspheric specimens in 1994-95 samples from 18m 46

PAGE 57

Evolutionary significance of trimodality Because successive asexual generations produce genetically-identical offspring, a mechanism is provided for the propagation of favorable genotypes. In a stressed population such as Conch Reef since 1991, successive asexual generations in the A. gibbosa life cycle enabled the rapid amplification of surviving genotypes This adaptation has implications not only for survival of environmental perturbations, but also for speciation Isolated subpopulations geographically cut off from the parent population could undergo relatively rapid genetic divergence in response to local environmental conditions A type of allopatric speciation could result, for example, during colonization of a newly emergent island or via a newly opened seaway. MacArthur and Wilson ( 1967) argued that successful colonizers are initially "r-selected" as they rapidly increase in population, but then shift to "k selection once the population is established The capacity for successive asexual generations provides Amphistegina and Heterostegina with this potential. Successive asexual generations also provide the potential for amplification of the "founder effect (MacArthur and Wilson 1967). Theoretically one foraminifer could found a new relatively isolated population. Such a population would carry a very limited gene pool with a high potential to be significantly different from the parent population. Further isolation could lead to the emergence of a new species 47

PAGE 58

While the two hypothetical examples just given are examples of possible allopatr i c speciation Rottger eta/. (1986) postulated that a subpopulation of ind i viduals undergoing repeated schizogony could result in sympatric speciation For example in an extreme case if mutagenic stress produced genotypes that were viable but reproductively isolated from parent stock schizogonous reproduction might allow that new genotype to establish a new species or (if the genotype was sufficiently different from the parent stock) even higher taxa All of these mechanisms may partly account for the rapid diversification of larger foraminifera at certain times in the geologic r ecord 48

PAGE 59

CONCLUSIONS 1. In Amphistegina gibbosa variability in test shape was greater in spec i mens collected in 1994-95 during a prolonged bleaching event than in specimens collected in 1984 2 Sources of increased variability in test morphology included high frequencies of deformed individuals prevalence of a schizont-like morphotype and a significant increase in variability of the spiral radius. 3 Variability in protoconch diameter was significantly different in specimens collected in 1994-95 than in those collected in 1984. 4 Significant variability was shown to exist between protoconch size distributions of field specimens collected from 18m depth and those collected from 30m depth 5 Conspicuously fewer microspheric (sexually-produced) specimens were found in 18m samples from 1994-95 than were found in 1984 Reproduction at this depth in 1994-95 appeared to be predominantly schizogonous 6 The presence of probable schizonts in the 1984 population (shown by protoconch size distributions) indicates that schizogony was common in Amphistegina gibbosa long before it was recognized by this study 49

PAGE 60

REFERENCES Biekart J.W. T. Bor R. Rottger, C W Drooger, and J .E. Meulenkamp 1985 Megalospheric Heterostegina depressa from Hawaii in sediments and laboratory cultures Proc Koninklijke Nede r landse Akademie van Wetenschappen, Ser B 88:1-20 Cushman J .A. 1928 Foraminifera : Their Classification and Economic Use Special Publications Cushman Laboratory for Foraminiferal Research. 1 : 1-401 Cushman J A. 1948 Foraminifera : Their Classification and Economic Use 4th Edn, Harvard University Press, Cambridge Mass D Orbigny Alcide Dessalines 1839 Voyage dans /'Amerique Meriodionale Foraminiferes: v.5. Pitois-Levrault et ce, Paris 86pp Drooger, C W. 1993 Radial Foraminifera: Morphometries and Evolution North Holland Press, New York 242pp Glaessner, M.F 1945 Principles of Micropaleontology. Melbourne University Press Melbourne. Hallock P 1979. Trends in test shape with depth in large, symbiont-bearing foraminifera J. Foram Res 9 : 61-69 Hallock P 1981. Light dependence in Amphistegina J Foram Res 11:42-48 Hallock, p 1985 Why are larger foraminifera large? Paleobiology 11: 195-208. Hallock, P 1988. Interoceanic differences in foraminifera with symb i otic algae : a result of nutrient supplies? Proc 6th lnternat. Coral Reef Symp 3 : 251-255 50

PAGE 61

Hallock P 1995. Amphistegina (Foraminiferida) densities as a practical reliable, low-cost i ndicator of coral reef vitality Submitted to : A coral reef symposium on practical, reliable low cost monitoring methods for assessing the biota and habitat conditions of coral reefs Not paginated Hallock P., T L Cottey, LB. Forward and J Halas. 1986a Population biology and sediment production of Archaias angulatus (Foraminiferida) in Largo Sound Florida. J Foram Res 16 : 1-8 Hallock P., LB. Forward and H J Hansen 1986b Influence of environment on the test shape of Amphistegina J Foram Res. 16:224-231 Hallock P., H.K Talge E.M Cockey and R.G Muller 1995 A new disease in a reef-dwelling foraminifer : Implications for coastal sedimentation J Foram Res 25 : 280-286 Hallock P., H .K. Talge K. Smith and E.M Cockey 1993 Bleaching in a reef dwelling foraminifer, Amphistegina gibbosa Proceedings 7th International Coral Reef Symposium Guam 1992 1 : 42 47 Hofker J 1964 Foraminifera from the tidal zone in the Netherlands Ant i lles and othe r West Indian islands. Studies on the Fauna of Curac;ao and other Caribbean Islands 21: 1-119 Lamarck J .B. 1801. Systeme des animau x sans vertebres The Author Paris 432pp. Larsen, A.R. 1976 Studies of recent Amphistegina taxonomy and some ecological aspects Israel J Earth Sci. 25 : 225-239 Le Galvez Jean 1935. Les gametes de quelques Foraminifereres Acad Sci. Paris Comptes Rend us v 210 p 1505-1507 Lee J J and O.R. Anderson 1991. Biology of Foraminifera Academic Press New York 368pp Lee J J W W Faber Jr., O .R. Anderson, and J Pawlowski. 1991. Life cycles of foraminifera In Lee J .J. and O .R. Anderson, eds., Biology of Foraminifera. Academic Press New York p 285-334 Leutenegger S. 1977 Reproduction cycles of larger foraminifera and depth distribution of generations Utrect Micropaleontological Bulletin 15 : 26 34 Lister J J 1895 Contributions to the life history of the Foraminifera Royal Soc London Philos Trans., ser B v. 186, p 401-453 51

PAGE 62

MacArthur Robert H and Edward 0 Wilson 1 967 The Theory of Island B i ogeography. Princeton Univers i ty Press New Jersey 203pp McKee E.D., J Chronic and E.B Leopold 1959 Sed i mentary belts in the lagoon o f Kapingimarangi Atoll. Bull. Am Assoc Petrol. Geol. 43 : 501-562 Montfort Denys de. 1808 Conchyliologie systematique et classificat i on methodique des coquilles : v 1, lxxxv i i 409pp Mur r ay J.W. 1991. Ecology and distr i but i on of benthic foramin i fera In Lee J.J and O.R Anderson eds. Biology of Foraminifera Academic Press New York p 221-253 Pecheux, M J .-F. 1995 Ecomorphology of a recent large foraminifer Operculina ammonoides. GEOBIOS 28 : 529-566 Rottger, R 1990. Biology of larger foraminifera : Present status of the hypothesis of trimorphism and ontogeny of the gamont of Heterostegina depressa In BENTHOS '90. Tokai Unive r sity Press Sendai p 43-54 R6ttger R M Fladung R Schmal j ohann M Spindler and H Zacharias 1986 A new hypothesis: The so-called megalospheric schizont of the larger foraminifer Heterostegina depressa D'Orbigny 1826 is a separate species J. Foram. Res 16: 141-149. Rottger R R Kruger and S de Rijk. 1990a Trimorph i sm i n foraminifera (protozoa): Verification of an old hypothesis Europ J Protistol. 25 : 226-228 Sagan D and L Margul i s 1987 Cannibal s relief : the origins of sex. New Scientist. 6 Aug : 36 40 Saraswati P K 1995. Biometry of early Oligocene Lepidocyclina from Kutch India. Ma r Micropaleontol. 26 : 303 3 1 1 Spencer R 1992. Quantified intraspec i fic variat i on of common benthic Foraminifera from the northwest Gulf of Me x ico : a potential pa l eobathymetric indicator J Foram Res. 22 : 274-292 Talge H K and P Hallock 1995 Cytological examination of symbiont loss in a benthic foraminifer Amphistegina gibbosa Mar Micropaleontol. 26 : 1 07113 Zar, J .H. 1984. Biostatistica/ analysis Prent i ce-Hall New Jersey 718pp 52

PAGE 63

APPENDICES 53

PAGE 64

01 Rs / Ru: ANOVA 1 (LONG TERM VARIABILITY) Analysis ofVaria n ce : One Wa y S u mmary Groups M 01 8 -58 4 CR18m An a l y sis o f Varia n ce Sourc e of Vari a tion Belween Groups Groups Total Count S um Average Vari a n ce 48 3 5 .765 9 9 0.74512 5 0 018092 1 60 12 0.499 3 0 .753121 0 .0 5 5836 ss d f 22 43669 28 38333 50 82002 M S F P -value F-crlt 1 22 43669 261. 3 754 4 02E-42 3 870867 3 1 8 0 089256 319 R s /Ru: AN OVA 3 (SEASONAL VARIABILITY18m) Ana lysis o f V ar i ance : One W ay Summary G r oups C R 1 8 -694 CR1 8-994 CR1 8-195 CR18-395 CR18-695 Analysis o f Variance Source of Variation Belween Groups Groups T otal C o unt S um A vera g e Varia n ce 30 23.4 7999 0 7 82666 0 088275 30 22 1087 0 .736957 0.0 75741 30 20. 52246 0 68408 2 0.033044 37 28. 27628 0.764224 0 04465 1 33 26.1119 1 0 .79127 0 04 1442 s s df 1 04426 20 09725 21.14151 M S F P va l u e F -crlt 4 0 261065 2 338214 0 057028 2 .42184 3 180 0 .111651 184 Rs /Ru: ANOVA 2 ( BATHYMETRIC VARIABILITY) Analysi s of Vari ance:On e Wa y Summary Groups CR1 8 m CR30 m A n a l ysis of V a r i anc e Source of V a riation Belween Groups Within Groups Total Counr Sum A ve r age V a r ia n c e 160 120. 49 9 3 0 753 1 2 1 0.0 55836 1 50 95.4 5 4 1 5 0.63636 1 0.0 92763 ss df 1 960 1 93 26.49612 28.45631 M S F P-valu e F-crlt 1 1 96019 3 23 6 2677 1 93E-06 3 870867 318 0 .083321 3 1 9 Rs/Ru : ANOVA 4 ( SEASONAL VARIAB I LITY 3 0m) A n alysis ofVariance: O n e W a y Summ a ry Grouos CR30-69 4 C R 30-99 4 C R30-19 5 CR30-395 CR30-695 Analysis of Varia n ce Source of Varia t ion B e lween Groups Groups To t al Count Sum A v e r a o e Var ia n ce 30 1 6 29855 0 543285 0 083576 30 20 3635 1 0 6787 84 0 1 3 1287 30 1 8 6269 0.6 20897 0 05300 1 30 20 32082 0 67736 1 0.1 3 068 2 30 1 9 84438 0 661479 0 064604 ss df 0 390416 13. 43132 1 3 .8217 4 MS F P va/u e F-crlt 4 0 09 7 604 1 .063699 0 .381781 2 434066 145 0 09263 149 '"'C)> m-u ::U'"'C .,m oz ;:uo m)> 0 0() zo I'"'C mr ;:um en-! _m ::U;:u em -uen )>C mO -111 m)> ::Uz )> en m en 0 ;:u )> z () m

PAGE 65

APPENDIX A. (cont.) COMPLETE RESULTS OF ANALYSES OF VARIANCE PERFORMED ON THE RS/RU PARAMETER Rs/Ru: ANOVA 5 Analysis of Variance:One Way Summary Groups M018-584 M018-595 Analysis of Variance Source of Variat i on Between Groups Within Groups Total Count Sum Average Variance 48 35.76599 0 745125 0 018092 30 20.46736 0 682245 0 055044 ss df 2.438 7 682998 10 .121 MS F P-va/ue F-crit 1 2.438 29.82846 3 8E-07 3.942303 94 0 .0817 34 95 Rs/Ru: ANOVA 6 Analysis of Variance:One Way Summary Groups M018595 CR18-694 Analysis of Variance Source of Variation Between Groups Within Groups Total Count Sum Average Variance 30 20.46736 0 682245 0.055044 30 23.47999 0 782666 0.088275 ss df 0 151265 4 156246 4 307511 MS F P-va/ue F-crit 1 0 151265 2.110891 0 151643 4.006873 58 0 071659 59 55

PAGE 66

(J"l 0> PROTOCONCH : ANOVA 1 (LONG TERM VARIABILITY) Analysis of Variance: One Way Summary Groups CR18m M018-584 Analysis of V ari ance Source of Varia t i on Between G r oups Within Groups T otal Count Sum Average Variance 60 2300 27 38 33783 111. 1688 21 737. 03 35.09667 168. 9993 ss df 20364 33 26752 69 4711 7 02 MS F P-vaiue F-crit 1 20364 .33 89. 8224 3.47 E -16 3 921478 118 226.7177 119 PROTOCON C H : ANOVA 3 (SEASONAL VARIABILITY18m) Analysis of V a r la n ce:""bne Way Sum mary Groups CR18-694 CR18-994 CR181 95 CR18-395 CR18-695 Analysis of Variance Source of Variation Between Groups Within Groups Total Count Sum Averaae Variance 16 592.71 37 04438 5 1 52875 13 435 26 33 48154 32 62113 7 279. 84 39 97714 126 4969 13 533. 6 7 41.05 1 54 123. 966 1 10 458. 79 45 .879 73. 51303 ss df 1753 449 59922 .62 61676 06 MS F P-vaiue F-crit 4 438 3622 1 1 33898 0 342645 2 430002 155 386 5975 15 9 PROT OCO NCH : ANOVA 2 (BATHYMETRIC VARIABILITY) Analysis ofVariance:One Way Summary Groups M01 8-584 CR-18m Analysis o f V a r i ance Source of Vari at ion Be twee n G r oups Within Group s Total Count Sum Average Variance 85 3111. 6 36 60 706 78 70043 60 2300.27 38 33783 111. 1688 ss df 3872 096 39107 26 4 2979 35 MS F P-vaiue F-crit 1 3872 096 16.63405 6 99 E -05 3 897407 168 232 7813 169 PROTOCON C H : ANOVA 4 (SEASONAL VARIABILITY30m) Analysis of Vari ance: One Way S ummary Groups CR30-994 CR3 0 1 95 CR30-395 C R30-695 A n alysis o f Va ria nc e S ource of Va ria tion Betwee n Groups Groups Tota l Count Sum Averaae Vari ance 17 644 5 37. 91176 76 6711 1 9 647.6 34 .08421 49 36918 11 392.4 35. 67273 47 03418 20 735.3 36.765 76 08134 17 691 8 40 69412 49 7 2184 ss df 2386 206 53584 1 2 55970 33 MS F P-vaiue Fcr/t 4 596 5516 1 614284 0 173778 2.434065 145 369 5457 149 ""O)> m-o :::0""0 "Tim oz :::co s: mx o9J Oo zo -os: :::0""0 or -lm o-l om o:::c zm (")(/) Ic r 0--l )>(/) s:o m"TI -l)> mz :::0)> r -< (/) m (/) 0 "TI :::0 )> z (") m

PAGE 67

APPENDIX B (cont.) COMPLETE RESULTS OF ANALYSES OF VARIANCE PERFORMED ON PROTOCONCH DIAMETER PROTOCONCH: ANOVA 5 Analysis of Variance: One Way Summary Groups M018-584 M018-595 Count Sum Average Variance Analysis of Variance Source of Variation Between Groups Within Groups Total 20 737 03 36 8515 109 8222 12 469 7 39 14167 112 .0117 ss df 1786 633 10672 69 12459 32 MS F P-value F-crit 1 1786 633 6.361291 0 .015973 4.098172 38 280 .8602 39 PROTOCONCH: ANOVA 6 Analysis of Variance:One Way Summary Groups M018-595 CR18-694 Analysis of Variance Source of Variation Count 12 16 Sum Average Variance 469 7 39 14167 112 .0117 592 .71 37.04438 51. 52875 ss df MS F P-value F-crit Between Groups Within Groups Total 472 .8581 6601.271 7074 129 1 472 8581 2 148941 0 153072 4.170877 30 220 0424 31 57

PAGE 68

01 CX> Rs/Ru: t-test !-Test: Two-S ample Mean Variance Observations Pea r son Correlation Poo l ed Variance df t P(T <=t) one-tail t Cri tical one-tail P(T <=t ) two-tail t Critical two-tail Rs/R u: ANOVA C R-Tvoe S TL-1991 0 77093481 0 917867 0.034 60867 0 10692 16 19 N A 0 1 5206572 32 0 40962183 0 34240628 1 69388874 0 68481257 2 03693334 Analysis o f Va riance:One W ay Summary GroiJ p_s CRType S TL-1991 Analys i s of Variance Source of Variation Between Groups W i thin Groups Total Coun t S u m A vera g e Varia n c e 16 12 33496 0 770935 0 034609 19 17.43947 0 917867 0 10692 ss df 0 .5 4283506 8 94831334 9 4911484 MS F Pv a/u e Fcrlt 1 0 5428 3 5 2 7 9 0516 0 1 0 1615 4 051749 46 0 .194 52 9 47 PROTOCONCH: t-test !-Te s t : T w o-Sample Mean Variance O b servations Pearson Corr ela t i o n Pooled V a riance d f P(T<=t) o n et ail t Critical one-tail P(T <=t) two-tail t Critical two-t ail CR Tvo e S TL 1991 47. 64 53. 78077 63. 2376 230.9521 8 1 3 NA 7 8 1.644851 33 0 33871084 1 6923603 0 67742168 2 0345153 PROTOCONCH : ANOVA A nalysis o f V a r i ance:One Way Summary G r o ups CR-TYPE S TL1 991 Analysis of V a riance Source o f V a riation Between Group s Within Groups T ota l Count S u m V a ria nce 8 381 12 4 7 .64 63. 2376 13 699.15 53. 78077 230 9521 ss df 2107 1475 2 32552 1709 34659 .3 184 MS F P-value F -crlt 1 2107 148 2.9 77 644 0 091135 4 051749 46 707. 6559 47 :::OO)> mz-u (1):;'\'"U -uzm m0z ()< --IZ() o....-... :::orJo ;Qcos: )>.._... r z)>m Oz--1 om '"U(/);;o :::ocm 0(/)(1) --1-uC Om ()()-I 0--i(/) ZmO Oo"'TI I..--.. (I) oO--i -:::0)> )>1--i s::--1m-<(/) --i'"U--1 mmo ;;o(l))> ......... (/)-I om I (I) t5(/) Z'"U --lm (1):::0 -:::0 :is: m 0

PAGE 69

APPENDIX D COMPLETE RESULTS OF LINEAR REGRESSIONS PERFORMED ON DMAX DINT AND DMIN FROM M018-584 M018-584: Correlation results x =Dma x y-Dint Regression Output: Constant Std Err of Y Est R Squared No. of Observations Degrees of Freedom X Coefficient(s) Std Err of Coef r x-Dma x y=Dmin 0 905233 0 023259 Regression Output: Constant Std Err of Y Est R Squared No. of Observations Degrees of Freedom X Coefficient( s) Std Err o f Coef r 0 392945 0 020517 59 0 00437 0 04142 0 970527 48 46 0 .98 5153 0 002285 0 0365 38 0 888563 48 46 0 942636

PAGE 70

APPENDIX E. COMPLETE RS/DMIN CORRELATION RESULTS FOR EACH SAMPLE Rs I Dmin: Correlation results M01 8-584 M018-595 Regression Output: Regression Output: Regression Output: Constant -0 .001 4586 Constant 0.026 1 2958 Constant 0 09359591 Std Err of Y Est 0 01903241 S td Err of Y Est 0.05123371 S t d Err of Y Es t 0 04617296 R Squared 0 65666056 R Squared 0 216046 R Squared 0 11640292 No. of Observations 46 No of Observa t ions 30 No of Observations 30 Degrees of Freedom 46 Degr ees of Freed om 26 Degrees of Freedom 26 X Coefficie nt(s) 0.42509236 X Coefficient(s) 0.33597926 X Coefftcient(s) 0.20437571 Std Err of Coef. 0.02563792 Std Err of Coef. 0 12094993 Std Err of Coef. 0 10641319 r 0 .925 5596 r 0 46460749 r 0 34117672 C R18-S 94 C R18-195 CR1 8-3 95 Regression Output: Regression Output: Regression Output: Constant 0 07417 Co n stant 0 01264764 Co nstant 0 04635672 Std Err of Y Est 0 03008879 Std Err of Y Est 0.03531941 Std Err of Y Est 0 03995476 R Squared 0 40455462 R Squared 0 50962062 R Squared 0 39670462 No of Observations 30 No of Observations 30 No of Observations 37 Degree s of Freedom 26 Degrees of Freedom 26 Deg r ees of Freedom 35 X Coefficie nt (s) 0.22942656 X Coefftcient(s) 0 36955297 X Coefficient(s) 0 3 143 7012 S td Err of Coe f 0.05 260 163 S td E rr of Coef 0.06650764 Std Err of Coef. 0.06552976 r 0 63604606 r o 7 136n16 r 0 62984492 CR30-994 R egre ssion O utput: R egression Output: Regression Output: Constant 0 04243 3 73 Cons tant 0.06636739 Constan t 0 02415542 Std E r r of Y Est 0 02646817 Std Err of Y Est 0.0570902 1 Std Err of Y Est 0 056 1 6921 R Squared 0.5904295 R Squared 0 17310053 R Squa red 0 36606892 N o o f Observations 33 N o of O b servat ion s 30 No o f Observations 30 Degrees of Freedom 31 Degrees of Freedom 26 Degrees of F reedom 26 X Coefficient(s) 0.32355 1 0 1 X Coefficient(s) 0 16466799 X Coefficient(s) 0 33443409 Std Err of Coer 0.04639965 St d Err of Coef. 0 0 763672 Std Err or coer 0 06317064 r 0 7663941 r 0.41605352 r 0.6050363 C R 3 0 -195 C R30 -395 Regression Output : Regression Output : Regression Output : Co n stant 0 .047 65116 Constant 0 23493073 Constant 0 04476756 S td Err o f Y Est 0.04214197 Std Err o f Y Est 0 05957969 Std Err of Y Est 0 04566 194 R Squared 0 11431356 R Sq ua red 0.0 1 576 553 R Squared 0 34762611 No of Observations 30 No of Observations 30 No of Observa tion s 30 Degrees of Freedom 26 Degrees of Freedom 26 Degrees of Freedom 26 X Coefficient(s) 0 .2567663 X Coefficien t (s) -0 0730665 X Coefficient(s) o .265no16 Std Err of Coef. o .1350m6 Std E rr of Coef. 0 1 0913234 Std Err of Coef 0.07396267 r 0 3361 0292 r 0.12556065 r 0 56959626 60



PAGE 1

MORPHOLOGICAL VARIABILITY IN STRESSED AMPHISTEGINA GIBBOSA (FORAMINIFERIDA) IN THE FLORIDA KEYS by JODI N HARNEY A thesis submitted in partial fulfillment of the requirements for the degree of Master of Science Department of Marine Science University of South Florida May 1996 Major Professor : Pamela Hallock Muller Ph .D.

PAGE 2

Graduate School University of South Florida Tampa Florida CERTIFICATE OF APPROVAL Master s Thesis This is to certify that the Master's Thesis of JODI N. HARNEY with a major in Marine Science has been approved by the Examining Committee on March 25 1996 as satisfactory for the thesis requirement for the Master of Science degree Examining Committee : Major Pamela Hallock-Muller Ph .D.

PAGE 3

ACKNOWLEDGMENTS Field support Associates SEM and darkroom Graphics Statistics USF Marine Science Department Graduate coordinator Committee members Major professor NOAA-NURC/UNCW Key Largo Subcontract Nos. 9120 9204.4 9322 9515 Geological Society of America Cole Award (to P Hallock-Muller) Dana Williams Helen Talge Rob Walker Sandy Nettles Dawn Olson Strawn Toler Tony Greco Chad Edmisten Dr Robert Muller Dr. Ted Van Vleet Dr Franco S Medioli Dr. Lisa L. Robbins Dr. Pamela Hallock-Muller

PAGE 4

LIST OF TABLES LIST OF FIGURES ABSTRACT INTRODUCTION Purpose TABLE OF CONTENTS Life cycles in foraminifera METHODS Sample collection External morphology Internal morphology Additional samples Limitations of methods RESULTS External morphology Protoconch diameter Additional samples DISCUSSION External variability Sphericity and elongation Variability in the spiral radius Sources of morphologic variability summarized Protoconch size variability Evidence for trimodality in A. gibbosa Advantages of schizogony Evolutionary significance of trimodality Ill iv v 1 4 5 10 10 12 13 16 17 19 19 28 32 35 35 35 38 40 41 42 45 47

PAGE 5

CONCLUSIONS 49 REFERENCES 50 APPENDICES 53 APPENDIX A. COMPLETE RESULTS OF ANALYSES OF VARIANCE PERFORMED ON THE PARAMETER RS/RU 54 APPENDIX B COMPLETE RESULTS OF ANALYSES OF VARIANCE PERFORMED ON PROTOCONCH DIAMETER 56 APPENDIX C COMPLETE RESULTS OF STATISTICAL TESTS PERFORMED ON KNOWN (TL-1991) AND SUSPECTED (CR-TYPE S) SCHIZONTS WITH RESPECT TO RS/RU AND PROTOCONCH DIAMETER 58 APPENDIX D COMPLETE RESULTS OF LINEAR REGRESSIONS PERFORMED ON DMAX DINT AND DMIN FROM M018-584 59 APPENDIX E. COMPLETE RS/DMIN CORRELATION RESULTS FOR EACH SAMPLE 60 II

PAGE 6

LIST OF TABLES Table 1 Summary of sample sites depths and dates of collection 11 Table 2 Descriptive statistics summary of external morphologic parameters for each sample and for pooled Conch Reef samples 21 Table 3 Summary of variance analyses employed to test for significant variability in the Rs/Ru parameter 26 Tab le 4 Descriptive statistics summary of protoconch diameter measurements for each sample and for pooled Conch Reef samples 30 Table 5 Summary of variance analyses employed to test for significant variability in protoconch diameters 30 Table 6 Descriptive statistics summary for known megalospheric specimens (PR-1982) known schizonts (TL-1991 ) and Type "S" field specimens (CRType S) 33 Table 7 Summary of correlation analyses between Dmin (x) and Rs (y) 39 iii

PAGE 7

LIST OF FIGURES Figure 1 Normal dimorphic life cycle of Heterostegina Amphistegina 9 and many other foraminiferal genera F i gure 2 Atypica l trimorphic life cycle of Heterostegina depressa collected from marginal environmental conditions 9 Figure 3 Map of sample collection sites in the Flor i da Keys 13 Figure 4 Diagram of external and internal test measu r ements collected from each specimen 14 Figure 5 Scanning electron micrographs plate 24 Figure 6 Frequency h i stograms of the pa r ameter Rs/Ru 27 Figure 7 Frequency histograms of protoconch d i ameter 31 Figure 8 Frequency histograms for known microspheric specimens (a) known megalosphe r ic specimens (b), known schizonts (c) and suspected schizonts (d) 34 Figure 9 Illustration of linear relationship between Dmax and Dint (a) and between Dmax and Dmin (b) 37 iv

PAGE 8

MORPHOLOGICAL VARIABILITY IN STRESSED AMPHISTEGINA GIBBOSA (FORAMINIFERIDA) IN THE FLORIDA KEYS by JODI N. HARNEY An Abstract Of a thesis submitted in partial fulfillment of the requirements for the degree of Master of Science Department of Marine Science University of South Florida May 1996 Major Professor : Pamela Hallock-Muller, Ph.D v

PAGE 9

Since the summer of 1991, populations of Amphistegina gibbosa on coral reefs in the Florida Keys have exhibited anomalous symptoms of stress including diatom endosymbiont loss, progressive cytoplasmic deterioration impaired calcification, anomalous reproductive events, and reduced population densities Field and laboratory observations prior to 1991 indicated that a metagenic life cycle (alternation of sexual and asexual generations) was typical for Amphistegina spp. In 1991-92 four successive asexual generations were observed within a single cultured lineage, demonstrating departure from the normal dimorphic life cycle and the discovery of a third mode of reproduction in this spec i es (schizogony). Relatively rapid recovery of population densities beginning in late 1992, indicated that Amphistegina gibbosa may have altered its life history strategy from alternation of generations to successive asexual generations to obtain maximum reproductive benefit. In samples collected from Conch Reef in 1994-95 the high variability in shell morphology was attributed to a high frequency of deformed individuals the existence of a schizont-like morphotype and previously undescribed variations in the spiral side of some tests. Protoconch diameters were also highly variable and reflected the presence of schizonts in the natural population. Conspicuously few microspheric specimens were found at 18m sites (<1 %); reproduction at this depth appeared to be predominantly schizogonous At 30m depth, more microspheric specimens were present (-10%), along with megalospheric schizonts indicating that both sexual and schizogonous reproduction were successful at this depth Although biometric analyses of external and internal test vi

PAGE 10

features were unable to verify physical trimorphism in these foraminifera a triphasic life cycle is evident. The presence of probable schizonts in samples collected from Molasses Reef in 1984 indicate that trimodality was probably common in A. gibbosa even before it was recognized in stressed Conch Reef populations Abstract Approved: -"--------'-'---.....:....::'-------------Major Professor : Pamela Hallock-Muller, Ph.D Professor Department of Marine Science Date Approved : vii

PAGE 11

INTRODUCTION Shelled protozoans of the class Foraminiferida with their easily-fossilized agglutinated or calcium carbonate skeletons have been studied by micropa l eontologists for more than two centuries The i r great abundance and ease of collection are features wh i ch facilitate their use in an array of scientific r ealms includ i ng biostrat i graphy evolutionary studies and paleoenvironmental reconstruction efforts Nineteenth-century research on these microfossils initiated the description of hundreds of genera both modern and ancient (e.g. Lamarck 1801; Montfort 1808; D Orbigny 1839) Studies on living foraminifera began with Lister s (1895) work on life cycles and research on the biology of these protists continued into the 20th century (e g., Le Galvez 1935) Cushman (1927) developed the modern taxonomic classification of foraminifera and was the first to investigate their economic use in petroleum exploration In the last 60 years, investigative methods for larger' calcareous foraminifera have been developed and used in stratigraphic correlations evolutionary studies and biological investigations (Drooger, 1993). More recently particular attention has been paid to the benthic foraminifera inhabiting coral reef env i ronments of trop i cal and subtropical lati tudes c i rcling the globe Many extant larger species host photosynthetic algae (e.g., diatoms and 1

PAGE 12

chlorophytes) in a symbiotic relationship that is exceedingly beneficial in the calcification and growth of these marine protists (Lee and Anderson, 1991 ) In their ecological similarities to zooxanthellate corals there lies considerable value for use as sensitive environmental indicators and as proxies for studies of reef health (Hallock 1995). Although Amphistegina is relatively small and morphologically simple compared with most other larger taxa it is one of the most important genera Not only are they the most widespread and abundant extant genus with algal symbionts (Hallock 1988) but they have a fossil record extending back more than 50 million years providing the ancestral stock for at least one lineage of orbitoid larger foraminifera the Discocyclinidae (Giaessner, 1945). Through the i r history they have been prolific calcium carbonate sediment producers, constituting as much as 90% of the sand-sized sediments in some Pacific atolls (McKee eta/. 1959) Benthic habitat depths of these foraminifera range from very shallow waters to over 1OOm, with density distributions on the order of 104 -107 individuals per square meter on suitable hard or phytal substratum (Hallock et a/., 1986a) Members of this genus typically grow to 1-3mm in diameter, harbor diatom endosymbionts, and have a life span of three to six months depending on environmental conditions (Hallock 1981; Hallock eta/., 1986b) Populations of Amphistegina spp possibly worldwide, are currently subject to an unknown stress that was init i ally discovered in the Florida Keys in the summer of 1991 (Hallock et at., 1993) Color loss commonly called bleaching," was found in more than 80% of post-juvenile (>0.6mm in diameter) 2

PAGE 13

Amphistegina gibbosa specimens collected at 1 0-30m depth on Conch Reef and other sites in the Florida Keys (Hallock eta/., 1995). Cytological investigation revealed symbiont digestion empty pore vaults and irreversible cytoplasmic damage (Talge and Hallock 1995). Monthly monitoring of the Conch Reef population since the onset of this disease revealed that the proportion of individuals exhibiting symbiont loss increases each spring peaks near the summer solstice and declines in late summer and fall (Hallock eta/., 1995) Population densities in September 1992 were 5% of those in September 1991, indicating elevated rates of mortality reduced reproductive success or both Partial recovery of densities has occurred as the incidence of bleaching has declined (Hallock eta/., 1995). In addition to symbiont loss and reduced population densities other major symptoms observed include : an apparent increase in morphologic variability in both juvenile and adult size classes ; shell damage (breakage epiphytization parasitization) ; progressive irreversible cytoplasmic damage ; and reproductive anomalies, including successive asexual generations previously unknown in this genus Detailed study of this event is providing invaluable information on how foraminifera respond to stress However symbiont loss itself does not leave a permanent record in the sediments. Understanding morphologic responses will be the key to recognizing similar stress events in the fossil record 3

PAGE 14

Purpose The purpose of this study was to assess morphological responses using biometric analysis of selected internal and external parameters to prolonged stress in Amphistegina gibbosa populations Specific questions that I addressed include : 1 Has variability in external morphology {i e test shape) increased in stressed field populations of A. gibbosa as compared with populations sampled before 1991? 2 Does test shape in stressed populations show temporal or bathymetric trends? 3 Do embryonic chamber dimensions of specimens collected from stressed field populations e x hibit significant anomaly or increased variability when compared to specimens collected before 1991? 4 Does protoconch size-range in individuals from stressed field populations demonstrate a departure from the size-range demonstrated in a normal dimorphic life cycle? The significance of understanding these stress effects is two-fold It provides insight concerning the adaptability of foraminifera to ongoing global environmental change and it will be helpful to investigators studying episodes of environmental change in the fossil record. 4

PAGE 15

Life cycles in foraminifera To investigate the hypothesis that environmental stresses have had an impact on the morphology and reproduct i ve strategy of this foraminifer it is necessary to consider the embryonic elements of the test. Thus it is useful to present a selective review of foraminiferal life cycles as they relate to morphology A more complete review can be found i n Lee eta/. (1991 ) Heterostegina depressa a large orbito i dal, symbiont-bearing benthic foraminifer is similar to Amphistegina gibbosa in i ts circumtropical occurrence habitat, life cycle and general biology Both its normal and environmentally modified life-history strategies have been studied extensively by Rottger et a/. (1986, 1990a) and R6ttger (1990) Rottger s research is invaluable to this study, as it provides a possible explanation of the reproductive anomalies observed in Amphistegina gibbosa specimens collected from stressed field populations from the Florida Keys since 1991. Rottger eta/. (1986) reported that reproduct i on i n Heterostegina and many other foraminiferal genera is metagenic ; that is dimorphic generations of sexually and asexually-reproducing individuals alternate as shown in Figure 1. The agamont (b) is diploid multinucleate and reproduces only asexually Its reproductive nuclei undergo meiosis (unlike other protozoa) and the parental cytoplasm is evacuated from the test ; asexual multiple fission then commences outside the parent's empty shell (R6ttger et a/., 1986) An unknown mechanism efficiently separates the diatom endosymbionts from the parental cytoplasm 5

PAGE 16

during these events, then redistributes them to the daughter cells when fission is complete Each asexual reproduction normally produces several hundred haploid young of uniform size shape and color in both Heterostegina and Amphistegina (Rottger eta/. 1986 ; Hallock 1985). By this mechanism, the uninucleate offspring already harbor the diatom endosymbionts necessary for their survival, growth and calcification When they reach adult size these gamonts (a) undergo mitotic gametogenesis in wh i ch the entire cytoplasm in converted to gametes and shed (Rottger et a/ 1986) Successful gametic union produces a zygote which matures into the diploid agamont and begins another cycle Reproduction whether se x ual or asexual normally occurs in adult size classes Gamonts and agamonts differ in reproductive strategy nuclear state and morphology producing clear gene r ational dimorphism. The pr i mary mechanism to distinguish between these products of sexual and ase x ual cycles lies in differences in their embryonic chamber dimensions (Rottger eta/. 1990a) The first and second chambers of the embryon are called the protoconch and deuteroconch respectively The dimensions of the protoconch are establ i shed during initial calcification and are determined by the volume of protoplasm around which it is built. Thus the size of the embryon is recorded during ontogeny preserved through adulthood and can be extracted from equator i a l sections The agamont is microspheric as its protoconch is noticeably smaller than that of the megalospheric gamont. In a study of Heterostegina depressa (Rottger eta/. 1990a) microspheric spec i mens had protoconch d i ameters on the average 6

PAGE 17

of 281-Jm; protoconchs of megalospheric gamonts averaged 1251-Jm. A clear understanding of this distinction is reached when one considers that daughter cells produced by asexual multiple fission (immature gamonts) begin life much larger than do products of zygosis (primordial agamonts) Through this distinction, a researcher can discern the ontogenetic history of specimens by analyzing their chambers in cross-section Murray (1991) expressed these morphological distinctions by designating Type A individuals as those with large proloculi (first chambers) and smaller tests and Type B individuals as those with small proloculi and larger tests. In addition he states that Type A morphologies (megalospheric gamonts) are far more numerous in natural populations than Type B morphologies (microspheric agamonts) Microspheric specimens are generally rare in field collected samples (Drooger 1993) because asexual reproductions contribute far greater numbers of offspring to the population than do gamete-broadcasting events Figure 2 illustrates the atypical trimorphic life cycle that has been observed in H depressa specimens collected from shallow environments with very low population densities (Rottger eta/., 1986) The broken arrows indicate how the megalospheric, haploid products of asexual reproduction (expected to be gamonts) may deviate from classical dimorphic alternation of generations by undergoing asexual multiple fission instead of gametogenesis once they reach maturity These "schizonts differ from gamonts in more than just their reproductive activity Observations in culture revealed that they reproduce at 7

PAGE 18

smaller s i zes have slower growth rates and possess protoconchs with smaller more variable diameters (R6ttger 1990 ; R6ttger eta/., 1990b) Because of the adaptive complexity and difficulty in studying complete life cycles there was some confusion initially in R6ttger s work. The suppression of the sexual cycle seemed so persistent that he published (R6ttger et a/., 1986) and later retracted (R6ttger et a/., 1990a) the hypothesis that what he was seeing was the emergence of a new apogamic species. This mistake was beneficial however, as it led to the realization that several generations of agametic schizogony may take place before sexual reproduction occurs It is clear that this cycle of multiple schizont formation exists, although the mechanism(s) controlling its initiation duration and closure are still largely unknown It may be an adaptive life history strategy during stressed conditions when population densities are low and the probabilities for gametic union and zygotic success are limited The existence of trimorphic life cycles in foraminifera has been postulated since the early part of this century by researchers who found some specimens didn t fit into classical life cycles (Leutenegger 1977; Drooger, 1993). The occurrence of three biologically-different forms in a single life cycle has been reported in several foraminiferal genera but complete generations suites must be observed from birth through reproduction over several cycles in order to prove trimorphism This has been documented only in Heterostegina depressa (R6ttger eta/., 1990a) 8

PAGE 19

AGAMONT microspheric d i p loi d Figure 1 Normal dimorphic life cycle of Heterostegina Amphistegina and many other foraminiferal genera In metagenesis generations of sexual reproduction (via gamete broadcasting a) and asexual reproduction (via multiple fission b) alternate (Adapted from Rottger et a/ .. 1986, 1990a .) ZVGOTE ;-----, .. ::_ '-._ haplo i d SCHIZONT mega l ospheri c haploid AGAMONT megalospheri c diploid Figure 2 Trimorphic life cycle of Heterostegina depressa collected from margina l environmental cond i tions (Adapted from Rottger et a/.. 1986 1990a ) 9

PAGE 20

METHODS Sample collection Field populations of Amphistegina gibbosa were sampled in June and September 1994 and in January March and June 1995 from sites at 18m and 30m depth on Conch Reef an extensive spur and groove formation lying 8km offshore of Key Largo in the Florida Keys (Figure 3). Field samples were also collected at 18m depth in May 1995 from nearby Molasses Reef for comparison with archive specimens from the same site collected in May 1984 Table 1 summarizes the sites depths and dates of the field collections Additional samples used in this study are also listed in the table, and the significance of each is discussed separately in a later sect i on Sample collection was carr i ed out following procedures established by Hallock eta/. (1993). For each sample SCUBA divers collected pieces of reef rubble into labe led plastic bags and transported them to the surface Sediment filamentous algae and attached foraminifera were later removed by scrubbing the rubble with a small brush and washing the pieces in seawater at the shore lab in Key Largo 10

PAGE 21

SOURCE/SITE DEPTH DATE R18-694 Conch Reef 18m 6/94 R18-994 Conch Reef 18m 9/94 R18-195 Conch Reef 18m 1/95 R18-395 Conch Reef 18m 3/95 R18-695 Conch Reef 18m 6/95 R30-694 Conch Reef 30m 6/94 R30-994 Conch Reef 30m 9/94 R30-195 Conch Reef 30m 1/95 R30-395 Conch Reef 30m 3/95 R30-695 Conch Reef 30m 6/95 R-lYPE S Conch Reef 30m 1994-95 Mo l asses Reef" 18m 5/84 Molasses Reef 18m 5/95 Puerto Rico* NA 11/82 Laboratory culture NA 199 1 -92 Micros herics Laborato culture NA 1994 Table 1 Summary of sample sites, depths and dates of collection An asterisk (*) indicates specimens were taken from archive samples -M olasses Reef Conch Reef Figure 3 Map of sample collection sites in the Florida Keys 11

PAGE 22

The resultant sediment slurry was transported to our laboratory in St. Petersburg decanted into a 150 x 20mm petri dish and placed in an environmental chamber ( cf. Hallock et a/., 1986b for discussion of light and temperature conditions) Living A. gibbosa specimens move to the top of the sediment when left for several hours and these were picked from sediment samples with the aid of a stereomicroscope From each sample examined a minimum of 30 individuals with test 0 8mm were randomly selected rinsed in deionized water dr i ed marked on their umbilical sides and glued to numbered squares on a micropaleontological slide labeled with the date depth and site of collection These specimens were then used for morphometric analysis. External morphology Each individual was placed under a stereomicroscope fitted with a M i croComp Image Analysis System which projected the specimen's image onto a computer monitor The equipment was calibrated using a Zeiss stage micrometer and repeated measurements had a precision of 0.001 mm A digitizing pad and mouse were used to first trace the outline of each specimen s equatorial view Maximum (Dmax) and intermediate (Dint) diameter lines were then drawn perpendicular to one another through the protoconch region as shown 12

PAGE 23

in Figure 4b The whole specimen v iew of each i nd i vidual was described and any u n usual features were noted ( e.g breakage and general morphotype ) Each spec i men was removed one at a t i me from the m i cropaleontological slide and set on its apertural edge beneath the imaging camera While several techniques were employed to p r event the specimens from falling over glue stick smeared onto a glass slide worked best to secure them The outline of each individual s radial view was traced and the measurements shown i n Figure 4a were collected The endpoint of the umb i lical rad i us line (Ru) was used as the start i ng point for the spiral rad i us l ine (Rs) such that these two measurements could s i mply be added together to obtain the specimen s minimum diameter (Dmin) The umbilical faces of each specimen had been marked for identificat i on and this prevented confusing the two radii. Each spec i men was glued back to its numbered square awaiting sectioning The equatorial and rad ial measurements were entered i n Quattro 4 0 spreadsheets and analyzed graph i cally and stat isti cally The morphometr i c parameters were later subjected to corre lati on and variance ana l yses to determine their relat i ve significance Internal morphology Following external test measurements spec i mens were sectioned and prepared for embryonic chamber analys i s To obtain med i an sect i ons each i nd i v i dual was or i ented umb ili ca l -side up i n a drop of warm Lakes i de 13

PAGE 24

I< )al Ru Rs I< )al Dmin (a) Radial view (c) Internal v i ew AB=Proto c on c h diamet e r CD=Deut erc on c h diamet e r EF=Embry o n he i ght Dint I< )al Dmax (b) Equatorial view (umblical side) B D Figure 4 Diagram of external and internal test measurements collected from each specimen 14

PAGE 25

thermoplastic cement atop a labeled glass slide. After the embedded specimen had cooled the umbilical face of the foraminifer was gently rotated by hand over wet fine-grained (400 grit) sanding paper The section was then rinsed blotted dry, dabbed with blue food coloring and examined under both a binocular scope and a transmitted light microscope to view the embryonic chambers. Several stepwise repetitions of sanding and examination were necessary to obtain satisfactory sections through the protoconch Unusual features were also noted for each specimen during sectioning Specimens successfully sectioned were removed from the thermoplastic by rinsing with methanol. They were then rinsed in deionized water, dried, mounted on an aluminum stub and sputter-coated with gold-palladium. A DS-130 scanning electron microscope (SEM) equipped with a 35mm camera was used to examine and photograph each spec i men at magnifications of approximately 270x and/or 360x Automatic numbering of photographs and the use of a keypad ensured identification of numbered specimens Images on the resultant negative strips were viewed one at a time over a small light source beneath the MicroComp video system previously described such that the photographic image appeared on the computer monitor. The equipment was calibrated using the scale bar visible on each micrograph, and repeated measurements had a precision within 11Jm. The internal biometric measurements (cf Drooger 1993 for discussion) illustrated in Figure 4c were collected using the mouse and digitizing pad from those specimens with clearly visible embryonic 15

PAGE 26

chambers Diameter lines were drawn through the protoconch and across the k i dney shaped deuteroconch ; when the entire embryon was visible a line perpendicular to these diameters was drawn to measure total he i ght. It was often difficult to measure the deuteroconch diameter and embryon height with certainty, thus protoconch diameters were collected for a greater number of ind i viduals Additional samples Three of the samples listed i n Table 1 (denoted PR 1982 TL-1991 and CR-Type S) have not yet been mentioned PR-1982 is a sample of 10 A. gibbosa specimens produced in culture by the multiple fission of a microspheric parent that was collected from Puerto Rico in November 1982 The biometric data on these specimens represents a known megalospheric generation. The TL-1991 sample is composed of 19 A. gibbosa specimens known to be schizonts These indiv i duals were part of a lineage cultured in 1991-92 by Talge (pers comm 1995) in which four successive asexual reproductions occurred The sample denoted CRType S consists of specimens selected from various field collections in 1994-95 based on their schizont-like morphology that was very similar to individuals from Talge s cultured lineage In addition four clone parents from a culture in March 1991 that were found to be microspheric upon sectioning are used as a reference for comparing relative protoconch s i zes 16

PAGE 27

Limitations of Methods A complete description of the morphotypic range of the Western Atlantic Caribbean species Amphistegina gibbosa is needed to investigate the degree of polymorphism present. The objectives of this study did not initially include the investigation of test morphology In retrospect more detailed examination of external parameters by Eigenshape analysis of both equatorial and radial views would have been more useful. Unfortunately, destruction of the shell by sectioning prevents reanalysis of features Problems are inevitably encountered during the delicate process of sectioning and cleaning specimens for examination of internal parameters. Brittle broken and deformed individuals often did not survive the sectioning process or displayed poor internal chamber preservation Occasionally micrographs of sections were masked by debris or were otherwise unusable Typically 50% of the specimens in each sample yielded data on protoconch size For statistical analysis these measurements were pooled by depth (18m and 30m) and by year (all Conch Reef collections from both depths in 1994-95) to provide satisfactory sample sizes Preservation of the entire embryon (nucleoconch) structure was even more difficult to obtain Data collected from the deuteroconch and the nucleoconch were not analyzed because even pooled sample sizes were i nsufficient. 17

PAGE 28

Conventionally protoconch diameters are optically measured us i ng a stereomicroscope (Drooger 1993) as it is a cheaper less time-consuming method than SEM analysis The use of scanning electron microscopy prov i des greater accuracy and more detailed examination but it does so at the expense of the number of specimens that yield internal data 18

PAGE 29

RESULTS External morphology Maximum diameters are consistently used by researchers to indicate size in Amphistegina and other larger foraminifera (e g., Larsen 1976; Hallock 1985; Drooger, 1993) Other measurements were statistically analyzed as the following morphologic ratios: Dmax/Dmin (elongation), Dmax/Dint (sphericity), Rs/Ru, Rs/Dmin and Ru/Dmin. These "secondary data" parameters are frequently used in biometric studies of foraminifera (Drooger 1993) Descriptive statistics of these parameters were compiled and are summarized in Table 2 The samples examined included archived specimens from 18m depth on Key Largo s Molasses Reef in 1984 (M018-584) specimens collected live from the same site in 1995 (M018-595), and five samples each from the 18m and 30m sites on Conch Reef collected quarterly in 1994-95 (e g CR18-694 was collected in June 1994 from 18m on Conch Reef) Data from the five Conch Reef 18m samples were pooled and analyzed (denoted CR-18m), as were those from the 30m site (denoted CR-30m), and from all ten samples collected from Conch Reef (denoted CR-all) 19

PAGE 30

Table 2 (facing page) Descriptive statistics summary of external morphologic parameters for each sample and for pooled Conch Reef samples Shown are : number of observations (n) mean, standard error (SE), minimum value (min) and maximum value (max) Dmax=maximum diameter ; Dmax/Dmin=elongation ; Dmax/Dint=sphericity ; Rs/Ru=spiral radius/umbilical radius ; Rs/Dmin=spiral radius/minimum diameter ; Ru/Dmin=umbilical radius/minimum diameter

PAGE 31

Dmax Dmax/Dm in Sa mole n Mean SE Min Max n Mean SE Min M018-584 48 0 6917 0 0268 0 3677 1 1251 M018-584 48 2 5450 0 0330 2. 0411 3 1093 M018-595 30 0 8984 0 0234 0 5871 1 2155 M018-595 30 2 4778 0 0427 1 9601 2 9703 30 1 1212 0 0391 0 6882 1 3971 CR18-094 30 2 5596 0 0540 2 0340 3 3271 30 1 1363 0 0440 0 7331 1 5830 CR18-994 30 2 6604 0 0551 1 9866 3.4166 30 1 .0899 0 0386 0 7418 1 5986 CR18-195 30 2 3995 0 0505 1 8344 3 1071 37 1 1745 0 0339 0 7384 1 5273 CR18-395 37 2 5926 0 0514 2. 1119 3 3956 33 1 0699 0 0322 0 7025 1 6093 CR18-095 33 2 7073 0 0523 2. 1922 3 5332 30 1.1098 0 0416 0 6941 1 5753 CR30-094 30 2 3208 0 0638 1 .7217 3 2920 CR30-994 30 1 1938 0 0413 0 7543 1 5691 CR30-994 30 2 2994 0 0579 1 8749 3 2548 30 1 0417 0 0248 0 8334 1 3380 CR30-195 30 2.4475 0 0475 1 8942 2 9414 30 1.3203 0 0318 0 9150 1 6349 CR30-395 30 2 5144 0 0734 1 6530 3 3746 "R30-095 30 1 0974 0.0373 0 6793 1 4639 CR30-095 30 2 3643 0 0814 1 5379 3 5742 160 1 1199 0.0168 0 6882 1 6093 160 2 5866 0 0247 1 8344 3 5332 lcR-30m 150 1 1526 0 0177 0 6793 1 6349 150 2 3892 0 0298 1 5379 3 5742 lcR-all 310 1 1357 0 0122 0 6793 1 6349 310 2.4911 0.0200 1 5379 3 5742 Dmax/D int Rs/Ru !sample n Mean SE Min n Mean SE M in M018-584 48 1 1135 0 0080 1 0155 1 2953 M018-584 48 0 7451 0.0194 0.4140 1 1068 M018-595 30 1 0837 0 0087 1 0033 1 1841 M018-595 30 0 .6822 0 0428 0 1042 1 1015 icR18-094 30 1 .0749 0 0141 0 8934 1 2874 30 0 7827 0 0542 0 2056 1 5803 icR18-994 30 1 0967 0 0128 0 9835 1 2134 30 0 7370 0 0502 0 3407 1 3791 CR18-195 30 1 0951 0 0106 1.0000 1 2233 icR18-195 30 0 6841 0 0332 0.2 183 0 9767 37 1 0887 0 0095 0 9534 1 1810 37 0 7642 0 0347 0.1776 1.2729 jcR18-095 33 1 1059 0 0118 1 0080 1 2871 33 0 7913 0 0354 0 3777 1 2619 icR30-094 30 1 1192 0 0103 1 0175 1 2596 30 0 5433 0 0528 0 1705 1 2726 icR30-994 30 1 1021 0.0102 1 0043 1 2122 30 0 3838 0 0192 0 1461 0 .6867 icR30-195 30 1 1137 0 0087 1 0065 1 1876 30 0 6209 0 0420 0 2236 1 0369 icR30-395 3o 1 0850 0 0084 0 9698 1 1681 30 0 6774 0 0660 0 1419 1 8287 lc;R30-$95 30 1 0758 0 0091 1 0000 1 1635 lc;R30-095 30 0 6615 0 0464 0 2519 1 2193 lcR-18m 160 1 0923 0 0052 0 8934 1 2873 160 0 7531 0 0187 0 1776 1 5803 lcR-30m 150 1 0992 0 0043 0 .9698 1 .2596 jcR-30m 150 0 6364 0.0249 0 1419 2 1916 lcR-all 310 1 0956 0 0034 0 8934 1 2873 lrR-all 310 0 6967 0 0157 0 1419 2 1916 Rs/Dmin Ru/Dmin n Mean SE Min Ma) !sample n Mean SE Min Ma M018-584 48 0 4236 0 0065 0 2928 0 5254 M018-584 48 0 5764 0 0065 0.4746 0 7072 M018-595 30 0 3926 0 0178 0 0944 0 5241 M018-595 30 0 6074 0 0178 0.4759 0 9056 lcR18-094 30 0.4232 0 0186 0.1705 0 6124 icR18-094 30 0 5768 0 0186 0 3876 0 8294 CR18-994 30 0.4113 0.0157 0 2541 0 5797 icR18-994 30 0 5887 0 0157 0.4203 0 7459 CR18-195 30 0 3990 0 0128 0.1792 0.4941 icR18-195 30 0 6010 0 0128 0 .5059 0 8208 CR18-395 37 0.4242 0 0129 0 1508 0 5600 icR18-395 37 0 5758 0 0129 0.4400 0 8492 CR18-$95 33 0.4346 0 0114 0 2741 0 .5579 33 0 5654 0 .011 4 0.4421 0 7259 CR30-694 30 0 3312 0 0212 0 1457 0.5600 30 0 6688 0 0212 0.4400 0 8543 CR30-994 30 0 3838 0 0192 0 1461 0 6867 30 0 6162 0 0192 0 3133 0 8539 CR30-195 30 0 3701 0 0174 0 1827 0 5090 30 0 6299 0 0174 0.4910 0 8173 icR30-395 30 0 3781 0 0234 0 1243 0 6465 30 0 6219 0 0234 0 3535 0 8757 lc;R30-095 30 0 3845 0 0170 0 2012 0 5494 lc;R30-095 30 0 6155 0 0170 0.4506 0 7988 lc;R18m 160 0 4190 0 0064 0 1508 0 6124 160 0 5810 0 0064 0 3876 0 8492 lc;R-30m 150 0 3696 0 0089 0 1243 0 6867 150 0 6304 0 0089 0 3133 0 8757 lcR-all 310 0 3951 0 0056 0 1243 0 6867 rR-all 310 0 6049 0 0056 0 3133 0 8757 21

PAGE 32

Normal A. gibbosa specimens generally possess biconvex or planoconvex morphotypes (Hofker 1964). Biconvex tests (Figure 5a) are characterized by nearly equal radii such that the ratio of the spiral side to the umbilical side (Rs/Ru) range from 0 8 to 1 0 Planoconvex tests (Figure 5b) are usually characterized by very low Rs/Ru values (down to 0.1) due to the distended nature of the umbilical side. While performing measurements on the radial views of specimens I noted that specimens in which the spiral radius was greater than the umbilical radius were about twice as common in the Conch Reef samples as in the archive Molasses Reef samples Analysis of variance (ANOVA) techniques were employed to test the significance of variability between several samples with respect to Rs/Ru (cf. Zar, 1984 for discussion) Table 3 summarizes the six analyses performed and lists the experimental F-values critical F-values and significance at the 95% confidence level of each test. Complete ANOVA results are provided in Appendix A. To ensure the validity of sample comparison between Molasses and Conch Reefs ANOVAs 1 and 2 were performed ANOVA 1 demonstrated significant variability in Rs/Ru existed between MO 18-584 and MO 18-595 ANOVA 2 found no significant difference in Rs/Ru values between M018595 and CR 18-694 thus concluding that the Molasses and Conch samples are constituents of the same geographic population and are statistically comparable ANOVA 3 revealed significant variability between the archive sample collected from 18m depth on Molasses Reef in 1984 (M018584) and those 22

PAGE 33

Figure 5 (facing page) Scanning electron micrographs plate Amphistegina gibbosa specimens shown : A: normal biconve x morphotype scale bar=500j..Jm; B : normal planoconvex morphotype scale bar=5001Jm ; C: schizont-like morphotype (Type "S"), scale bar=500j..Jm; D-E : deformed specimens scale bar=5001Jm ; F-L: variability of morphotypes found in 1994-95 samples from Conch Reef, scale bar=5001Jm ; M : equatorial section of test scale bar-1 OOj..Jm; N : megalospheric protoconch scale bar=100j..Jm; 0 : microspheric protoconch scale bar=1001Jm

PAGE 34

24

PAGE 35

collected from 18m depth on Conch Reef in 1994-95 (pooled as CR-18m) Histograms of Rs/Ru frequencies were constructed to illustrate those differences The shape of the histogram for M018-584 (Figure 6a) is similar to that for CR18m (Figure 6c} but highly planoconvex specimens 0 3) and specimens with spiral radii greater than umbilical radii (Rs/Ru > 1.0) were more common in the 1994-95 CR-18m samples Similarly ANOVA 4 tested for significant variability between samples collected in 1994-95 from 18m and 30m depth on Conch Reef Comparison ofF values indicated that the bathymetric variability in Rs/Ru is also significant. Comparison of histograms of Rs/Ru frequencies (Figures 6c and 6d) demonstrated that planoconvex specimens (Rs/Ru < 0 5) were far more common at the deeper site ANOVA 5 was performed to test for significant variability between the five quarterly samples collected from 18m depth on Conch Reef in 1994-95 Comparison of F values revealed that no significant seasonal variability exists with respect to the Rs/Ru parameter. A s i milar conclusion was gathered for the five quarterly Conch Reef 30m samples from the results of ANOVA 6 2 5

PAGE 36

AN OVA Scale of variability Samples included F-value F-crit Signif? 1 Long term M018-584 and M018-595 29 .83 3 .84 yes 2 Geographic M018-595 and CR18-694 2 .11 4 .01 no 3 Long term M018-5 84 and CR-18m 251. 38 3 87 yes 4 Bathymetri c CR-18m and CR-30m 23 53 3 87 yes 5 Seasonal (18m) CR18-$94 CR18-994 CR181 95, 2.34 2 42 no CR18-395 and CR18-$95 6 Seasonal (30m) CR30-$94, CR30-994, CR30-195, 1 05 2.43 no CR30-395, an d CR3 0-$95 Table 3 Summary of variance analyses employed to test for significant variability in the Rs/Ru parameter. Significance level 95% Complete ANOVA results are provided in Appendix A 26

PAGE 37

a b M018-584 d CR-18m 40 "' u ; 20 :I r:r I!! u. 10 0 30 25 C2o "' u i 1 5 :I r:r .t 10 5 0 M018-595 CR-30m Figure 6 Frequency histograms of the parameter Rs/Ru Shown : samples collected from 18m depth on Molasses Reef in 1984 (a) and 1995 (b); samples collected from Conch Reef during 1994-95 at depths of 18m (c) and 30m ( d ) 27

PAGE 38

Protoconch diameter Measurement of protoconch s i ze is the primary method used by researchers to distinguish between products of asexual and sexual reproduction (Rottger eta/., 1986 ; Drooger 1993) Individuals produced by asexual multiple fission are called megalospheric," while those resulting from gametic union are termed microspheric" (Rottger eta/. 1986 ; Lee eta/., 1991) The internal morphology of equatorial sections was examined as illustrated in Figure 4c. Only protoconch diameters are reported here as t he number of useable sections revealing deuteroconch diameter and embryon height were insufficient for analysis Scanning electron micrographs of megalospheric and microspheric sect i ons of A. gibbosa specimens are shown in Figure 5 Plates Nand 0 respectively Descr i ptive statistics compiled for the protoconch diameter data collected from each of the samples are summarized in Table 4 The samples examined and the statistics shown are the same as for the external data, though sample sizes are smaller because only -50% of the sect i ons yielded useable protoconch diameters Analysis of variance techniques were again employed to test the sign i ficance of variability between samples w i th respect to protoconch size. Table 5 summarizes the six analyses performed and lists the experimental F values, critical F-values, and significance of each test at the 95 % level. The ANOVA tests performed on protoconch data (Appendi x B) had the same purpose and simi lar results as those discussed i n the prev i ous section on 28

PAGE 39

external morphologic variability The results of ANOVA 1 established significant protoconch diameter variability between Molasses Reef samples from 1984 and 1995 ANOVA 2 established that Molasses and Conch Reefs are geographic subpopulations that can be reliably compared. ANOVA 3 revealed that significant variability existed between the archive Molasses Reef 18m sample ( 1984) and the pooled Conch Reef 18m samples (1994-95). The frequency histograms of protoconch diameters for these two samples are difficult to interpret. The M018-584 sample (Figure 7a) is vaguely trimodal with nearly 8% of the protoconchs < in diameter. The CR-18m sample (Figure 7c) is unimodal with< 1% of the protoconchs < in diameter and with a greater proportion of protoconchs > in diameter. Strong peaks over the protoconch diameter size-range are present in both samples. ANOVA 4 showed that the depth-related variability in protoconch size that existed between pooled 18m and pooled 30m Conch Reef samples collected during 1994-95 was also significant. The protoconch diameter histogram for CR30m (Figure 7d) approaches a normal distribution The deeper site contains more specimens with very small protoconchs and the peak is stronger in the 30m samples. As with the Rs/Ru parameter no significant seasonal variability (between the five quarterly samples at each depth) could be detected in ANOVA 5 (18m) or ANOVA 6 (30m). 29

PAGE 40

Protoconc h diameter Sample n Mean SE Min Max M018-584 20 36. 9 1 5 15 4 60. 0 M018595 12 39. 1 1 9 26. 1 55 0 16 37 0 1 3 28. 6 55 9 13 33. 5 1 0 27 0 44 4 7 40. 0 2 1 25.4 54 3 13 41. 1 1 8 22 5 60.4 10 45 9 1 5 37 0 57.4 17 37 9 1 6 22 0 52 3 pR30-994 19 34 1 1 3 21. 0 49.7 11 35. 7 1 3 22. 2 46. 8 20 36. 8 1 6 23 3 54 3 17 40. 7 1 3 25. 6 53 6 59 39. 0 1 2 22 5 60.4 84 37 0 0 9 2 1 0 54 3 143 37 9 0 7 2 1 0 60. 4 Table 4 Descriptive statistics summary of protoconch diameter measurements for each sample and for pooled Conch Reef samples Shown are : number of observations (n), mean, standard error (SE), minimum value (min), and maximum value (max) AN OVA Scale of variability Samples i ncluded F-value F-crit Signif! 1 Long term M018584 and M018-595 6 36 4 .10 yes 2 Geographic M018-595 and CR18-694 2 15 4 .17 no 3 Long term M018-584 and CR-18m 251. 38 3 .87 yes 4 Bathymetric CR-18m and CR-30m 23 53 3 87 yes 5 Seasonal ( 18m ) CR18-694, CR18 -99 4 CR18-195 2 34 2.42 no CR18-395, and CR18-695 6 Seasonal (30 m ) CR30-694, CR30-994 CR30-195 1 05 2.43 no CR30 -3 95. and CR30-695 Table 5. Summary of variance analyses employed to test for significant variability in protoconch diameters Significance level 95% Complete ANOVA results are provided i n Appendix B 30

PAGE 41

25 r: .. & 10 :!! ... 5 M018-584 CR-18m {b) ------M018-595 {d),_ ________ _, CR-30m Figure 7 Frequency histograms of protoconch diameter. Shown : samples collected from 18m depth on Molasses Reef in 1984 (a) and 1995 (b); samples collected from Conch Reef during 1994 95 at depths of 18m (c) and 30m (d) 31

PAGE 42

Additional samples Descriptive statistics on the external and internal parameters (Dmax, Dmax/Dmin Dmax/Dint Rs/Ru Rs/Dmin Ru/Dmin, and protoconch diameter) were compiled for samples PR-1982 TL-1991, and CR-Type S ; they are shown in Table 6. To i nvestigate the presence of the schizont morphology" in field populations I compared the Rs/Ru parameter of known schizonts (TL -1991) to that of suspected schizonts (CRType S) At-test of means detected no significant difference in mean Rs/Ru values between the two samples I also performed an AN OVA to compare sample variances of the Rs/Ru parameter and no significant difference was detected between known and suspected sch i zonts. The complete experimental results of these tests are prov i ded in Appendix C Protoconch size-frequency distributions were also constructed for known microspher ics (Figure 8a), for known megalospheric specimens from the PR1982 sample (Figure 8b), for known schizonts of Talge s lineage (Figure 8c) and for suspected schizonts from field collections (Figure 8d) The microspheric and megalospheric histograms illustrate the normal dimorphic variability present in Amphistegina gibbosa protoconch diameters. Known schizonts can possess a wide range of protoconch diamete r s but a conspicuous peak is present over the 35-451-Jm range. The protoconchs of specimens possessing a schizont-like morphotype all fall within the 35-551-Jm range 1 compared the protoconch diameters from TL-1991 and CR-Type S using the same statistical methods : a t-test of means and an analysis of variance No 32

PAGE 43

significant difference in protoconch size could be detected between known and suspected schizonts (Appendix C). These results demonstrate that the field morphotype we believe to be "schizont-like" is not significantly different from that of known schizonts from Talge's lineage (with respect to the external parameter Rs/Ru and protoconch size). Parameter Sample n Mean SE Min Max Dmax PR 1982 10 1 1708 0.0532 0 9361 1 3927 TL-1991 19 0 9906 0 0312 0 8177 1 4840 CR Type S 16 0 8999 0.0405 0 6401 1.1877 Dmax/Dmin PR-1982 10 2.4347 0.0554 2 0923 2 6752 (Elongation) TL-1991 19 2 9330 0 1000 2 2808 3 9619 CR Type S 16 3 1650 0 1045 2 5732 4 0874 Dmax/Dint PR -1982 10 1 1294 0 0173 1 0579 1 2227 (Sphericity) TL-1991 19 1 1444 0 0307 0 9882 1 6100 CR Type S 16 1 0974 0 0141 1 0100 1 1969 Rs/Ru PR-1982 10 0 6079 0 0661 0 3945 0 9547 TL-1991 19 0 9179 0 0668 0.4638 1 7298 CR-Type S 16 0 7709 0 0465 0 3944 1 0717 Rs/Dmin PR-1982 10 0 3689 0 0249 0 2829 0.4884 TL1991 19 0 4645 0 0181 0 3168 0 6337 CR -Type S 16 0 4575 0 0159 0 3364 0 5672 Ru/Dmin PR-1982 10 0 6311 0 0249 0 5116 0 7171 TL1991 19 0 5355 0 0181 0 3663 0 6832 CR Type S 16 0.6180 0 0298 0 4324 0 8528 Protoconch diam. PR-1982 4 77.8 5.4 65. 5 103 0 TL-1991 13 53 8 3 1 37 1 85 2 CR -Type S 8 47 6 2 0 37.4 58 6 Table 6 Descriptive statistics summary for known specimens (PR-1982), known schizonts (TL-1991), and "TypeS" field spec1mens S) Shown are: number of observations (n) mean standard error (SE), m1mmum value (min), and maximum value (max) 33

PAGE 44

a c 60 50 !!40 ,., u :; 30 20 u.. 10 0 25 20 e15 .. 6-10 u.. 5 Known microspherics Known schlzonts [11_) J b d 25 20 c .. 610 u.. 5 Suspected s chlzonts (CRType S) Figure 8 Frequency histograms for known microspher i c spec i mens ( a) known megalospheric specimens (b) known schizonts (c) and suspected schizonts (d). 34 l

PAGE 45

DISCUSSION External variability Shell biometry is a widely-used tool in the study of foraminifera Morphotypes have previously been analyzed with digitizing equipment for use in bathymetric (e.g., Spencer, 1992) stratigraphic (e.g., Saraswati 1995), and evolutionary (e.g., Drooger, 1993) studies Because the developmental stages of foraminiferal life histories are encapsulated in the test biometry of internal chambers is useful in life cyc l e investigations of both modern (e g., Rottger eta/. 1986) and fossil (e.g. Biekart eta/., 1985) foraminifera. Drooger (1993) summarized his lifelong research on radial foraminifera and prov i ded guidelines for biometric studies of external and internal morphology. Sphericity and elongation Of the morphometric ratios (secondary data) the parameters of sphericity ( Dmax/Dint) and elongation (Dmax/Dmin) are of common use in foraminiferal test biometry (Drooger 1993) Both parameters are functions of test size which 35

PAGE 46

reflects age in these foraminifera (Hallock 1985) The relationship between size and sphericity is demonstrated by the linear plot of Dmax vs Dint (Figure 9a) Similarly, the relat i onship between size and elongation is demonstrated by the plot of Dmax vs Dmin (Figure 9b) Both graphs were constructed using data from the Molasses Reef 18m sample from 1984 and are correlated 94% and 98% respectively (see Appendix 0 for complete linear regression results) The sphericity and elongation of Amphistegina gibbosa shells are thus relatively constant functions of test diameter Other researchers (e.g., Larsen 1976; Hallock 1979) have reported that relationships between thickness and diameter for Indo-Pacific species of Amphistegina (and other rotaliine foraminifera) are approximately linear but tend towards a maximum thickness. This trend is especially evident in soritacean genera (e g., Archaias and Sorites) whose thickness to diameter plots are actually curvilinear (Hallock 1979) In A gibbosa spec i mens collected in 1984, the relationship between Dmax and Dmin remained linear even i n very large specimens (Figure 9b). Elongation (Dmax/Dmin) and its inverse (thickness-to-diameter ratio) are frequently used in depth-related morphocline studies (e.g., Spencer, 1992). Orooger ( 1993) Hallock et a/ ( 1986) and Pecheux ( 1995) are among the researchers who have reported that trochospira l and plan i spiral tests of larger benthic foraminifera become flatter and elongation values increase (T/0 values decrease) with increasing depth of habitat. This change is interpreted as a shade adaptation attributed to lower light levels and the corresponding 36

PAGE 47

decreased act i vity of photosynthetic symbionts (Drooger 1993) Hallock eta/. ( 1986b) showed fo r A. lessoni that it is also related to decreased water motion These studies show that biconvex (Ru"'Rs) and planoconvex (Ru>>Rs or Rs>>Ru) forms were generally more abundant in shallower habitats, while compressed forms dominated in deeper waters Contrary to these studies of Amphistegina in the Indo-Pacific the Caribbean species A. gibbosa demonstrates an opposite trend I found that planoconvex specimens consistently occur in greater numbers at the 30m site on Conch Reef and this is supported by what Hallock has observed since the Keys r esearch began in 1991 (pers comm., 1995) Thus the depth related morphoc l ine of A. gibbosa is not influenced by the shade adaptation observed fo r other species a 1.6 .---------, c: e. 1 4 (j; Qi 1 2 E "' i5 1 08 Q) E .2l 06 "" ,. . ""':. : ..!: 0.4 0.6 0.8 1 1.2 1.4 1.6 1 8 Maximum diameter {Om ax) b 0.8 ,.-----------------, c E 0 7 e. (j; 0 6 Qi 05 i5 E 0.4 :::1 E c 0 3 . .... ..:. : -. .. 0 2 0.4 0.6 0 8 1 1 2 1.4 1 6 1 .8 Maxim u m diame ter { Om ax ) Figure 9 Illustration of the linear relationship between Dma x and Dint (a) and between Dmax and Dmin (b) Data are from M018-584 37

PAGE 48

In field samples from Conch Reef biconvex tests were typically lenticular in shape when viewed along the test axis, but a continuous gradient existed between several sub-morphotypes (Figure 5, Plates F-L) Some specimens with nearly equal radii (Ru Rs) were lobate, pointed, or curved (Figure 5, Plates F-J), while others were twisted (Plate K) or globular (Plate L). One of the shape variations observed for biconvex tests was that of the schizont-like morphotype (Plate C) The "S" morphotype possesses a very wavy perimeter (axis or "keel") and an extremely thin, discoid test that is often pointed at both ends and abnormally curved Many of the schizonts from Talge's lineage of successive asexual reproductions demonstrated similar features. The irregular presence of Type S individuals ranged from 1 0-50% abundance in the samples with no perceptible trends. Elongation values for these individuals range from 2 6 to 4.1 (Table 6) which are much higher than the normal range of 2.0-2.7 observed for this species (Hallock eta/., 1986b). Known schizonts from Tatge s lineage were also highly elongate, possessing values of 2.3-4.0 (Table 6). These morphologic similarities indicate the previously undescribed "S" morphotype from the natural population possesses features of known schizonts. Variability in the spiral radius Normal A. gibbosa specimens are often unequally biconvex," and increased thickness (Dmin) is usually attributed to greater umbilical radii, as the spiral side often remains flat (Hofker 1964). The descriptive statistics in Table 2 38

PAGE 49

however, show that the field samples from Conch Reef in 1994-95 demonstrate a greater range in Rs/Dmin ratios (0. 1243-0.6867) compared to the archive sample from Molasses in 1984 (0. 2928-0 5254) Correlation coefficients between minimum diameter and spiral radius are listed for each field sample in Table 7. Specimens analyzed from the Molasses Reef 1984 sample exhibited 93% correlation between Rs and Dmin indicating the spiral side of the test was generally a "stable" feature dependent on minimum diameter. In contrast the spiral radius and minimum diameter measurements from 1994-95 field samples demonstrated reduced and highly variable correlations of 13-71% (Appendix E). Some of the variation observed in the elongation and Rs/Ru parameters therefore, must be attributed to the distended spiral radii of many specimens in the Conch Reef samples. This effect is occasionally pronounced to the point of deformity, as in very globular morphologies (Figure 5, Plate L). SAMPLE r M018-584 0 93 M018-595 0 46 CR18-694 0 34 CR18-994 0 .64 CR18-195 0 .71 CR18 395 0 .63 CR18-695 0 77 CR30-694 0.42 CR30-994 0.61 CR30-195 0.35 CR30-395 0.13 CR30-695 0 59 Table 7. Summary of correlation analyses between Dmin (x) and Rs (y) Complete results are provided in Appendix E. 39

PAGE 50

Sources of morphologic variability summarized Populations of some species of foraminifera tend to adapt to the environmental conditions of their local habitat, and distinct morphologies (ecophenotypes) often arise as a result (Drooger, 1993). The previously undescribed "flexibility of the spiral side of the test could be an ecophenotypic adaptation found in stressed populations as could be the novel Type S morphology thought to be indicative of schizonts. Both observations whatever their cause are significant sources of test shape variability in the samples collected from Conch Reef in 1994-95 The high degree of variability in external morphology present within and between 1994-95 Conch Reef samples is also due to the breakage calcification damage, and shell deformities that are among the observed stress symptoms (e. g., Figure 5 Plates D and E). Bathymetric variability is due to these factors plus the observed shift in dominant morphology with depth (i. e., planoconvex forms dominate in deeper habitats) Depth-related morpho l ogic trends in test shape of the Western Atlantic-Caribbean species Amphistegina gibbosa should be stressed as they are vastly different from those documented in Indo-Pacific species and other rotaliine genera Increased levels of UV-b (280-320nm) radiation are suspected of inducing the symbiont loss observed at Conch Reef as well as at other sites in the Flor i da Keys Australia Jamaica Hawaii and Palau (Hallock et. a/. 1995) If high levels of UV are to blame for the bleaching observed in these foraminifera since 1991 40

PAGE 51

perhaps they are causing increased mutation rates in the population If so the range of morphologic variability observed may be a partly genotypic response Protoconch size variability Reproductive anomalies and their ensuing effects have been noted in both field and culture studies since the onset of the stress in 1991 They persist throughout the year at all depths sampled but increase in frequency during the summer months Laboratory observations of reproductive activity show: reproduction in pre-adult size classes (0 5-0 9mm diameter) ; reduced numbers of offspring produced by asexual events ; highly variable sizes and shapes of offspring produced by ase x ual events including high incidence of juvenile deformity ; high juvenile mortality ; and successive asexual reproduction (previously undescribed i n this genus). Perhaps the most significant finding of the reproductive studies carried out since the onset of the degenerative disease involves Talge s lineage of successive asexual reproduction (Talge, pers comm. 1995) Specimens collected in November 1991 were cultured and monitored for reproduct ive activity as usual. When asexual reproductions occurred offspring which were maintained in culture and monitored for growth began to reproduce asexually at unusually small sizes (diameters 0.5-0 8mm) In the following year four 41

PAGE 52

successive asexual generations were observed within this single lineage Schizont production in the Amphistigenidae is undocumented, and th i s was the first observation of such activity in Amphistegina by Hallock and her students in 20 years of research on this genus (Hallock, pers comm., 1995) Megalospheric schizonts have been postulated in other genera of modern larger'' foraminifera (Leutenegger 1977) but until Talge s 1992 A. gibbosa lineage their actual formation by agamonts had been observed only in Heterostegina depressa (Rottger eta/., 1986 ; Rottger eta/., 1990a) Having positively established the existence of a third reproductive mode (the schizont) i n A. gibbosa questions arose as to whether or not this "biological trimodality could be detected in environmentally stressed populations as seen in H. depressa Evidence for trimodality in Amphistegina gibbosa Talge (pers. comm., 1995) observed four successive asexual generations in an A. gibbosa lineage whose original stock was collected from the profoundly impacted Conch Reef population in fall 1991. The external morphology of these known schizonts was similar enough to the field-collected S morphotype that the test measurements of the two samples could not be statistically distinguished Rottger et a/ ( 1986 1 990a) recognized that schizonts of Heterostegina depressa reproduced at smaller sizes and had slower growth rates than megalospheric gamonts Over the last four years of study on stressed A. gibbosa from the Florida Keys Hallock and T alge have observed that most of the indiv i duals that 42

PAGE 53

reproduced or attempted reproduction in culture were small specimens (0 50.9mm) and many morphologically appeared to be Type S individuals (i.e. suspected schizonts). Although slower growth rates were not detected in either schizonts or suspected schizonts recognition of a "schizont" morphology suggests the existence of trimodality in natural populations of A. gibbosa. Beyond the morphologic similarity between known schizonts and field collected Type "S" individuals further evidence that schizonts are present in the natural population can be inferred from the protoconch size distributions given in Figures 7 and 8. From Figure 8a we see that known microspheric A. gibbosa specimens have minute first chambers falling into the size range 15-251-.Jm. Protoconchs of known megalospheric specimens are much larger ranging from 651-.Jm to over 1001-.Jm (Figure 8b) Rottger eta/. (1986 1990a) demonstrated that protoconch diameters of Heterostegina depressa schizonts span a wide range overlapping and falling in between microspherics and normal megalospherics A similar result is illustrated in Figure 8c for Amphistegina gibbosa : the distribution of protoconch sizes observed for the known schizonts occupies a range from 351-.Jm to as large as 801-.Jm. Individuals possessing the Type S morphology have a similar distribution as shown in Figure 8d occupying the size range 35-551-.Jm. The conspicuous peaks that appear in the distributions of known and suspected schizonts at 35-451-.Jm were thus interpreted to be indicative of products of schizogony Surprisingly the telltale schizont peak was not only observed in the protoconch size distributions of the 1994-95 stressed field 43

PAGE 54

samples (i e., M018-595 CR-18m and CR-30m) but was present in the archived sample from 1984 as well (i. e M018-584) This ind i cates that schizogony was occurring commonly long before it was detected The M018-584 histogram indicates that schizonts were accompanied by both m i crospher i c products of sexual reproduction (agamonts) and megalospheric products of normal asexual reproduction (gamonts) in the natural population Thus from protoconch diameter ranges I concluded that the life cycle of this species is typically triphasic (although not obligatory) Variances of protoconch d i ameters compared between depths on Conch Reef were significant although the means for pooled samp l es were very similar (from Table 2) The normality of the 18m curve indicates that schizogony is the dominant mode of reproduction at this depth whereas at 30m both schizogonous and sexual reproduction take place Many researchers have noted that microspheric specimens are conspicuously absent in shallow habitats (e g., Drooger 1993 ; Lee eta/. 1991 ) and even in deeper waters (where se x ual reproduct i on is generally more common) they are uncommon Microspheric specimens have m i nute protoconchs and were easily distinguishable during sect i oning by the very small size of the first second and subsequent chambers Therefore I was able to tabulate the number of microspheric specimens encountered in each sample even if no embryonic data was later collected during SEM analys is. Of the 48 i ndividuals p r esent in the Molasses Reef 18m sample from 1984 nine were lost or ground th r ough before relative protoconch s i ze was noted Of 44

PAGE 55

the remaining 39 individuals three were classified visually as microspheric upon sectioning (-8%). The field sample from Molasses Reef 18m collected in 1995 contained no microspheric specimens in 30 sections Similarly out of the 137 sections from Conch Reef 18m only one microspheric specimen was found If microspheric specimens were occurring in 1984 proportions 13-14 specimens should have been found in modern 18m samples Microspheric individuals were found in "normal proportions in the 30m samples from Conch Reef ; 118 sections yielded 12 microspheric individuals (-10% ) This indicates that if se x ual reproduction is occurring at all at 18m the offspring survived to adulthood less frequently in modern samples than either in 1984 or at present at 30m depth In summary I conclude that the megalospheric schizont morphology is more common in the 1994-95 Conch Reef population than in 1984 particularly at 18m In addition the relative infrequency of microspheric specimens at 18m depth on Conch Reef during the sampling year indicates a shift to predominantly schizogonous reproduction possibly in response to stress The adaptive significance of this stress response i s discussed in the next section. Advantages of schizogony The hypothes i s that A. gibbosa has altered its life history strategy in response to environmental stress is supported intuitively when one considers the advantages of asexual reproduction in stressed conditions. Offspring produced by multiple fission begin life larger and receive their alga l symb i onts d i rectly from 45

PAGE 56

the parent. In contrast products of gametic union must find and incorporate a sufficient number of symbionts Drooger (1993) suggested that high light levels or other environmental stresses may inhibit the photosynthesis and free life" of algal cells, thus making it even more difficult for microspheric juveniles to acquire algal symbionts. Juvenile mortality rates are very high but decrease exponentially with increasing size of young (Hallock 1985) thus megalospheric progeny of asexual reproductions have energetic advantages and far greater chances of survival than do their microspheric counterparts. Reproduction by multiple fission is an effective method for maintaining high population densities. Population densities of A. gibbosa on Conch Reef experienced a relatively rapid recovery beginning in the fall of 1992 indicating a switch to successive asexual generations. Such a rapid expansion of numbers by asexual activity may provide maximum reproductive benefit for stressed populations. This calls for reinterpretation of the premise that sexual reproduction is advantageous under stressed conditions. Although sexual reproduction has the advantage of contributing genetic variability necessary for adaptation (e .g., Cushman, 1948; Sagan and Margulis, 1987) it is energetically expensive and has a high probability of failure in species that are gamete broadcasters. The odds for gametic union and survival to adulthood are limited in stressed environments with low population densities This may explain the reduced frequency of microspheric specimens in 1994-95 samples from 18m 46

PAGE 57

Evolutionary significance of trimodality Because successive asexual generations produce genetically-identical offspring, a mechanism is provided for the propagation of favorable genotypes. In a stressed population such as Conch Reef since 1991, successive asexual generations in the A. gibbosa life cycle enabled the rapid amplification of surviving genotypes This adaptation has implications not only for survival of environmental perturbations, but also for speciation Isolated subpopulations geographically cut off from the parent population could undergo relatively rapid genetic divergence in response to local environmental conditions A type of allopatric speciation could result, for example, during colonization of a newly emergent island or via a newly opened seaway. MacArthur and Wilson ( 1967) argued that successful colonizers are initially "r-selected" as they rapidly increase in population, but then shift to "k selection once the population is established The capacity for successive asexual generations provides Amphistegina and Heterostegina with this potential. Successive asexual generations also provide the potential for amplification of the "founder effect (MacArthur and Wilson 1967). Theoretically one foraminifer could found a new relatively isolated population. Such a population would carry a very limited gene pool with a high potential to be significantly different from the parent population. Further isolation could lead to the emergence of a new species 47

PAGE 58

While the two hypothetical examples just given are examples of possible allopatr i c speciation Rottger eta/. (1986) postulated that a subpopulation of ind i viduals undergoing repeated schizogony could result in sympatric speciation For example in an extreme case if mutagenic stress produced genotypes that were viable but reproductively isolated from parent stock schizogonous reproduction might allow that new genotype to establish a new species or (if the genotype was sufficiently different from the parent stock) even higher taxa All of these mechanisms may partly account for the rapid diversification of larger foraminifera at certain times in the geologic r ecord 48

PAGE 59

CONCLUSIONS 1. In Amphistegina gibbosa variability in test shape was greater in spec i mens collected in 1994-95 during a prolonged bleaching event than in specimens collected in 1984 2 Sources of increased variability in test morphology included high frequencies of deformed individuals prevalence of a schizont-like morphotype and a significant increase in variability of the spiral radius. 3 Variability in protoconch diameter was significantly different in specimens collected in 1994-95 than in those collected in 1984. 4 Significant variability was shown to exist between protoconch size distributions of field specimens collected from 18m depth and those collected from 30m depth 5 Conspicuously fewer microspheric (sexually-produced) specimens were found in 18m samples from 1994-95 than were found in 1984 Reproduction at this depth in 1994-95 appeared to be predominantly schizogonous 6 The presence of probable schizonts in the 1984 population (shown by protoconch size distributions) indicates that schizogony was common in Amphistegina gibbosa long before it was recognized by this study 49

PAGE 60

REFERENCES Biekart J.W. T. Bor R. Rottger, C W Drooger, and J .E. Meulenkamp 1985 Megalospheric Heterostegina depressa from Hawaii in sediments and laboratory cultures Proc Koninklijke Nede r landse Akademie van Wetenschappen, Ser B 88:1-20 Cushman J .A. 1928 Foraminifera : Their Classification and Economic Use Special Publications Cushman Laboratory for Foraminiferal Research. 1 : 1-401 Cushman J A. 1948 Foraminifera : Their Classification and Economic Use 4th Edn, Harvard University Press, Cambridge Mass D Orbigny Alcide Dessalines 1839 Voyage dans /'Amerique Meriodionale Foraminiferes: v.5. Pitois-Levrault et ce, Paris 86pp Drooger, C W. 1993 Radial Foraminifera: Morphometries and Evolution North Holland Press, New York 242pp Glaessner, M.F 1945 Principles of Micropaleontology. Melbourne University Press Melbourne. Hallock P 1979. Trends in test shape with depth in large, symbiont-bearing foraminifera J. Foram Res 9 : 61-69 Hallock P 1981. Light dependence in Amphistegina J Foram Res 11:42-48 Hallock, p 1985 Why are larger foraminifera large? Paleobiology 11: 195-208. Hallock, P 1988. Interoceanic differences in foraminifera with symb i otic algae : a result of nutrient supplies? Proc 6th lnternat. Coral Reef Symp 3 : 251-255 50

PAGE 61

Hallock P 1995. Amphistegina (Foraminiferida) densities as a practical reliable, low-cost i ndicator of coral reef vitality Submitted to : A coral reef symposium on practical, reliable low cost monitoring methods for assessing the biota and habitat conditions of coral reefs Not paginated Hallock P., T L Cottey, LB. Forward and J Halas. 1986a Population biology and sediment production of Archaias angulatus (Foraminiferida) in Largo Sound Florida. J Foram Res 16 : 1-8 Hallock P., LB. Forward and H J Hansen 1986b Influence of environment on the test shape of Amphistegina J Foram Res. 16:224-231 Hallock P., H.K Talge E.M Cockey and R.G Muller 1995 A new disease in a reef-dwelling foraminifer : Implications for coastal sedimentation J Foram Res 25 : 280-286 Hallock P., H .K. Talge K. Smith and E.M Cockey 1993 Bleaching in a reef dwelling foraminifer, Amphistegina gibbosa Proceedings 7th International Coral Reef Symposium Guam 1992 1 : 42 47 Hofker J 1964 Foraminifera from the tidal zone in the Netherlands Ant i lles and othe r West Indian islands. Studies on the Fauna of Curac;ao and other Caribbean Islands 21: 1-119 Lamarck J .B. 1801. Systeme des animau x sans vertebres The Author Paris 432pp. Larsen, A.R. 1976 Studies of recent Amphistegina taxonomy and some ecological aspects Israel J Earth Sci. 25 : 225-239 Le Galvez Jean 1935. Les gametes de quelques Foraminifereres Acad Sci. Paris Comptes Rend us v 210 p 1505-1507 Lee J J and O.R. Anderson 1991. Biology of Foraminifera Academic Press New York 368pp Lee J J W W Faber Jr., O .R. Anderson, and J Pawlowski. 1991. Life cycles of foraminifera In Lee J .J. and O .R. Anderson, eds., Biology of Foraminifera. Academic Press New York p 285-334 Leutenegger S. 1977 Reproduction cycles of larger foraminifera and depth distribution of generations Utrect Micropaleontological Bulletin 15 : 26 34 Lister J J 1895 Contributions to the life history of the Foraminifera Royal Soc London Philos Trans., ser B v. 186, p 401-453 51

PAGE 62

MacArthur Robert H and Edward 0 Wilson 1 967 The Theory of Island B i ogeography. Princeton Univers i ty Press New Jersey 203pp McKee E.D., J Chronic and E.B Leopold 1959 Sed i mentary belts in the lagoon o f Kapingimarangi Atoll. Bull. Am Assoc Petrol. Geol. 43 : 501-562 Montfort Denys de. 1808 Conchyliologie systematique et classificat i on methodique des coquilles : v 1, lxxxv i i 409pp Mur r ay J.W. 1991. Ecology and distr i but i on of benthic foramin i fera In Lee J.J and O.R Anderson eds. Biology of Foraminifera Academic Press New York p 221-253 Pecheux, M J .-F. 1995 Ecomorphology of a recent large foraminifer Operculina ammonoides. GEOBIOS 28 : 529-566 Rottger, R 1990. Biology of larger foraminifera : Present status of the hypothesis of trimorphism and ontogeny of the gamont of Heterostegina depressa In BENTHOS '90. Tokai Unive r sity Press Sendai p 43-54 R6ttger R M Fladung R Schmal j ohann M Spindler and H Zacharias 1986 A new hypothesis: The so-called megalospheric schizont of the larger foraminifer Heterostegina depressa D'Orbigny 1826 is a separate species J. Foram. Res 16: 141-149. Rottger R R Kruger and S de Rijk. 1990a Trimorph i sm i n foraminifera (protozoa): Verification of an old hypothesis Europ J Protistol. 25 : 226-228 Sagan D and L Margul i s 1987 Cannibal s relief : the origins of sex. New Scientist. 6 Aug : 36 40 Saraswati P K 1995. Biometry of early Oligocene Lepidocyclina from Kutch India. Ma r Micropaleontol. 26 : 303 3 1 1 Spencer R 1992. Quantified intraspec i fic variat i on of common benthic Foraminifera from the northwest Gulf of Me x ico : a potential pa l eobathymetric indicator J Foram Res. 22 : 274-292 Talge H K and P Hallock 1995 Cytological examination of symbiont loss in a benthic foraminifer Amphistegina gibbosa Mar Micropaleontol. 26 : 1 07113 Zar, J .H. 1984. Biostatistica/ analysis Prent i ce-Hall New Jersey 718pp 52

PAGE 63

APPENDICES 53

PAGE 64

01 Rs / Ru: ANOVA 1 (LONG TERM VARIABILITY) Analysis ofVaria n ce : One Wa y S u mmary Groups M 01 8 -58 4 CR18m An a l y sis o f Varia n ce Sourc e of Vari a tion Belween Groups Groups Total Count S um Average Vari a n ce 48 3 5 .765 9 9 0.74512 5 0 018092 1 60 12 0.499 3 0 .753121 0 .0 5 5836 ss d f 22 43669 28 38333 50 82002 M S F P -value F-crlt 1 22 43669 261. 3 754 4 02E-42 3 870867 3 1 8 0 089256 319 R s /Ru: AN OVA 3 (SEASONAL VARIABILITY18m) Ana lysis o f V ar i ance : One W ay Summary G r oups C R 1 8 -694 CR1 8-994 CR1 8-195 CR18-395 CR18-695 Analysis o f Variance Source of Variation Belween Groups Groups T otal C o unt S um A vera g e Varia n ce 30 23.4 7999 0 7 82666 0 088275 30 22 1087 0 .736957 0.0 75741 30 20. 52246 0 68408 2 0.033044 37 28. 27628 0.764224 0 04465 1 33 26.1119 1 0 .79127 0 04 1442 s s df 1 04426 20 09725 21.14151 M S F P va l u e F -crlt 4 0 261065 2 338214 0 057028 2 .42184 3 180 0 .111651 184 Rs /Ru: ANOVA 2 ( BATHYMETRIC VARIABILITY) Analysi s of Vari ance:On e Wa y Summary Groups CR1 8 m CR30 m A n a l ysis of V a r i anc e Source of V a riation Belween Groups Within Groups Total Counr Sum A ve r age V a r ia n c e 160 120. 49 9 3 0 753 1 2 1 0.0 55836 1 50 95.4 5 4 1 5 0.63636 1 0.0 92763 ss df 1 960 1 93 26.49612 28.45631 M S F P-valu e F-crlt 1 1 96019 3 23 6 2677 1 93E-06 3 870867 318 0 .083321 3 1 9 Rs/Ru : ANOVA 4 ( SEASONAL VARIAB I LITY 3 0m) A n alysis ofVariance: O n e W a y Summ a ry Grouos CR30-69 4 C R 30-99 4 C R30-19 5 CR30-395 CR30-695 Analysis of Varia n ce Source of Varia t ion B e lween Groups Groups To t al Count Sum A v e r a o e Var ia n ce 30 1 6 29855 0 543285 0 083576 30 20 3635 1 0 6787 84 0 1 3 1287 30 1 8 6269 0.6 20897 0 05300 1 30 20 32082 0 67736 1 0.1 3 068 2 30 1 9 84438 0 661479 0 064604 ss df 0 390416 13. 43132 1 3 .8217 4 MS F P va/u e F-crlt 4 0 09 7 604 1 .063699 0 .381781 2 434066 145 0 09263 149 '"'C)> m-u ::U'"'C .,m oz ;:uo m)> 0 0() zo I'"'C mr ;:um en-! _m ::U;:u em -uen )>C mO -111 m)> ::Uz )> en m en 0 ;:u )> z () m

PAGE 65

APPENDIX A. (cont.) COMPLETE RESULTS OF ANALYSES OF VARIANCE PERFORMED ON THE RS/RU PARAMETER Rs/Ru: ANOVA 5 Analysis of Variance:One Way Summary Groups M018-584 M018-595 Analysis of Variance Source of Variat i on Between Groups Within Groups Total Count Sum Average Variance 48 35.76599 0 745125 0 018092 30 20.46736 0 682245 0 055044 ss df 2.438 7 682998 10 .121 MS F P-va/ue F-crit 1 2.438 29.82846 3 8E-07 3.942303 94 0 .0817 34 95 Rs/Ru: ANOVA 6 Analysis of Variance:One Way Summary Groups M018595 CR18-694 Analysis of Variance Source of Variation Between Groups Within Groups Total Count Sum Average Variance 30 20.46736 0 682245 0.055044 30 23.47999 0 782666 0.088275 ss df 0 151265 4 156246 4 307511 MS F P-va/ue F-crit 1 0 151265 2.110891 0 151643 4.006873 58 0 071659 59 55

PAGE 66

(J"l 0> PROTOCONCH : ANOVA 1 (LONG TERM VARIABILITY) Analysis of Variance: One Way Summary Groups CR18m M018-584 Analysis of V ari ance Source of Varia t i on Between G r oups Within Groups T otal Count Sum Average Variance 60 2300 27 38 33783 111. 1688 21 737. 03 35.09667 168. 9993 ss df 20364 33 26752 69 4711 7 02 MS F P-vaiue F-crit 1 20364 .33 89. 8224 3.47 E -16 3 921478 118 226.7177 119 PROTOCON C H : ANOVA 3 (SEASONAL VARIABILITY18m) Analysis of V a r la n ce:""bne Way Sum mary Groups CR18-694 CR18-994 CR181 95 CR18-395 CR18-695 Analysis of Variance Source of Variation Between Groups Within Groups Total Count Sum Averaae Variance 16 592.71 37 04438 5 1 52875 13 435 26 33 48154 32 62113 7 279. 84 39 97714 126 4969 13 533. 6 7 41.05 1 54 123. 966 1 10 458. 79 45 .879 73. 51303 ss df 1753 449 59922 .62 61676 06 MS F P-vaiue F-crit 4 438 3622 1 1 33898 0 342645 2 430002 155 386 5975 15 9 PROT OCO NCH : ANOVA 2 (BATHYMETRIC VARIABILITY) Analysis ofVariance:One Way Summary Groups M01 8-584 CR-18m Analysis o f V a r i ance Source of Vari at ion Be twee n G r oups Within Group s Total Count Sum Average Variance 85 3111. 6 36 60 706 78 70043 60 2300.27 38 33783 111. 1688 ss df 3872 096 39107 26 4 2979 35 MS F P-vaiue F-crit 1 3872 096 16.63405 6 99 E -05 3 897407 168 232 7813 169 PROTOCON C H : ANOVA 4 (SEASONAL VARIABILITY30m) Analysis of Vari ance: One Way S ummary Groups CR30-994 CR3 0 1 95 CR30-395 C R30-695 A n alysis o f Va ria nc e S ource of Va ria tion Betwee n Groups Groups Tota l Count Sum Averaae Vari ance 17 644 5 37. 91176 76 6711 1 9 647.6 34 .08421 49 36918 11 392.4 35. 67273 47 03418 20 735.3 36.765 76 08134 17 691 8 40 69412 49 7 2184 ss df 2386 206 53584 1 2 55970 33 MS F P-vaiue Fcr/t 4 596 5516 1 614284 0 173778 2.434065 145 369 5457 149 ""O)> m-o :::0""0 "Tim oz :::co s: mx o9J Oo zo -os: :::0""0 or -lm o-l om o:::c zm (")(/) Ic r 0--l )>(/) s:o m"TI -l)> mz :::0)> r -< (/) m (/) 0 "TI :::0 )> z (") m

PAGE 67

APPENDIX B (cont.) COMPLETE RESULTS OF ANALYSES OF VARIANCE PERFORMED ON PROTOCONCH DIAMETER PROTOCONCH: ANOVA 5 Analysis of Variance: One Way Summary Groups M018-584 M018-595 Count Sum Average Variance Analysis of Variance Source of Variation Between Groups Within Groups Total 20 737 03 36 8515 109 8222 12 469 7 39 14167 112 .0117 ss df 1786 633 10672 69 12459 32 MS F P-value F-crit 1 1786 633 6.361291 0 .015973 4.098172 38 280 .8602 39 PROTOCONCH: ANOVA 6 Analysis of Variance:One Way Summary Groups M018-595 CR18-694 Analysis of Variance Source of Variation Count 12 16 Sum Average Variance 469 7 39 14167 112 .0117 592 .71 37.04438 51. 52875 ss df MS F P-value F-crit Between Groups Within Groups Total 472 .8581 6601.271 7074 129 1 472 8581 2 148941 0 153072 4.170877 30 220 0424 31 57

PAGE 68

01 CX> Rs/Ru: t-test !-Test: Two-S ample Mean Variance Observations Pea r son Correlation Poo l ed Variance df t P(T <=t) one-tail t Cri tical one-tail P(T <=t ) two-tail t Critical two-tail Rs/R u: ANOVA C R-Tvoe S TL-1991 0 77093481 0 917867 0.034 60867 0 10692 16 19 N A 0 1 5206572 32 0 40962183 0 34240628 1 69388874 0 68481257 2 03693334 Analysis o f Va riance:One W ay Summary GroiJ p_s CRType S TL-1991 Analys i s of Variance Source of Variation Between Groups W i thin Groups Total Coun t S u m A vera g e Varia n c e 16 12 33496 0 770935 0 034609 19 17.43947 0 917867 0 10692 ss df 0 .5 4283506 8 94831334 9 4911484 MS F Pv a/u e Fcrlt 1 0 5428 3 5 2 7 9 0516 0 1 0 1615 4 051749 46 0 .194 52 9 47 PROTOCONCH: t-test !-Te s t : T w o-Sample Mean Variance O b servations Pearson Corr ela t i o n Pooled V a riance d f P(T<=t) o n et ail t Critical one-tail P(T <=t) two-tail t Critical two-t ail CR Tvo e S TL 1991 47. 64 53. 78077 63. 2376 230.9521 8 1 3 NA 7 8 1.644851 33 0 33871084 1 6923603 0 67742168 2 0345153 PROTOCONCH : ANOVA A nalysis o f V a r i ance:One Way Summary G r o ups CR-TYPE S TL1 991 Analysis of V a riance Source o f V a riation Between Group s Within Groups T ota l Count S u m V a ria nce 8 381 12 4 7 .64 63. 2376 13 699.15 53. 78077 230 9521 ss df 2107 1475 2 32552 1709 34659 .3 184 MS F P-value F -crlt 1 2107 148 2.9 77 644 0 091135 4 051749 46 707. 6559 47 :::OO)> mz-u (1):;'\'"U -uzm m0z ()< --IZ() o....-... :::orJo ;Qcos: )>.._... r z)>m Oz--1 om '"U(/);;o :::ocm 0(/)(1) --1-uC Om ()()-I 0--i(/) ZmO Oo"'TI I..--.. (I) oO--i -:::0)> )>1--i s::--1m-<(/) --i'"U--1 mmo ;;o(l))> ......... (/)-I om I (I) t5(/) Z'"U --lm (1):::0 -:::0 :is: m 0

PAGE 69

APPENDIX D COMPLETE RESULTS OF LINEAR REGRESSIONS PERFORMED ON DMAX DINT AND DMIN FROM M018-584 M018-584: Correlation results x =Dma x y-Dint Regression Output: Constant Std Err of Y Est R Squared No. of Observations Degrees of Freedom X Coefficient(s) Std Err of Coef r x-Dma x y=Dmin 0 905233 0 023259 Regression Output: Constant Std Err of Y Est R Squared No. of Observations Degrees of Freedom X Coefficient( s) Std Err o f Coef r 0 392945 0 020517 59 0 00437 0 04142 0 970527 48 46 0 .98 5153 0 002285 0 0365 38 0 888563 48 46 0 942636

PAGE 70

APPENDIX E. COMPLETE RS/DMIN CORRELATION RESULTS FOR EACH SAMPLE Rs I Dmin: Correlation results M01 8-584 M018-595 Regression Output: Regression Output: Regression Output: Constant -0 .001 4586 Constant 0.026 1 2958 Constant 0 09359591 Std Err of Y Est 0 01903241 S td Err of Y Est 0.05123371 S t d Err of Y Es t 0 04617296 R Squared 0 65666056 R Squared 0 216046 R Squared 0 11640292 No. of Observations 46 No of Observa t ions 30 No of Observations 30 Degrees of Freedom 46 Degr ees of Freed om 26 Degrees of Freedom 26 X Coefficie nt(s) 0.42509236 X Coefficient(s) 0.33597926 X Coefftcient(s) 0.20437571 Std Err of Coef. 0.02563792 Std Err of Coef. 0 12094993 Std Err of Coef. 0 10641319 r 0 .925 5596 r 0 46460749 r 0 34117672 C R18-S 94 C R18-195 CR1 8-3 95 Regression Output: Regression Output: Regression Output: Constant 0 07417 Co n stant 0 01264764 Co nstant 0 04635672 Std Err of Y Est 0 03008879 Std Err of Y Est 0.03531941 Std Err of Y Est 0 03995476 R Squared 0 40455462 R Squared 0 50962062 R Squared 0 39670462 No of Observations 30 No of Observations 30 No of Observations 37 Degree s of Freedom 26 Degrees of Freedom 26 Deg r ees of Freedom 35 X Coefficie nt (s) 0.22942656 X Coefftcient(s) 0 36955297 X Coefficient(s) 0 3 143 7012 S td Err of Coe f 0.05 260 163 S td E rr of Coef 0.06650764 Std Err of Coef. 0.06552976 r 0 63604606 r o 7 136n16 r 0 62984492 CR30-994 R egre ssion O utput: R egression Output: Regression Output: Constant 0 04243 3 73 Cons tant 0.06636739 Constan t 0 02415542 Std E r r of Y Est 0 02646817 Std Err of Y Est 0.0570902 1 Std Err of Y Est 0 056 1 6921 R Squared 0.5904295 R Squared 0 17310053 R Squa red 0 36606892 N o o f Observations 33 N o of O b servat ion s 30 No o f Observations 30 Degrees of Freedom 31 Degrees of Freedom 26 Degrees of F reedom 26 X Coefficient(s) 0.32355 1 0 1 X Coefficient(s) 0 16466799 X Coefficient(s) 0 33443409 Std Err of Coer 0.04639965 St d Err of Coef. 0 0 763672 Std Err or coer 0 06317064 r 0 7663941 r 0.41605352 r 0.6050363 C R 3 0 -195 C R30 -395 Regression Output : Regression Output : Regression Output : Co n stant 0 .047 65116 Constant 0 23493073 Constant 0 04476756 S td Err o f Y Est 0.04214197 Std Err o f Y Est 0 05957969 Std Err of Y Est 0 04566 194 R Squared 0 11431356 R Sq ua red 0.0 1 576 553 R Squared 0 34762611 No of Observations 30 No of Observations 30 No of Observa tion s 30 Degrees of Freedom 26 Degrees of Freedom 26 Degrees of Freedom 26 X Coefficient(s) 0 .2567663 X Coefficien t (s) -0 0730665 X Coefficient(s) o .265no16 Std Err of Coef. o .1350m6 Std E rr of Coef. 0 1 0913234 Std Err of Coef 0.07396267 r 0 3361 0292 r 0.12556065 r 0 56959626 60