Habitat Utilization by Harpacticoid Copepods: a Morphometric Approach

MARINE ECOLOGY PROGRESS SERIES Vol. 35: 59-64, 1987 - Published January 27 Mar. Ecol. Prog. Ser. l

Habitat utilization by harpacticoid copepods: a morphometric approach

Susan S. Bell, Keith Walters & Margaret 0. Hall

Department of Biology, University of South Florida, Tampa. Florida 33620. USA

ABSTRACT Examination of harpacticoid copepod morphology was conducted to determine whether morphological resemblance provides a reasonable index of habltat utilization and movement. Discri- minant analysis was used to explore the relationship between body form and habitat utilization for copepod species collected from 3 subhabitats within seagrass beds in Tarnpa Bay, Florida. To examine the accuracy of our procedure the discriminant function derived for Tarnpa Bay copepods was apphed to copepods collected in worldwide studies. Three morphological characteristics - ratlo of the length of Pereopod 1 first endopod segment to the remaining endopod segments, area of the cephalosome, and length of the first antennule - significantly contributed to vanation in habitat util~zationby identifiable groups. Habitat utilization suggested from the literature corresponded well to that predicted by the dlscnrninant function derived for Tarnpa Bay harpacticoids. Our findings will be useful to suggest which copepods should (1) be associated with vegetation (2) display active migration and (3) be linked to sediment processes.

INTRODUCTION ment cores migrate from the sediments into the overly- ing water column. Thus differences in habitat usage by Extant families of harpacticoid copepods may have the wide variety of harpacticoid copepods in Tampa evolved from an epibenthic ancestor to exploit not only Bay seagrass systems may be related to behavioral sediment but also phytal and planktonic habitats (Por traits. Accordingly, species which are found Living on 1984). A wide range of harpacticoid body forms has seagrass blades, in the sediments, or moving into the accompanied the exploitation of these diverse habitats. water column from either blades or sediments, should While most sediment harpacticoids are vermiform or exhibit variation on one or more morphological charac- torpedo shaped, phytal or epibenthic species display a teristics given that each of these habitats requires host of body shapes including lateral and dorsal-ven- specialized features. We ask here whether specific tral compression of the body, extremely rounded or habitat utilization is related to, or can be predicted by, broadened cephalosomes and larger overall size morphological characteristics of harpacticoid cope- (Noodt 1971).The ecological literature on harpacticoid pods. copepods has traditionally focused upon studies of sediment-dwelling species (see Coull & Bell 1979, Hicks & Coull 1983 for reviews). Available information METHODS AND MATERIALS concerning copepod species living phytally, epibenthically or in the water column suggests that these Discriminant analysis was used to examine the rela- morphologically diverse forms also have ecological tionship between copepod morphology and habitat traits which differ from those of sediment-dwelling utilization (Norusis 1985). The SPSSX (Nie 1983) pack- species (Marcotte 1983, Hicks 1985). age of statistical procedures was used for all analyses. Studies on the distribution and abundance of sea- First, copepods from Tampa Bay seagrass beds were grass meiofauna in Tampa Bay, Florida, have iden- assigned a prion to one of 3 groups (phytal, water tified 3 habitats utilized by harpacticoid copepods column or migrator, sediment) based upon their dis- (Bell et al. 1984). Harpacticoids have been collected tribution in the field. Twenty of the most common from seagrass blade surfaces and within sediments. In copepod species representing 13 famihes from collec- addition, some copepods sometimes recovered in sedi- tions in seagrass habitats from Tampa Bay, Florida,

O Inter-Research/Printed in F. R. Germany 60 Mar Ecol. Prog. Ser. 35: 59-64, 1987

(Table 1) were used in this first analysis. We assumed able feature of characters used in such analyses. Mea- that collection of copepods from different subhabitats surement of cephalosome length and width and meta- of the seagrass bed reflected differences in habitat some length were considered in preliminary analyses. utihzation. Some species were collected from more Cross-correlations between these and the other mea- than one habitat but were assigned to the category surements were often above 0.9, thus the number of where they were relatively most abundant. Body mea- independent morphological variables was reduced to surements for each species were made on females and those given above. Measurements were made on either included length of antennule (AIL),length and width mounts of copepods with an ocular micrometer and of cephalosome (CL and CW, respectively), length of compound microscope or from llterature descriptions, urosome (UL),length of caudal rami including setae mostly Lang (1948), using vernier calipers. All mea- (CRL) and length of Pereopod 1 (P,) endopod first, surements, except those of P,, were converted to mm second, and third (if present) segments (Fig. 1).Length using information on the largest size of females. Val- and width of cephalosome were multiplied to obtain ues for P, endopod lengths were expressed as ratios of cephalosome area (CA). The ratio of length of P, measurement units because not all descriptions of Endopod 1 to length of PI Endopod 2 plus Endopod 3 pereopods contained scale dimensions (e.g. Lang (PIR) was calculated. The particular combination of 1948). measurements AIL, CA, PIR,UL, and CRL was chosen The 5 morphological variables for the 20 species because these specialized morphological charactens- from Tampa Bay seagrass habitats constituted the tics may be related to habitat utilization (Noodt 1971) independent variables which were used to discrirni- and provided the best set of non-correlated morpholog- nate between the habitat groups. Variables were log ical traits for dscriminant analyses, the latter a desir- transformed to meet assumptions of normdty and

Table 1. Species and famihes of copepods used in discriminant analyses from (A) Tampa Bay seagrass beds and (B)llterature reports. A pnori classification(P: phytal, M: migrator,S: sediment dweller)is noted in parentheses

A. Tampa Bay B. Literature Longipediidae Harpacticidae Longipedia americana (M) Harpacticus chelifer (P) Canuellidae Zaus spinatus (P) Scottolana canadensis(S) Tachidiidae Ectinosomatidae Microarthridion littorale(S) Ectinosoma rnelaniceps(M) Tisbidae Harpacticidae Tisbe furcata (P) Harpacticussp. A (P) Sacodiscuslittoralis (P) Zausodes arenicolus (M) Peltidiidae Tegastidae Neopeltopsis pectinipes(P) Parategastes sp.A (P) Eupelte beckleyae(P) Thalestridae Thalestridae Paradactylopod~abrevicornis (M) Dactylopodia vulgaris (P) Idomene forficata (P) Thalestrislongirnana (P) Dactylopodia tisboides (P) Parathalesh-sharpactoides (P) Dactylopodopsis sp.A (P) Phyllothalestris rnysis (P) Diosaccidae Parathalestris clausi (P) Diosaccus sp. A (P) Diosaccidae hphlascussp. A [P) Amphiascusminutus (P) Metidae Amonardia normani (P) Metis holothuriae (M) Ameiridae Ameiridae Ameira minuta (P) Njtocra sp. A (S) Tetragonicepsidae Leptomesochra sp. a (S) Ph yllopodopsyllusbradya (S) Canthocamptidae Canthocamptidae Mesochra pygmaea (P) Orthopsyllus Bneans(P) Louriniidae Cletodidae Loum-aarmata (S) Huntemanniajadensis (S) Cletodidae Nannopus palustris (S) Enhydrosorna herrerai (S) Enhydrosomapropmquum (S) Laophontidae Cylindropsylhdae Laophonte cornuta (P) Stenocans minuta (S) Heterolaophonte stroemi (P) Bell et al.: Habitat utilization by harpacticoid copepods 61

habitat affinity was reported. Groupings predicted by our discriminant function were then compared to the reported habitat utilization. A total of 21 copepods, including species from literature on fauna inhabiting seagrass beds (Decho et al. 1985), living associated with plants (Hicks 1976, 1977, 1982) or moving from sediments (Eckman 1983, Chandler & Fleeger 1983, Palmer & Gust 1985) (Table 1) were used. Our methods for collecting measurements for the literature species followed that of the Tampa Bay fauna. A priori category assignment (i.e. phytal vs water, column vs sediment) for literature copepods was extracted from the data provided in the studies. Discriminant scores for literature species were calculated using the functions generated for copepods in Tampa Bay seagrass beds.

RESULTS

Three morphological characteristics of Tampa Bay copepods were identified as significantly contributing to the variation between habitat utilization by groups (Table 2). P,R was the first variable to enter the model

rable 2. Summary table of (A) stepwise procedure using Fig. 1. Diagrammatic illustration of morphological measure- minimization of Wilk's lambda and (B) unstandardized ments made on harpacticoid copepods used in discriminant canonical discriminant coefficients. P, R: (length P, Endopod analyses. AIL: antennule length; CW: cephalosome width; 1) / (length P, Endopod 2 t 3); CA: cephalosome area; AIL: CL: cephalosome length: UL: urosome length; CRL: caudal First antennule length rami length; P,: Pereopod 1; END,: endopod first segment length; END,: Endopod 1 second segment length A. Stepwise procedure Step Variable Wilk's Significance assure a non-singular covariance matrix (Williams entered lambda (p value) 1983). Initially a stepwise procedure was run to iden- tify variables which provided the best separation 1 P1 R 0.604 0.013 between groups (i.e. phytal vs water column vs sedi- 2 CA 0.536 0.036 ment) such that between-group variabhty was greater 3 AI L 0.415 0.029 than within-group variability. A forward entry of var- B. Discriminant coefficients iables based upon the minimization of Wilk's Lambda Function 1 Function 2 was employed. Next, a discriminant function was generated for the 20 species from Tampa Bay using varia- CA bles identified as important in the stepwise pro- A, L cedures. The discriminant function provided coeffi- P1 R cients for each variable which codd then be used to Constant generate a discriminant score for each species. The proportion of cases in each a pnori category (see Table 1) was used as an estimate of the 'prior probabil- significantly (Norusis 1985). CA and AIL were the ity'. By using the discriminant function, each species second and third variables to enter the model, respec- could be assigned to one of the 3 groups (phytal, water tively (Table 2A). column, sediment) and then compared to the original Two discriminant functions were generated using grouping based on our arbitrary assessment of habitat the 3 variables identified from above (Table 2B). The utilization. first function accounted for 73.3 % and the second To examine the accuracy of our classification of function accounted for 26.7 % of the between-group copepods from Tampa Bay seagrass sites, we applied variability, respectively. The coefficients for CA, AIL the same discriminant analyses to copepods selected and PIR used in Functions 1 and 2 to calculate discri- from the literature for whom group membership or minant scores are presented in Table 2. We a pnon 62 Mar. Ecol. Prog. Ser. 35: 59-64. 1987

correctly classified 75 % of Tampa Bay seagrass utihzation as defined in the studes and that identified species into habitat groups based upon the calculated from morphological traits and our discriminant func- functions. Five species were 'misclassified' by our tion. Only phytal or sediment-dwelling copepods were quantitative method (i.e. had a higher probability of obtained from the literature. Four species were fittmg into another category) and their classifications categorised into a group by the discriminant scores according to the discriminant function are presented in that did not match the category assigned previously in Table 3A. the Literature. The copepods misclassified from the literature are presented in Table 3B along with their calculated group assignment. A combined plot of the Table 3. Specles misclassified by discr~mlnantfunctions from discriminant scores and group centroids for the 2 func- Tampa Bay data set (A) and literature cases (B). A prion classification is from Table 1 and discriminant classification is tions for each species from both the Tampa Bay and based upon discriminant scores. Abbreviations as in Table 1 literature data sets is presented in Fig. 2.

Species A pnori Dlscnminant classification classification DISCUSSION

A. Tampa Bay Overall, the discriminant analysis provided good Mesochra pygmaea P M separation of copepods into groups based upon 3 mor- Zausodes arenicolus M P phological traits: the ratio of the length of PI Endopod Paradactylopodia brevicornis M P Ectinosoma melaniceps M S 1 to P, Endopod 2 plus 3 (PIR),the length of the Lourinia armata S M antennule (AIL),and the area of the cephalosome (CA) (Tables 2 & 3, Fig. 2).Such multivanate analyses more B. Literature precisely summarize the qualitative observations that Tisbe furcata P S copepods living in different habitats have different Amonardia normani P M Huntemannla ladensis S P forms (Hicks & Coull 1983) and Illustrate that mor- Ph yllopodopsyllus bradya S P phological resemblance provides a reasonable index to habitat utilization or movement. Because many of the species or genera used in our analyses are often abun- Applylng the discriminant function derived from dant or represented in other studies on harpacticoid Tampa Bay copepods to harpacticoids from the litera- copepods (Por 1964, &to 1977, Coull et al. 1979, ture resulted in good correspondence between habitat Pallares 1979), we would expect our results to provide

Fig. 2. Plot of species by group for discriminant Functions 1 and 2. 1: Phytal; 2: M~grator;3: Sediment. 2 4 Group centroids for phytal (+), migrators (++) and sedment FUNCTION 1 (+ ++) forms are dlso provided Bell et al.: 'Habitat utilization by harpacticoid copepods 63

insight into the habitat utilization of many world-wide by its discriminant score. Although this species does copepod taxa. not inhabit seagrasses it has been reported in high The best morphological characteristic upon which to abundances living epibenthically on scallop shells (K. segregate species was PIR, although both CA and AIL Sherman pers. comm.), and coralline algae (Hicks were also important. The P1 character state deserves 1977), presumably among the entrapped sediments, particular attention. As has been repeatedly noticed for and occasionally enters water column traps in Tampa copepods, those that are phytal associates generally Bay (Walters & Bell 1986). Ectinosoma melaniceps is possess a modified P, which is strongly prehensile (P, clearly itinerant, being found in phytal, sediment and Endopod 1 > P1 Endopod 2 plus 3), modified presum- water column samples. These findings suggest that ably for grasping the substrate. Our results confirm there is not a stnct separation among habitats by some that phytal forms have a strongly prehensile P, and species i.e. some migrators divide their time between a migrators have moderately prehensile P, endopods substratum and the water column. The results for while sediment forms rarely have a first segment Zausodes arenicolus are enigmatic and not easily exceeding the length of the remaining segments. The explained since this species with 'phytal-like' mor- exact relation between AIL and CA and copepod utili- phology is abundant in sediments (Kern et al. 1984) zation of habitat is still unknown, although Marcotte and is a dominant migrator (Service 1986). In this last (1983) reports that an enlarged cephalosome, charac- case, behavior/habitat umzation do not follow the teristic of migrators and phytal forms, serves as an aid same relation as other copepods. to swimming. Interestingly, only one species from the literature Some of the cases of copepod misclassification from was placed into the category of migrator based upon the literature can be explained by inspection of the P1 morphology. No species from the literature was desig- arrangement. Specifically, while Huntemannia jaden- nated a prion as a migrator in published studies. In sis has been reported as a sediment form (Eckman contrast, at least 5 species were collected in the water 1983), it was assigned to the phytal category by the column in Tampa Bay. This migratory group of species discriminant scores. This is a result of the P, endopod was morphologically discernible. Thus, the noted having only 1 segment which when expressed as a active movement of the Tampa Bay forms in contrast to ratio, makes it appear infinitely 'prehensile'. Pro- the passive resuspension suggested for some sediment portionately, the P, endopod is comparable to that of forms (e.g. Eckman 1983, Palmer & Gust 1985) is other burrowing, sediment dwellers. Likewise, Tisbe reaffirmed by morphological evidence. The consistent furcata has a P, endopod that is elongated for both the agreement of the discriminant analysis to previous first and second segment. Such a relation of approxi- literature descriptions (above) and the world-wide dis- mately equal segment sizes within the PI is typical of tribution of copepod species similar to those investi- sediment-dwelling forms and is the reason T. furcata is gated here suggests that our findings may be useful to classified as a sediment species when it clearly utilizes predict which copepods should (1) be associated with phytal or water column habitats. Direct measurement vegetation, (2) display active migration and/or (3) live of the relative length of each segment of the P1 among sediments and be closely Linked to flow endopod, if available, could avoid these problems. regimes and sediment entrainment. Because intersti- Amonardia normani was the only species from the tial copepods were not present in the original data set literature that was classified as a migrator based upon we have not included them in our model, however they morphology. A. normani has been reported from algae, would serve as an additional test for the discrimination and Caste1 & Lasserre (1977) reported the species from power of the morphological traits used in our model. both sediment and submerged plants in lagoons. No Studies on harpacticoid copepod morphology are extensive distributional data are available on common, with some addressing the interesting ques- Phyllopodysyllus bradya to evaluate its classification. tion of morphological variation within a species (e.g. Four of the 5 species misclassified in the Tampa Bay Cod & Fleeger 1977) or family (e.g. Montagna 1982). data set represent those species which were often col- Additional morphological investigations of copepod lected in more than one habitat. For example, functional groups comprising many species have also Mesochra pygmaea was often captured in the water been reported (Marcotte 1977, Thistle 1982) but column although it was periodically very abundant on categories have not been quantitatively denved. Our seagrass blades. Paradactylopodia brevicornis was approach here has differed in that we chose to examine common on plant structure as well as in the water a broad range of copepod species and by using quan- column, but also dominated sediments at other sites titative analyses on a combination of traits, denved a within Tampa Bay seasonally (e.g. Kern & Bell 1984). predictive model. Furthermore, Lourinja armata was classified as a sediment dweller in Tampa Bay but designated a migrator 64 Mar Ecol. Prog. Ser. 35: 59-64, 1987

Acknowledgements. The work was supported in part by Hicks, G. R. F., Coull, B. C. (1983). The ecology of marine grants from thc National Science Foundation from the Biolog- meiobenthic harpacticoid copepods. Oceanogr, mar. Biol. ical Oceanography Program (OCE 80017261A) and Interna- A. Rev. 21: 67-175 honal Programs (84-NZ-02) (S. S. Bell, Principal Investigator). Kern, J. C., Bell, S. S. (1984). Short-term temporal variation in We thank G. R. F. Hicks and anonymous reviewers for their population structure of two harpacticoid copepods, insightful comments. Zausodes arenicolus Wllson and Paradactvlo~odia, . brevicornis (Claus). Mar. Blol. 84: 53-63 Kern, J. C., Edward, N. A., Bell. S. S. (1984). Precocious clasping of early copepodite stages: a common occurrence In Zausodes arenicolus Wilson (Copepoda: Harpac- LITERATURE CITED ticoida). J. Crust. Biol. 4: 261-265 Kito, K. (1977). Phytal animals in the Sargassum confusum Bell, S. S., Walters, K., Kern, J. C. (1984). Meiofauna from region in Oshoro Bay, Hokkaido: phenology of harpac- seagrass habitats: a review and prospectus for future ticoid copepods J. Fac. Scl. Hokkaldo Univ. V1 Zool. 20: research. Estuaries 7: 331-338 691-696 Castel, J., Lasserre, P. (1977). Colonisation et &stribution Lang, K. (1948). Monographie der Harpacticiden. 0. Koeltz spatiale des copepodes dans les lagunes semi-arbficielles. Science, Konigstein In: Keegan, B. F., Cedgh, P. O., Boaden, P. J. S. (ed.) Marcotte, B. M. (1977). The ecology of meiobenthc harpac- Biology of benthic organisms. Pergamon Press, New York, ticoids (Crustacea: Copepoda) in West Lawrenceton, Nova p. 129-146 Scotia. Ph. D. dissertation, Dalhousie University, Halifax Chandler. T. G., Fleeger, J. W. (1983). Meiofaunal coloniza- Marcotte, B. M. (1983). The imperatives of copepod diversity: tion of azoic sediment in Louisiana: mechanisms of disper- perception, cognition, competition and predation. In: sal. J. exp. mar. Biol. Ecol. 69. 175-188 Schram, F. R. (ed.) Crustacean phylogeny. A. A. Balkema Coull, B. C., Bell, S. S. (1979). Perspectives of Maine Publishers, Rotterdam, p. 47-72 meiofauna ecology. In: Livingston, R. J. (ed.) Ecological Montagna, P. A. (1982). Morphological adaptation in the processes in coastal and marine systems. Plenum Press, deep-sea benthic harpacticoid copepod family Cer- New York, p. 189-216 viniidae. Crustaceana 42: 37-43 Coull, B. C., Bell, S. S., Savory, A. M,, Dudley, B. W. (1979). Nie, N. H. (1983). SPSSX user guide. McGraw-Hill, New York Zonation of meiobenthic copepods in a southeastern Noodt, W. (1971). Ecology of the Copepoda. In: Hulu~gs,N. C. United States salt marsh. Estuar. coast. mar. Sci. 9: (ed.) Proceedings of the first international conference on 181-188 meiofauna. Smithson. Contr. Zool. 76: 97-102 Coull, B. C., Fleeger, J. W. (1977). A new species of Pseudo- Norusis, M. J. (1985). SPSSX advanced statistics gwde. stenhelia, and rnorpholo~cal variation in Nannopus McGraw-fill, New York palustns (Copepoda, Harpactlcoida). Trans. Am. microsc. Pallares, R. E. (1979). Copepods harpacticoides marinos de SOC.96: 332-340 Tierra Del Fuego (Argenhna), Isla de Los Estados 111. Decho, A. W., Hummon, W. D., Fleeger, J. W. (1985). Contribuci6n Cientifica #l42 del Centro de lnvestigacion Meiofauna-sediment interactions around subtropical sea- de Biologia Marina grass sedirnents using factor analysis. J. mar. Res. 43: Palmer, M. A, Gust, G. (1985). Dispersal of meiofauna in a 237-255 turbulent tidal creek. J. mar. Res. 43: 174-210 Eckrnan, J. E. (1983). Hydrodynamic processes affecting Por, F. D. (1964). A study of the Levantine and Pontic Harpac- benthic recruitment. Lmnol. Oceanogr 28: 241-257 ticoida (Crustacea, Copepoda). Zool. Verh., Leiden 64: H~cks,G. R. F. (1976). Neopeltopsis pectinipes, a new genus 1-128 and species of seaweed-dwelling copepod (Harpac- Por, F. D. (1984). Canuellidae Lang (Harpacticoida, Poly- ticoida: Peltidiidae) from Wellington, New Zealand. N. 2. arthra) and the ancestry of the Copepoda. Crustaceana 43: Jl mar. Freshwat Res. 10: 363-370 1-24 H~cks,G. R. F. (1977). Species associations and seasonal Service, S. K. (1986). ~Manipulahve studies of the effects of population densities of marine phytal harpacticoid harpacticoid copepod density on dispersal. M. S. thesis, copepods from Cook Strait. N. 2. J1 mar. Freshwat. Res. 11 Univ, of South Florida, Tampa 621-643 Thlstle, D. (1982). Aspects of the natural history of harpac- Hicks, G. R. F. (1982). Porcellidiidae and Peltidiidae ticoid copepods of San Diego Trough. Biol. Occanogr 1. (Copepoda: Harpacticoida) from the marine algae of St. 225-238 Croix Island, Alyoa Bay, South Africa. J. Linn. Soc. 75: Walters, K.,Bell, S. S. (1986). Dlel patterns of active vertical 49-90 migration in seagrass meiofauna. Mar. Ecol. Prog. Ser. 34: Hicks, G. R. F. (1985).Meiofauna associated with rocky shore 95-103 algae. In: Moore, P. G., Sneed, R. (ed ) The ecology of Williams, B. K. (1983).Some observations on the use of discn- rocky coasts. Hodder and Stoughton, London, p. 3&56 minant analysis in ecology. Ecology 64: 1283-1291

This artlcle was submitted to the editor; it was accepted for printing on October 21, 1986