Marine Biology ( 1994) 120:467-478 © Springer-Ve rl ag 1994
D. M. Anderson · D. M. Kulis · G. J. Doucette J. C. Gall agher · E. Balech Biogeography of toxic dinoflagellates in the genus Alexandriu~ #l from the northeastern United States and Canada
Received: II April 1994 I Accepted: 20 June 1994
Abstract Twenty-eight strains of toxic dinoflagell ates in widespread, homogeneous populati on, but rather is com the genus Alexandrium from the northeastern Uni ted States prised of several sub-populati ons, each with its own phys and Canada were characterized on the basis of morphol iological characteri tics and hi story. Two ce narios are ogy, biolu minesce nce capacity, mating compatibil ity, and considered with respect to this regiona l biogeography. The toxin composition. T he distributions of these characters first invokes recent and continuin g dispersal of isolates to were evaluated in the context of regional patterns of para the south from a center of origin in the north, fo ll owed by lytic hell fi sh poisoning (PSP) and coastal hydrography. recombination and strong selection. The second holds that Two morphospecies were identified - A. tamarense Lebour the northern and southern popul ati ons di verged from a and A. j imdyense Balech. The two are interspersed geo common ancestor (vicari ance), but now repre ent localized graphicall y though there are areas, such as the Gulf of population with little mixing of genotypes. Neither hy Maine, where apparently only A.f undyense occurs. South pothesis can be comp lete ly refuted by the data presented ern waters (Cape Cod, Connecticut, and Long Island) have here, though the weight of the evidence favors the latter. especiall y diverse populations. The two species are sexu The correct cenario may be a combination of both, with ally compatible. Virtua ll y all northern i olates are biolu recent and continuous spreading occuring within the G ulf minescent, whereas southern isolates inc lude biolumines of Mai ne and perphaps the Gulf of St. Lawrence, but with cent and non-bioluminescent strains. Cluster analyses, endemic localized populations persisting without geneti c based on high performance li quid chromatography (HPLC) exchange in most outhern locations. T hese data also in determinati ons of the sui te of toxins produced by each iso dicate that although morphological cri te ri a separate toxi c late, revealed two and perhaps three distinct groups. One Alexandrium isolates from the study region into two mor is comprised almost exc lu ively of northe rn strains, and phospecies, these as ignments do not coincide with c lu - the other of southern strains. A Cape Cod c lu ster may be terings based on toxin compo ition or a ll ozyme electro separable from the south ern group. These analy es explai n phoresis, and they are further violated by mating results. a previously reported north-to-south trend of decreasing A revision of taxon designations to the varietal level could Lox icity, a the northern isolates produce greater propor be justified. ti ons of the more potent toxins than do southern forms. The overall perspective is that the biogeography of toxic Alex andrium spp. in the study region is not that o f a single, Introduction Communicated by J. P. Grassle, New Brunswick Some species in the dinofl agell ate genus Alexandrium pro 1 D. M. Ande r so n ( ~)· D. M. Kulis· G. J. Doucette duce pote nt neurotoxins which are well known because Biology Department, Woods Hole Oceanographi c Institution, Woods Hole, Ma. sachusetts 02543. USA they cause paralyti c he ll fish poisoning (PSP), a syndrome of human illness and death foll owing the consumpti on of I 1. C. Gallagher Biology Department. City College of CUNY, contaminated shell fish. Alexandrium spp. toxins a lso move I 38th St. at Convent Ave .. New York, to other levels of the food chain. impacting zooplankton, New York I 003 I , USA fish, and e ven marine mammals. PSP is a global phe nom E. Balech enon that affects many coastal coun tries. Along the west Casilla de Correo 64. 7630 ecochea, Argentina ern Atlantic, it represents a real or potential threat from 1 New Je rsey in the United States north to the Sr. Lawrence Present address: National Marine Fisheries Service, Southeast Fisheries Science Center. C harleston Laboratory, P.O. Box 12607, estuary and Newfoundland in Canada (Fig. I A). She ll fish C harleston, South Carolina 29422, USA toxic ity ha been a problem for hundreds of years in east- 468 Lawrence (Cembella et al. 1988), as well as in scattered A Canada locations on ewfoundland (White and White 1985). Far to the south, toxic Alexandrium spp. blooms occur in iso so· lated embayments and estuari es in Cape Cod, Connecticut, Long Island, and New Jersey (Anderson et al. 1982; Schrey et al. 1984; Cohn et al. 1988) and PSP has recentl y been detected at hi gh levels in offshore shellfish on Georges ' Bank (White et a l. 1993). Between these southern and northern extremes, the Gulf of Maine (inc luding the mid dle and lower reaches of the Bay of Fundy) experiences United States PSP along virtually its entire coastline. Toxicity is local ized and infrequent along the eastern coast of Nova Scotia. There is considerable potential for genetic exchange among Alexandrium spp. popul ations fo r two reasons. First, the organi sms undergo a sexual phase resulting in the 70° so• production of re istant, overwintering cysts during the late stages of blooms (A nder on and Wa ll 1978). Furthermore , the areas affected by PSP are linked hydrographicall y B Canada (Smith and Schwing 199 1) and there is a general north-to south fl ow in the region (Fig. I B). orth of Cape Cod. 50' Alexandrium spp. blooms are geographically widespread phenomena, affecting hundreds of kil ometers of coastline (Cembe lla et al. 1988; Martin and W hite 1988; Franks and Anderson 1992), facilitating populati on dispersal over large distances. In contrast, southern blooms are locali zed within shallow e mbayments and estuaries (Anderson et al. United 1982; Schrey et al. 1984) . with little or no exchange with States local coastal waters. Little is known of the geneti c and ecological relati on shi ps that link different populations of Alexandrium spp. within this large region. Whether the 1972 outbreak was large enough to disper e these pecie as far south as Con 70° so· necticut or New Jersey in one spreading event is uncertain, nor is it known whether toxic Alexandrium specie were Fig. 1 T he study region. A Origins of the isolates examined in the present study. Strain names fol lowed by " F" or "T". designating Alex present in southern waters prior to that date , or whether andrium flmdyense and A. tamarense, respecti vely. B General pat there have been spreading events on a smaller scale since tern of non-tidal surface circulation in the study region (arrows). 1972. The first compari son of regional Alexandrium spp. Coastline areas affected by paralytic shellfish poisoning (PSP) are populations was a survey of the toxin content of isolates blac ke ned. Not shown: PSP toxicity also occurs in offshore shell fish within the central Gul f of Maine. on Georges Bank, and Nantucket originating between Nova Scotia and New York (Maranda Shoals (located j ust to the west of Georges Bank). Qualitati ve now et al. 1985). That study revealed a systematic decrease in patterns deri ved in part from Smith and Schwing ( 199 1) and re fer toxic ity from north to south, an o bservation that appears e nces therein inconsistent with the hypothesized recent dispersal of toxic cells from a common source during the 1972 bloom. Are search program was establ ished to compare isolates from ern Ca nada, but the western Gulf of Maine and areas fu r throughout the region using fi ve differe nt characters: mor the r so uth have only been affected during the last two phology. mating compatibility. bioluminescence capacity, decade. (P rakash et al. 197 1; A nderson e t al. 1982). A toxin composition. and all ozyme e lectrophoresis. Electro com mon belief is that the recent history of PSP in south phoretic results are given by Hayhome et al. ( 1989). Here ern waters reflects the di spersal of Alexandrium spp. from we present informati on on the other character and provide e ndemic populati ons in the north during a massive 1972 an overview o f the entire study. red tide, an event whi ch caused PS P toxicity for the fi rst time in southern New England (Hartwe ll 1975). She ll fi sh toxic ity is now a recurrent annual p roble m in the e previ Materials and methods ously una ffected area . PSP is not evenly distributed throughout the region (Fig. Isolation and culture I B). At the northern extreme, it is associated with she ll Twenty-eight clones of Alexandrium tamarense and A. frm dyense fish along the lower estuary of the St. Lawre nce Ri ver, in from waters between the Gulf of St. Lawrence. Canada and Long Is c luding the Gaspe peninsula a nd the western Gulf of Sl. land, NY were established in culwre (Table I). GT2 and GT7 were 469
Table I Alexam/rium ramarense and A. ftmdyense. List of i ~o l a t es and their characteristics. ordered from north-to-south. (a archived sam ple. Due to evaporatio n and chemical conversio ns during long-te rm ~ t orage . toxin contents not reported)
Clo ne Species Locatio n Toxi n Biolum
(fmol ce11 )1 (fgST X cell )
ATSLI2 A. wmarense Baie Comeau. St. Lawerance Estuary (49° 03'N. 68° 20'W) 187.9 58695 Yes AF FA4 A. jimdyense Harbour Grace. ew Foundland, (47° 30' . 53° I 0' W) 88.8 38075 Yes GT2 A. ftmdyense Campobello Island, Bay of Fundy (44° 46' N. 67° II'W a a Yes GT7 A. ftmdyense Campobello Is land. Bay of Fundy (44° 46' . 67° II ' W) 3 1.9 10927 Yes GTME05 A. ftmdyense Greenlaw Neck. Deer Isle. Maine (44° 13' N. 68° 40' W) a a Yes GTME20 A. f tmdyense Monhegan Island , Maine (43° 45'N, 69° 19'W) 50.7 17 507 Yes GTCA29 A. ftmdyense Gulf of Maine. N H (43° 00'N, 70° 19'W) 69.8 25627 Ye~ GTCA28 A. ftmdyense Gulf of Maine, NH (43° 00' N. 70° 19' W) 61 .5 2 1 372 Ye · GTCA08 A. ftmdyense Gulf of Maine. NH (43° 00'N. 70° 19'W) 53.8 191 95 Yes GTCA04 A. ftmdyense Gulf of Maine. NH (43° 00' N. 70° 19'W) 67.3 26894 Yes GTCAOI A. ftmdyense Gulf ofMaine. H (--B 0 00' . 70° 19' W) a a 0 GT429 A. f undyense Ipswich Bay. MA (42° 40' N. 70° 45' W) 34.5 11435 Yes GTSPI A. ramarense Salt Pond, Eastham, MA (4 1° 50' . 69° 58' W) a a Yes GTMP14 A. ramarense Mill Pond. Orleans. MA (41 ° 47' N. 69° 57' \V) 74.0 19071 0 GTMP02 A. ramarense Mill Pond. Orleans. MA (41 ° 47' , 69° 57'W) a a 0 GTMMPI OI A. Jundyense Mill Pond. Orleans. MA (41 ° 47' t . 69° 57' W) a a 1 0 GTMMP103 A. ftmdyense Mill Pond. Orleans. MA (41 ° 47' N. 69° 57' \V) a a Ye. GT MROI A. Jundyense Mitchell Ri,er. Chatham, MA (41 40'N, 69° 58' W) 11 8.3 29859 Yes AFNS85 A. fundyense antucket Shoals. MA (41 ° 36' . 69° 45'W) 13 1.5 53698 Yes GTPPOI A. ramarense Perch Po nd. Falmouth, MA (41 ° 30' . 70° 35' W) 29.3 781 4 No GTPP02 A. wmarense Perch Pond. Falmouth. MA (41 ° 30'N. 70° 35'W) a a /A GTPPIO A. ramarense Perch Pond , Falmo uth. MA (4 1° 30' N. 70° 35'W) 72.3 15894 0 GTC 02 A. fundyeme Palmer Cover. Groton, C (41 ° 20' . 72° 00' W) 127.6 -1 6 553 0 GTCNI 5 A. ftmdyen\·e Mumford Cove, Groton. CN (41 ° 20' . 72°0l'W) 48.0 11621 No GTC 16 A. tamarense Mumford Cove, Groto n, C (41 ° 20' N. 72° 0l ' W) 51. 3 7772 No GTLI2 1 A. tamarense Mud Creek, Moric hes Bay, Y (40° 45' , 72°49' W) 25.5 6199 Yes GTLI22 A. ramarense Mud Creek. Babylo n. NY (40° 40'N, 73° 20'W) a a 0 GTLI23 A. f wrdyense Mud Creek. Babylon. NY (40° 40' . 73° 20' W) a a N/A
obta ined from A. W. White. and GT429 was isolated in 1972 by C. bated under normal culture conditions. The germinated planomeio Martin. All others were i ~o l a t ed between 1978- 1991 by Ande rson cy te~ were allo wed to divide twice in the we ll ~ . and the resulting four and co-wo rk e r ~. Most c ultures (those termed ··non-archived'') were cells were individually isolated and placed into four eparate culture grown in duplicate 25-ml volumes off/2 medium (Guillard and Ryth tubes. In orne cases. all fo ur cultures survived, represent ing a tet er 1962) made with 0.2-!.lm filtered seawater [3 1 psu (practi cal sa rad. In other cases when o ne o r two cultures did not survive. incom linity units)J. The f/2 medium was modified by addin& H2Se0 3 ple te tetrads were obtained. In o ne instance. division proceeded too and reduc ing C uS0 4 , both to a fi nal concentratio n of I o- M . C ul quic kly and s ix cultures were established from o ne cyst. Clearly. two tures were incubated at 20°C o n a 14 h light: IO h dark cycle (ca. of these cultures are identica l to two others. T he presence or absence 2 1 250 1.1E m- s irradiance. cool white flu o rescent bulbs). Growth was of the ventral pore in cells from each culture was a~ce rt a in ed u ·ing monito red o n a Turner mode l I 0 fluorometer. When cultures reached Calcofluo r staining (Fritz and Triemer 1985) and microscopic obser mid-expone ntial stage. duplicate samples were removed for ce ll vation of at least 40 thecae. Mating type was determined by back counh and the remaining cells harvested by ccmrifugation at 1700xg cros!>ing these first fi lial generation (F 1) cultures with the parents. fo r 3 min at 23 °C. The supernatam was aspirated and the cell pellet Several F 1 cultures were a l o c rossed with each other and induced containing approximate ly 1.5 x I 05 cells was resuspe nded in 0.5 ml to form cysts, which were then induced to germinate. and the prog 0.05 M acetic acid and sonified with a Branson sonifi er for 3 min at eny c ultured to demonstrate the viability of the second fi lial (F2) gen a selling of 6 A in an ice water bath. Samples were frozen at - 20 °C eratio n. until analysis. then thawed. mi xed, recentrifuged as above and fil tered through 0.45-).lm syringe filters. Some isolates were lost from the culture collecti on and could onl y be analyzed using stored sam Mating compatibility ples prepared for a previo us swdy. The e archi ved samples had been grown and extracted as described by Hay home et al. ( 1989). A ~ ub Mating compatibility was assessed by inoculating individual isolates sample was processed fo r toxin analysis and the extracts frozen and alo ne, or in combination with another i olate, into f/2 medium with thawed three times prior to analysis to re lea e the toxins. nitrogen supplied o nl y as NH; at 25 ).IM. Initial concentrations were I 00 cell. ml- 1 of each srrai n. Crosses were incubated under the tem perature and lighting condition descri bed abo\e and c hecked for Pore morphology resting cyst formatio n after I mo.
Cultures grown under standard conditions were fi xed in 4% formal in and u ed fo r taxonomic analyses. Species designations were as Bioluminescence signed according to the criteri a of Balech ( 1985). The morphology of celb from selected cultures produced by mating Alexa11driw11 ram C ultures were gro wn under standard conditio n and tested for bio arel/se and A. f undyense were also examined to determine if a ven luminescence in a darkened room by the additio n of several mls of tral pore was present in the progeny. Cysts were isolated and placed 0.5 M acetic acid to a 25-ml culture in mid-expo nential growth . Neg individually into tis<, uc culltlre wells containing medium and incu- ative observatio ns were verified at a later date. 470
1 Toxin analysis pressed as total molar concentrations and as STX equivalents ceW • Observed differences in toxin composition of non-archived and ar Non-archived and archi ved samples were analyzed by the three-step chived samples of rhe same isolates led to a concern that inclusion isocratic elution method of Oshima et al. ( 1989) for high perfomance of the archived samples might systematically bias the analyses. liquid chromatography (HPLC) determination of the saxitoxins, with Therefore, each analysis was performed using the combined (ar some modifications. For the C I-C4 toxins, col umn temperature wa chived plus non-archived) data, and the non-archi ved data alone, to 30 oc and the post co lumn reaction temperature 45 oc to enhance determine if cluster patterns would be affected. T he loading of each separation and sens itivi ty. T he mobile phase was adjusted to pH 5.8 variable on the principal components is available from the authors with 0.05 M acetic acid. Gonyautoxins were analyzed wi th a column on request. temperature of23 °C and a post column reaction temperature of35 °C to enhance fluorescence. Column and post column reaction temper atures were 24 and 50 °C, respectively, for the saxitoxin group in or der to increase neosaxiroxin flurorescence. The column eluate was Results mixed with 7 mM periodic acid in 80 mM sodium phosphate buffer. The HPLC system consisted of a Waters 600E multi-solvent deliv ery system connected to a Waters Wisp 700 autosampler and a Shi Morphology madzu RF-535 fluorescence detector. A Beckmann C8 UILrasphere Octyl 55m analytical colu mn (25 x 46 mm) and a Brownlee MPLC All clones examined belong to the ''tamarensis/catenella" New Guard Column were used fo r the separation of the toxins. Ex ternal standard soluti ons were run prior to sample analysis and after group, with two species represented, i.e. Alexandriwn tam every fourth sample. Toxin composition profiles were determined arense Lebour and A. fundyense Balech. There were ten of from replicate analyses (two injections) from each of two separate the former species and 16 of the latter (Table I ). The dis samples. Abbreviati ons used throughout this text are: STX - saxi tribution of the two species across the study region was not toxin: NEO - neo ax itoxin: GTX I ,4- gonyautoxins I and 4; GTX I I I 2,3 - gonyau10x ins 2 and 3; GTX 5 - gonyautoxin 5 (or 8 I; Hall uniform (Fig. A), as of the 12 northernmost isolate et al. 1990); GTX 6 - gonyautoxin 6 (or 8 2); Cl.2 -toxins Cl and were A. jimdyense, the one exception being an A. tama C2; C3, C4- toxins C3 and C4; de- decarbamoy I. The terms GTX 1.4, rense isolate from the St. Lawrence estuary. All northern GTX2,3. C I ,C2 and C3,C4 are used to represent the pooled concen Gulf of Majne isolates were A.fundyense. In the outh, the trati ons of two toxins to account for possible epimerization. Toxic 1 two species were interspersed. ities (in STX equivalents cell- ) were calculated from the molar com position data using individual potencies (Oshima et al. 1992). Spe cific toxicities of the individual toxins (pg STX eq.) are as follows: C l = 15: C2=239: C3=32: C4= 143; GTXl =2468: GTX2 =896; Mating compatibi li ty GTX3 = 1286; GTX4= 1803:GTX5= 160; dcGTX2= 1617; clcGT X3 = 1872; NE0=2295; dcSTX= 1274; STX=2483. Non-archived toxin extracts were analyzed within 3 mo of ex None o f the seven Alexandriumfundyense and three A. tam traction, whereas archi ved materials had been stored fo r several years arense isolates cul tu red by themselves in low nutri ent me at 4 °C in 0.05 N aceti c acid. One concern was that prolonged stor dium formed resting cysts (hypnozygotes), but numerous age could alter the toxin compos ition: another was that evaporation cysts were formed from both inter and intra-specie could cause the calculated total tox in concentrations to increase. In some analyses described below [e.g. clustering and principal com crosses. Of the 16 crosses or A. fundyense isolates with ponent analyses (PCA)], archived and non-archived data were com other A. f undyense (not inc luding self-crosses), nine pro bined. For others, only non-archived samples were used. Justifica duced cysts (data not shown). Of 17 crosses ofA .fundyense tion for these decisions is given in the text. with A. tamarense, eight produced cysts. After a suitable maturation interval, cysts obtained from one A. fun dyense!A. tamarense cross (GTSP 1 and GT7) were germi Data analysis nated to produce typical planomeiocytes (von Stosch 1973; Toxin composition data were examined us ing multivariate matrix Anderson and Wall 1978) which divided to produce vege cluster analysis employing Pearson correlation coeffi cient s and the tative cell s that were i olated and cultured as tetrads. average linkage method (program CLUSTER, option JOI MA Crosses of these tetrad cultures yielded cysts 50% of the TRIX. in the SYSTAT package. Systat, Inc., Evanston, lL). This is time in a pattern consistent with the presence of two mat a hierarchial clustering procedure in wh ich rows and columns are permuted tO maximize indentificati on of clusters, and the cuto ff in g types in thi s heterothallic organism. poi nts bet ween ranges of the correlation coeFficients are determined Complete and incomplete tetrad culture produced by by Tukey's gapping method. This multivari ate clu tering method is this mating between Afexandriu.m tamarense and A. fun more appropriate than the more common uni variate procedures in dye/ue were examined to determine mating type, whether that it simultaneously yields information on the relationshi ps among the cell s were bioluminescent, and if a ventral pore was the isolates and among the toxins. This method was applied to tox in data transformed with the weighted-ratio method ofCembella et a l. present on the theca (Table 2). Both parents were biolumi ( 1987), where the maximum mole percent (mol%) of a given toxin nescent, and one (GTSP I, A. tamarense) had a prominent across isolates was assigned a value of I. and the relati ve mol % of vemral pore but the other (GT7, A. f undyense) did not. Of the toxin in other isolates was scaled relative to this maxi mum val the 14 tetrad progeny assayed for bioluminescence, 13 ue. T hi s treatment weight each toxin componem equally. avoiding bia against toxins present in low concentrations in all isolates. In were positi ve and one was negative. Matin g types were order to correct for the role of evaporation and possible cpimeriza equally di vided and howed no relationship to biol umines tion in archived samples, the values for the non-archived prepara cence or to the presence of a pore. Only one of the 2 1 prog ti ons were scaled against themselves whereas the archived sample e ny cultures contained cells with a prominent ventral pore. were scaled against the hi ghest archived samples. Toxin composition was also exami ned us ing PCA on the corre All others either had no pore what oever or had a very lation matrix (program FACTOR, option VA RIM AX. ElGE =0. small , generall y indistinct pore on a few (maximum 8%) in the SYSAT package). This method was used with toxicity ex- of the cell s. [Close examin ation of the clonal GT7 parent 471 Table 2 Alexandriw11 spp. Characteristics of progeny from mating ber and quantity of individual toxins that are pre ent in an between A. ramarense (GTSP /) and A. fimdyense (GT7). Isolate extract, typicall y determined using HPLC. grouped 10gether are from the same cyst (Biolum. biolumi nescent ; NIA data not available) During this extended study, a number o f Afexandrium i o lates were lost. For many, pre erved amples and toxin Isolate Biolum. Mating Ventral Comments extract were available, but it wa not clear whether epi type pore meri zation and evaporati on during storage would intro Parents duce errors into their analy es. HPLC analy is of isolates GT7 Yes + 3 of 59 cells, small pore which were still li vin g, but also represented in the archived GTSPI Yes + 40 of 40 cells, normal pore samples, demonstrated that storage of extracts in 0.05 N Progeny aceti c acid at - 20 °C caused change in the relati ve abun 17C8A Yes + + 38 of 40 cells, normal pore dance of each component of the epimer pairs GTX l ,4, 17C8B N/A 3 of 40 cells, small pore 17C8C Yes GTX2,3, C I ,C2, and C3,C4. This effect, previously noted 17C80 Yes + I of 40 cells. sma ll pore by Hall et al. ( 1990), does not alter the percent composi 1705A 3 Yes tion o f the toxin s if the concentrati on of the epimer pairs 170 58" N/A are combined (i.e., as GTX I ,4 rather than a individua l tox 170 sc• Yes + I of 40 cell s. small pore 170 50 " Yes + I of 40 cell s, small pore ins GTX I and GTX4). However, whe n toxin content is ca l 17 0 5E" Ye + 3 of 40 cell s, small pore cul ated from the concentrati ons of the individual toxins in 170 5F" Yes + 2 of 40 cells, small pore each i olate, epimeric conversions during long-term stor 17C3A Ye N/A age do introduce error because of the different potency of 17C3B /A /A 17C3C Yes N/A the epimers (Oshi ma et al. 1989). Furthermore. evapora 17C4A Yc /A ti on during torage can reduce the volume of the archived 17C4B N/A N/A sample , increasing the calculated molar concentration of 17C4C Yes /A the toxin and the ir overall toxicity. Analyses of non-ar 17C5A N/A N/A chi ved and archi ved extracts demon trated that evapora 17C5B /A /A 17C5C N/A N/A tion had occurred in the stored samples (data not shown). 17C IA o N/A N/A When five isolates were examined usin g both non-ar 17CIC Yes N/A N/A chi ved and archived samples in the cl uster analysis, the two "duplicate" samples always grouped in the same clu - " These cultures represent the fi rst ix cell s formed after a single cyst germinatio n. Ordi naril y. four are isolated, but an additional di vision ter with the same major toxin signature (Figs. 2 and 3). In occurred. T herefore, two of the isolates are identical to two others two case . the dupl icates were each other's closest ne igh (probabl y C,O =E.F) bors. The relative distances between archi ved and non-ar chi ved duplicates is indicated by the da hed lines in Fig. 3. Clustering analyses run on the non-archi ved samples alone revea led two main clusters wi th ATS L 12 being an outlier culture revealed that 3 of 59 thecae examined (5%) had a (data not hown), just as was the case when non-archi ved mall ventral pore.] and archived data were pooled for the analy e . PCA anal y e of only non-archi ved ample also produced s imilar patte rns to the combined analy i . Archived samples were Biolumine cence therefore inc luded in analyse that are based on relative toxin composition (i .e .. clusterin g and PCA analyses). but Of 26 culture from the regional survey examined for bi were excluded from all that rely on ab olute toxin concen oluminescence capacity, 18 were positi ve (Table I). Four trati ons (e.g. toxin content comparisons) due to the effects were Alexandriun1 tarnarense and 14 were A. f undyense. of evaporation. Five of the e ight non-bioluminesce nt isolates were A. tam The toxin composition o f Alexandriurn isolates varied arense. Al l but one of the northern isolates were biolumi dramaticall y over the region. Detail ed, tabular in formation nescent, but tho e from Cape Cod and areas to the south for all isolates is avai Ia ble fro m the senior author on re were mi xed. quest. Toxins C l and C2 were present in most isolates. though proportions ranged broadly from 0. 1 to 67.5% of the total toxin concentration . With a few exception , Toxin compo iti on outhern i olates had high concentrati on of C I ,C2 com pared to northern isolates. Toxins C3,C4 were detected in Individua l Alexandrium isolates can produce a suite of neu onl y ten isolates, all of which were either from Cape Cod rotoxins in the family of compounds collecti vely call ed the or Long Island; concemrati ons were always low, usuall y axitoxins. They include the parent compound, saxitoxin , < I% of the total. GTX I ,4 were pre ent in all i olates, rang and 17 deri vati ves (Hall et al. 1990) whose potency can ing from 5 to 46% of the total; no geographic patterns were vary by nearly two orders of magnitude. The term "toxin evident. GTX5 was found in highe t relative abundance in conte nt" re fer to the overall toxic ity of a cell, the inte Long Island. Connecticut and Cape Cod isolates ( I 0 to grated potency of all toxins present (ex pre sed as )..l.g STX 30%), and in trace amounts ( I%) in three nothern i olates. 1 equi valents ceW ) . "Toxin composition" refers to the num- GTX2,3 varied from 0.5 to 49%, with highest concentra- 472 tions typically in northern isolates. The pattern s for NEO 3 and for STX were simi lar, with proport ions ranging from A 0 to 40 and 0 to 29%, respectively, and the highest concen 2 trations in northern isolates. The decarbamoyl toxins dcGTX2,3 were present in very low concentrations in most of the isolates; dcSTX was found in three- one from the C\J St. Lawrence, and two from Cape Cod. 2 0 -1 -2 3 B 2 · AFNS85 F y NTS C\J GTCN02 F N CNL u5 GT7-A F y GM The first is comprised of isolates with strong signatures for 100000 NEO, STX, and GTX2,3. With one exception (GTCN02), the isolates in th is cluster originate in the Gulf of Maine ., 80000 or further north. All 17 members of the first c luste r are "....>< Vl Alexandrium fundyense and all but three are bi olu mines 60000 2 cent. Those three [AFNS85, GTCAO I, and GTCN02. iso ~ ..J t ..J lated from Nantucket Shoals (offshore of Cape Cod), ... 40000 ,~ u rt southwestern Gulf of Maine, and Palmer Cove, Connecti ...... z ! t ~ ~ ~~~~ ~~ ~~~ cut, respectively] share the same toxi n profiles as the north x ~ ~ 0 20000 ern group and also have A. fundyense morphology...... ~ .· . The second main cluster is comprised of isolates wi th fit•• • • strong signatures for C I ,C2. dcGTX2,3, and GTX I ,4. 40 46 60 These isolates a ll originated in Connecticut, Long Island, LATI TUDE and Cape Cod. Both Alexandriumfundyense and A. tama Fig. 4 Alexandrium spp. Toxin content of Alexandrium isolates as rense morphologies are present, with A. tamarense being a function of latitude of origin. Toxicity determined from high per more common. Most, but not all , are bioluminescent. Cape formance liquid chromatography toxin composition ana lysis, with Cod isolates also have strong signatures for C3,C4. and it each indi vidual toxin concentration converted to saxitoxin equiva is possible that these might belong to a third c lu ter that lents (STX eq) using the conversion. listed in the text. Dashed line represents linear regression. correlation coefficient r= 0.49. All sam could be berter resolved if larger sample sizes were avail ples analyzed in replicate: error bars denote I SD able. One isolate (ATSL12) had a unique profile containing both NEO and C l ,C2 in relatively high proportions: it con sequently did not fall in to e ither main cluster. Thi isolate had Alexandrium tamarense morphology, was biolumines toxicity, averaging 69.6 and 74.2 fmol cell- 1 (or 25 7 18 and 1 cent. and originated in the St. Lawrence estuary. 20 758 fg STX ceW ), respecti vely. When toxin content is PCA was performed on the toxin composition data ex plotted versus the latitude of the isolates' ori gins, a slight pressed as the molar content of each toxin and as STX increase in toxicity from south-to-north is sugge ted equi valents (Figs. 3A,B). In the former analysis, the first (Fig. 4). A linear regres ion of toxin content (expres ed as two principal components account for 62.3% of the vari molar concentration) versus latitude was not significant at ance, and the first three account fo r 75.3%. With toxicity the 95% confidence level (data not shown). A regression expressed as STX equivalents, the first two components expressing toxin content as STX equivalents ceW1 (Fig. 4) account for 66% of the variance, with the third accounting was significant (? <0.05). for an additional 13%. For both analyses, the highest cor relations (positi ve or negative) for the first princ ipal com ponent were STX, C3,C4, GTX2,3, and GTX5. Forthe sec ond axis, they were C I ,C2, and dcGTX2,3. The plots for Discussion the scores on the first two axes are very similar to there sults of the clustering analysis (Fig. 2) and do not support Genetic linkages between populations of closely related an hypothesis of a simple north -to-south progression of toxic d ino fl agell ates can be evaluated u ing several diffe r traits. One isolate from Connecticut (GTCN02) and one ent characters. In this, and in a previous study (Hayhome from Newfoundland (AFNFA4) appear to be very similar et al. 1989), toxic dinoflagellates of the genus Alexandrium to the Gulf of Maine isolates. The isolates from the St. Law from the northeastern United States and Canada have been rence (ATSLI2) and from Nantucket Shoals (AFNS85) are characterized on the basis of morphology, bioluminescence outliers. The isolates from Connecticut and Long Island capacity, mating compatibility, allozyme banding pattern s, have toxin signatures that are intermediate between the and toxin composition. Of these, toxin composition pro northern group and those from Cape Cod. vides the most usefu I insights into the genetic relationships between geographically separated populations. The othe r characters either provide no resolution at the regional scale Toxin Content under study here or add complementary information of use in interpreting the toxin composition results. The perspec When the molar concentrations of all toxins in an isolate tive on the PSP phenomenon that emerges is that the bio are totaled, or if they are converted to STX equivalent , es geography of toxic Alexandrium spp. in the study region timates of the total toxicity or toxin content are obtained. is not that of a single, widespread. homogeneous popula Among isolates, toxin content varied by nearly an order tion, but rather is comprised of several sub-populations, of magnitude (Table 1). The range was 25.5 to each with it own physiological characteristics and history. 1 1 187.9 fmol ceW (or 6 199 to 58695 fgSTX ceW ) in the M ixin g and genetic exchange among these populations non-archived toxin extracts. There was no appare nt rela may be infrequent. tionship between toxin content and species designation, as Four discrete toxic Alexandrium spp. populations can Alexandriwnfundyense and A. tamarense were similar in be propo ed on the basis of the timing of outbreaks, the 474 geographic separation between epi odes o f PSP. and an are inter persed geographically though there are areas, understanding of regional hydrography (Fig. I B). These such a the Gulf of Maine, where only one (A. fundyense) represent: ( I) the lower St. Lawrence estuary, including the occurs. Southern waters are especially heterogeneous. Far Gaspe peninsula, the western Gulf of St. Lawrence, and to the north where our data are limited to two isolates from Newfoundland: (2) the Gulf of Maine, including the Bay the St. Lawrence e tuary and Newfoundland, both species of Fundy; (3) Cape Cod; and (4) Connecticut and Long Is are represented. Turgeon et al. ( 1990) and Cembella et al. land. Further subdivisions are pe rhap possible (e.g. ( 1988) report that A. tamarense (= A. excavatum) is the Georges Bank), but these four provide a useful framework dominant morphotype in the lower St. Lawrence, but for analyses. A. f undyense al o occur . The regional taxonomic distri Sexual compatibility between Alexandrium tamarense bution is thus a mi xture of A. tamarense and A. fundyense and A. fundyense, demonstrated here for the first time. al in the Sr. Lawrence, Newfound land area (apparentl y dom lows gene flow between these populations as long as hy inated by A. tarnarense), a pure population o f A. fundyense drographic pathways facilitate intermixin g. The extenr to in the Gulf of Maine, and a mixture of the two species from whic h this ha. occurred is difficult to ascertain, however. Cape Cod southwards. Our data sugge t that exchange ha been limited and re productive isolation common. Below, we summarize the distributions of everal morphological and physiological Mating compatibility characters and use the e to evaluate the validity of the pro po ed population separations and the extent of inte rchange. Crosses between isolates and within species produced hyp We also consider our data in li ght of the species designa nozygoti c cysts approximately 50% of the ri me, whereas tions assigned to our strains on morphological grounds and self-crosses never yie lded cysts. This i the expected re argue that A. tamarense and A. fundyense could be varie sult for a heterothallic specie when crosse are made of ties of a single species and not separate species. i olates with unknown mating types. The arne rate o f suc ce s was obtained with cro · es between Alexandrium tam arense and A.fundyense, and there ulting cysts germinated Morphotaxono my to produce viable F 1 progeny. Additional cro es ofF 1 cul tures yielded cysts which were germinated to produce nor The taxonomy of the armoured, STX-produc ing dinoflag mal vegetative cells and viable cultures re presenting the ellates called the " tamarensis/catenella group" has long F2 generation. It is thus establi shed that A. tamarense and heen contentious. There is agreeme nt that these dinoflag A. fundyense are sexually compatible. e ll ates should be in the genus Alexandrium and nor Gon Maring experiment also provided information on the yaulax or Protogonyaulax (Ste idinger and Moestrup 1990), manner in which characters such as biolumine cence, the but concerns about species assignment remain. Some be ventral pore, and toxin composition are inherited. Vegeta li eve the "tamarensis/carenella" group represents a single tive cells and gametes of most dinoflagellates are hapl oid species complex comprised of numerou · biochemicall y (Pfi ester and Anderson 1987). and thus the presence or ab- distinct varieties who e allozyme and toxin composition cnce of any trait in the progeny cannot be attributed to patterns do not demon trate c lear di fferentiation (Taylor dominance. The most likely explanation for the absence of 1985: Cembella and Taylor 1986; Cembell a et al. 1987). a pore in almost all of the progeny of a cross between Alex An opposing view is he ld by tho e who differentiate ap andrium fundyense and A. tamarense (Table 2) is that this proximately 30 pecie on the basi of !> mall morphologi apparently simple trait is a quantitative character who e cal features such as the hapes of thecal plates or the pres expression is affected by more than one gene. Thi conclu ence or absence of pores (Balech 1985). The isolates in sion i also supported by the observation that the presence vestigated in the present tudy were examined and initially or absence of a pore was not always clear. In several cul categorized u in g these characte rs. tures, a few cells could be found that had a "vestigial" pore So me authors (e.g. Turgeon et al. 1990) have argued while the vast majority of cells had no pore. Since the pre that the St. Lawrence populations might be a separate pe cise pattern of inheritance and number of genes involved cies call ed Alexandrium excavatwn. Recentl y, however, is not known at present, it would be unwi e to conclude one o f u (E.B.) examined many natura l and cultured spec that all matings between A. tamarense a nd A. fwzdyense imens and concluded that it is possible to find examples o f wil l yield predominantly A. fundyense cell . all po sible tran itions between A. tamarense and A. exca vatum. On morphological ground . A. excavatum cannot be distingui shed from A. tamarense and there fore we use Bioluminescence the latter, older name throughout this texl. There are only two morphospecies responsible for the Loeblich and Loeblich ( 1975) used bioluminescence to production of PSP toxins within the region, i.e.. Alexan separate pecie within the '·tamarensis" group of Alex drium tamarense and A.fundyense. The critical feature dis andrium, but this di tinction wa subsequently rejected tingui hing the two taxa i the ventral pore on the margin (Schmidt and Loeblich 1979). Our re ult confirm this lat between plate I' and 4' in A tamarense that i absent in A. ter conclu ion, since biolumine cent and non-biolumine - fundyense (Balech 1985). A. tamarense and A. f undyense cent strain of both A. fundyense and A. tamarense occur 475 1 (Table I). evertheless, there is a geographic pattern to by HPLC and expressed in STX equivalen t ceW , since bioluminescent capacity within the region; virtually all of both are measures of potency. the northern isolates were biolumine cent, wherea south Cluster analysis and PCA of the toxin compo ition data ern i olate were mi xed, with a slight bias towards non both confirm the identification of at least two distinct biolumine cence. The underlying rea ons for this distribu groups of i olates. One is compri ed almo 1 exclu ively of tion are unclear. M ating between two bioluminescent par northern isolates (the exception being the Connecticu t ents produced at least one daughter cell that wa non-bio train GTC 02), all of which are Alexandrium fundyense. luminescent (Table 2). Crosses between biolumine cent The second cluster consists of southern isolates, which are and non-biolumine cent strains have not been attempted. a mixture of A. tamarense and A. fundyense. PCA analy sis (Fig. 3) identifies two clear outlier : ATSLI 2 from the St. Lawrence Estuary and AFNS85 from an tucket Shoa ls. Tox in composition and content The former strai n has a toxin profile that overlaps with both th e nonherly and southerly groups, while the Iauer has a A north-to-south trend of declining toxin content in Alex stronger signal for dcSTX than the rest of th e isolates in andriwn species within our study region was fir t reported the northern group (Fig. 2). PCA how that isolates from by Maranda et al. ( 1985) ba ed on mouse bioas ay results, Cape Cod are quite different from Connecticut and L ong and later augmented by Cembella et al. ( 1988) who pro fsland train , which more clo ely re emble th e northern vided HPLC toxin analy is of St. L awrence i olates not group. A larger sample ize would likely further delineate repre ented in the former study. Our HPLC analyse fur this Cape Cod cluster from the other two. ther support the ex istence of a geographic trend in toxic ity. though the relationship is weak. A regression between toxin content (expressed in total moles of all toxins per Biogeographic scenarios ce ll) and latitude was not significant (P > 0.05; data not shown), but when molar concentrations were converted to Two alternate hypotheses can account for the di rribution STX equiva lent (i.e., potency). the regression against lat of characters among toxic isolates of Alexandrium in our itude was ignifica nt at the 95% level (Fig. 4). The corre study region. The first is a scenario of recent and conti nu lation coefficien t was only 0.49, whereas a similar regres ing dispersal to the south from a center of origi n in the sion by Maranda et al. ( 1985) of many of the same strains north, followed by recombi nation and strong selection. The had a corre lation coe fficient of 0. 82. A partial explanation second hypothesis is that northern and south ern popula for the difference is th at five high toxicity, northern strains tions diverged from a common ancestor (vicariance) but that Maranda et al. analyzed had died and could not be in now repre ent loca lized populations. with relatively little cluded in our tud y. We also included a new i olate from mixing of genotypes among the regions delineated earlier. Nantucket Shoal , offshore of Cape Cod, which wa very ei ther hypothes is can be completely refuted by the data toxic relative to outhern isolates which were predomi pre ented in the pre ent study, but th e weight of evidence nantly of estu arine origin. If our rudy had included more leans towards the " local population/vicariance" hypothe train from the northern Gulf of Maine or from the St. sis. Lawrence. repre entatives of which have con i ten tly been The "recent and continuing di per al"' hypothe i is highly toxic (Cembella et al. 1988), the north-to- outh con istent with: ( I ) retention of exual compatibility trend in toxicity would likely have been stronger. Further among train regardless of origin or morpho pecie des re olution of this issue await toxin compo ition studie ignation; (2) the nonh-to-south flow of coastal waters in currently in progress on Alexandrium isolates from th e St. this region; (3) the ex istence of high toxicity forms in the L awrence region (Cembella per onal communication). sou th (e.g. strains GTCN02 and AFNS85); and (4) the rel This north-to-south trend of declining toxicity, though ative homogeneity of characteristics such as biolumines only weakly evident in our data, is probably a characteris ce nce and morphology in the north compared to th e het tic of the regional popu lation. It is not due to a higher in erogeneity in the south. trinsic toxicity of Alexandriumfundyense, the predominant Under this scenario, the heterogeneity in southern morphotype of northern isolates, since that species i , on popu lations could be accounted for by genetic recombina , average, similar in toxicity to A. tamarense (Table 1). Nor tion followed by selection and/or drift. The observed pat is it because northern isolates produce more mole of toxin, terns would have evolved very rapidly. since 1972 was pre si nce a regres ion between latitude and toxin content ex ·umably the first major introduction event. However, e pres ed in mole wa not significant. Rather, the trend lective forces capable of generating and maintaining this arises becau e outhern i olates produce proportionally gradient in characters have neither been identified nor hy more of the les -potent toxin than do northern strains (Fig. pothe ized. Another unappea ling aspect of thi hypothesi 2). The carbamate toxin that predominate within th e north is that recombination and selection would have to main ern i olates are much more potent (0 hima et al. 1989) th an tain morphologica l and bioluminescen t homogeneity in the the su lfamate toxin that are in highest proportion in the northern populations but be relaxed in some way to allow southern strains. Given this distribution of toxin composi heterogenei ty in southern form . Furthermore, ongoing tion , it follows th at toxin content would decrease from mating studies between northern and southern isolates have north-to-south when it is determined by mou se bioa say or yielded progeny with toxin composition patterns that are 476 generally equivalent to the parents' (D.M.A. unpublished tent with all avail able data. Neverthe less, the correct story data). Similar results were reported for Alexandrium cafe may well be a combination of both, with recent and con nella and for A. tamarense by Sako et al. ( 1992) and Ish tinuing spreading from the north occurring within the Gulf ida et al. ( 1993). Those workers demonstrated that simple of Maine (as clearly occurred during the 1972 New Eng Mende li an inheritance predominated and that on onl y one land red tide; Hartwell 1975) and possibly withi n the Gulf occasion, out of severa l dozen cyst germinations, did the of St. Lawrence, but endemi c, localized populati ons persist progeny o f a cross between strains with two different toxin in g without geneti c exchange in most southern locations. compositions produce a suite of toxins different from ei ther of the pare nts. If recent and continuing dispersal were occurring in our study region, matings between northern The species concept and southern populations should be yie lding both low and high toxicity proge ny in the south, yet the latter are very The cumulati ve data on Alexandrium spp. from the present rare. The validity of this hypothesis therefore depends upon study and other illu trate the conflict between traditional the existence of an additional type of environmental selec classification and the needs of ecologists. Traditional tax tion in southern waters- one that favors strains producing onomy is based primarily on morphological traits and bi sulfamate toxins and is ultimately lethal to strains produc ogeographic distributions, and for phytoplankton, the most ing mostly carbamates. This is again a theoreti cal possibil commonly used definition of a species is that group of in ity, but its like lihood is questionable . divi duals which maintains disc rete, morphological fea In the alternati ve " local population/vicariance" sce tures that are genetically fixed. There is no require ment nario, northern and southern populations would have been that this be a physicall y large feature, onl y that it be dis established from at least one common ancestor in the dis crete and constant. In this context, the stability of the ven tant past and would have evolved more o r less indepen tral pore in culture (Balech 1985) argues that A.fundyense dently since their initial separati on. Under thi s mode l, oc and A. tamarense are "good" species. Here we de monstrate casional dispersal from one area to the other could occur, that these two forms are sexually compatible, whi ch might but the populations would be generall y isolated from each be interpreted to mean just the opposite, i.e., that they a re other. The high toxicity strain AFNS85 could be an exam not good "biological" species as originally defined by ple of such an occasional introduction. It wa found in Mayr ( 1942). The retention of sexual compatibility in cul southern waters off Nantucket Shoals, an area to the east ture does not necessarily imply that genetic exchange is a of Cape Cod influenced by flow from the Gulf of Maine. regul ar event in nature, however, as compatibility can be and possibly from the Scotian She lf as well (Fig. I B). a primitive trai t retained by species whi ch are not c losely The evidence in favor of this alternati ve hypothesis in related, (e.g. hybridization; Costas I 986). Therefore, our c ludes: ( I) Cape Cod and Georges Bank act as hydro mating results do not by the mselves refute the classifica graphic barriers to force the predominant north-south flow tion of the two species of Alexandrium studied here as dis offshore, away from the estuaries and embayments where tinct taxa. A more seri ous challenge to their taxonomic sep southern Alexandrium spp. cells and cysts are localized; aration as species comes from the analyses reported here (2) the toxin composition profiles of most Cape Cod iso and by Hay home et al. ( 1989). all of which indicate that in lates cluster together and are generally distinct from those the toxic Alexandrium spp., morphology is a poor predic of the Gulf of Maine and Connecticut/Long Island; (3) ger tor of the pattern of differences in biochemical or physio m ination of Cape Cod Alexandrium spp. cysts is apparently logical traits suc h as all ozyme banding. toxin composition, regulated by a different mechanism than the intern a l clock or bioluminescence. Even stronger evidence challenging that controls cysts in the G ul f of Maine (Anderson and the utility of morphospecies designations for ecological Keafer 1987); (4) the geographic distribution of Alexan purposes in toxic Alexandrium spp. comes from the recent driwn spp. populations has several large gaps, such as the work of Scholin and Anderson ( 1994) on the ribosomal area a long the Scotian she lf between the St. Lawrence and RNA genes in the '·tamarense/catenella" group, whi ch the Gulf of Maine, or between Cape Cod and Connecticut c learly demonstrates that isolates cluster together more (Fig. 1 B); (5) there was a report of A. tamarense (=Gon logicall y on the basis of geographic origin that morpho yaulax tamarensis) in Vineyard Sound, southern Cape Cod taxonomy. long before the massive 1972 New England red tide ( Lil This discussion is not simply an academic exercise. Par lick 1937); (6) mating of parents wi th northern and south adoxically, the impact of species designati ons falls most ern toxin composition genotypes yields progeny resem heavily on ecologists, who commonly assume that a spe bling the parents (D.M.A. unpublished data), and strains cies definition predicts a pattern of physiological differ with " northern" toxin composition are rare ly observed in ences (in our case, toxicity or bioluminescence) that can the south; and (7) localized southern populations of toxic be used to define a ni che. Surprisingly. thi important as Alexandrium spp. could have persisted for many years sumption has rarely been te ted quantitati vely. as it has without detection because of their low toxici ty. been here. Other quantitative studies of the issue for ma Neither of these two biogeographic scenarios can be rine phytoplankton support our conc lusion that species complete ly discounted, but we favor the " local popula designations can sometimes be of little use fo r ecological tion/vicari ance" hypothesis since it does not require highl y purposes (Beam and Himes 1982; Gall agher 1986; Wood directional , unknown elective mechanisms and is consis- 1987; Brand 199 1). 477 We recognize that the science of taxonomy seeks only Baleeh E ( 1985) The genus Alexandrium or Gonyaulax of the tam to delimit stable groups of individuals that share common arensis group. In: Anderson DM, White AW, Bade n DG (eds) Toxic dinofl agell ates. Elsevier, 1ew York, pp 33-38 features (Manhart and McCourt 1992; Wood and Leatham Beam CA, Himes M ( I 982) Di stribution of members of the Cryp 1992) and that there need not be other declared purposes, rfl econdinium cofln ii (Dinophyceae) species complex. J Proto such as predictability or utility to other fi elds. To some ex zoo! 29:8- 15 tent, the problems of classification discu sed here are ar BrandLE ( 199 1) Review of gene tic variation in marine phytoplank tifact of what ecologists and evol utionary biologists want ton species and the ecological implications. Bioi Oceanogr (NY) 6:397-409 species to "do" rather than they often "are". A reasonable Cembella AD. Sullivan JJ , Boyer GL. Taylor FJR, Ande rson RJ solution is to reserve species-level designati ons for taxa ( 1987) Variation in paralyti c shell fi sh toxin composition within where there is consistency among morphological, bio the Protogonyaulax tamarensis/carenella species complex: red chemical, molecular, and physiological data. Using this ap tide dinofl agellates. Biochem Syst Ecol 15: 17 1- 186 Cembell a AD, Taylor FJR ( 1986) Electrophoreti c variability proach, the morphological variants identifiable as Alexan within the Prorogonyaulax tamarensislcarenella species com drium tamarense, A. carenel/a and A.fundyense might best plex: pyridine linked dehydrogenases. Biochem Syst Ecol 14: 3 1 I be described as "varieti es'· in the botanical sense (Taylor -32 1 1976; Steidinger et a l. 1980). Cembe lla AD, Therri ault JC, Beland P ( 1988) Toxicity of cultured isolates and natural populations of Protogonyaulax tamarensis Designation of Alexandrium ramarense and A. fun (Lebour) Taylor from the St. Lawrence Estuary. J Shellfish Res dyense as varieties would not solve the many problems in 7:611 -622 herent in their biology; it s imply displaces them to a dif Cohn MS. Olsen P, Majoney JB, Feerst E ( 1980) Occ urrence of the fe rent le vel. It informs users of this classifi cation that these dinofl agellate Gonyaulax tamarensis, in New Jer ey. Bull NJ Ac ad Sci 33:45 -49 entities are not completely isolated from each other and Costas E ( 1986) Aparicion de hibridos interspecificos de Prorocen may not be stable. In taking this position, we recognize that rrum micans y Prorocenlrum rriestinium en cultivos experimen recommending the use of trinomials i not ri sk-free. Many tales. Boln lnst exp Oceanogr 3:6 1 - 66 ecologists and systemati sts are reluctant to use infraspe Franks PJS, Anderson DM ( 1992) Alongshore transport of a toxic cific names and they are often dropped. Jn the case of the phymplankton bloom in a buoyancy c urrent: Alexandriwn tam arense in the Gulf of Maine. Mar Bioi 12: 153- 164 Alexandrium species, dropping the varietal designation Fritz L, Triemer RE ( 1985) A rapid simple technique utilizing Cal completely would result in a genuine loss of information. cafl uor White M2R for the visualization of dinofl agellate thecal The studies on Alexandrium spp. represent one o f the plates. J Phycol 21:662-664 most complete data sets on the patterns of species and pop Gall agher JC ( 1986) Population genetics of microalgae. Algal bio mass technologies: an interdisciplinary perspective. Nova Hed ulati on-level genetic diversity in any protist. When these wigia 83:6- 14 data arc combined with more fragme ntary com pi lations for Guillard RRL, Ryther JH ( 1962) Studies of marine plankton diatoms. oth er species, they indicate that many assumptions inher I. Cyclorella nana Hustedt and Detonula conf ervacea (Cleve) ent in fi eld and laboratory investigations of "species" by Gran. Can J Microbi ol 8:229- 239 Hall S. Stri chartz G, Moczydlowski E. Ravindran A, Reichardt PB phytoplankton ecologists may be incorrect. The problem ( 1990) The saxitoxins; sources. chemi stry, and pharmacology. ln: of speciati on and genetic diversity at lower taxonomic le v Hall S, Strichartz G (eds) Marine toxins: origin . structure, and els in protists need further study in order to determine if molecular pharmacology. Amer. C hem. Soc. Symp. Ser. No. 4 18, th is conclusion is correct. Amer. Chem. Soc. Washington. DC, pp 29-65 Hartwell AD ( 1975) Hydrographic factors affecting the distribution and movement of toxic dinoflagellates in the western Gulf of Acknowledgements We are pleased to acknowledge the generous Maine. In: Lo Cicero VR (ed) Proceedi ngs of the fi rst interna gift of toxin standards for HPLC analysis from Y. Oshima and the tional conference on toxic dinoflagellate blooms. Massachusetts help of A. White and J. Martin in updati ng PSP records for Canada. Science and Techno logy Foundati on, Wa ke fi eld, Massachu ells, This work was supported by National Science Foundation grants pp 47-68 OCE891 1226 (to DM A) and OCE-8809484 and DEB-9 1 19405 (to Hayhome BA, Anderson OM, Kulis OM, Whitten DJ ( 1989) Varia JCG). by the National Sea Grant College Program Office, Depart tion among congeneri c dinoflagellates from the northeastern ment of Commerce under Grant No. NA90-AA-D-SG480 (WHO I United States and Canada. I. Enzyme electrophoresis. Mar Bioi Sea Grant Projects R/B- 100 and 112 to DM A), and by a Canadian I 0 I :427-435 NSERC postdoctoral fell owship (to GJD). Contri bution No. 8643 Ishida Y, Kim CH, Sako Y, Hirooka N, Uchida A ( 1993) PSP toxin from the Woods Hole Oceanographic Institution. production is chromosome dependent in Alexandriwn spp. In: Smayda T, Shimizu Y (eds) Tox ic phytoplankton blooms in the sea. Elsevier. Amsterdam. pp 88 1-887 Lillick LC ( 1937) Seasonal studies of the phytoplankton off Woods Hole Mas achusetts. Bioi Bull mar bioi Lab, Woods Hole References 73:488-503 Loebl ich LA, Loeblich AR Ill ( 1975) The organis m causi ng New Anderson DM, Keafer BA ( 1987) An endogenous an nual clock in England red tides: Gonyaulaxexcavata . In: Lo Cicero V (ed) Tox the tox ic marine dinofl agell ate Gonyaulax ramarensis. Nature ic dinofl agellate blooms. Proceedings of the first international y 325:616-617 conference on toxic dinofl agellate blooms. Massachusetts Sci Anderson DM. Kulis DM , Orphanos JA, Ceurvels AR ( 1982) Dis ence and Technology Foundation, Wakefield. Massachusetts, tribution of the toxic red tide dinofl agell ate Gonyaulax tamaren pp 207-224 sis in the southern New England region. Estuar cstl Shelf Sci Manhart JR. McCourt RM ( 1992) Molecular data and species con 14:447-458 cepts in the algae. J Phycol 28:730-737 Anderson DM. Wall D ( 1978) The potential importance of benthic Maranda L, Anderson DM , Shimizu Y ( 1985) Comparison of toxic cysts of Gonyaulax tamarensis and Gonyaulax excavata in in ity between popul ations of Gonyaulax tamarensis of eastern itiating toxic dinofl agellate blooms. J Phycol 14:224-234 North American waters. Estur est! Shelf Sci 2 1:40 I -4 10 478 Martin JL, White AW ( 1988) Distribution and abundance of the tox iodoma. In: Grane li E. Sundstrom B. Edler L, Anderson 0 (eds) ic di noflagellate Gonyaulax excavata in the Bay of Fundy. Can Toxic marine phytoplankto n. Elsevier, New Yo rk, pp 522-523 J Fish aquat Sciences 45: 1968- 1975 Steidinger KA. Tester LS, Taylor FJR ( 1980) A redescription of Py Mayr E ( 1942) Systematics and the origin of species. Columbia Uni rodinium bahamense var. compressa (Bohm) stat. nov. from Pa versity Press, New York cific red tides. Phycologia 19:329-337 Oshima Y. S ugino K. Yasumoto T ( 1989) Latest advances in HPLC Taylor FJR ( 1976) Fl agell ate phylogeny: a study in conflicts. J Pro analysis of paralyti c hell fi sh toxin . In: atori S, Hashimo to K. tozool 23:28-40 Ueno Y (eds) M yco t ox in ~ and phycotoxin~. Elsevier, Amster Taylor FJR ( 1985) The taxonomy and relationships of red tide dino dam. pp 3 19 -326 flagellates. In : Ander on OM. White AW. Baden DG (ed~) Tox Pfiester L. Ander on OM ( 1987) Dinoflagellate reproduction. ln : ic dinoflagellate . El evier. ew York. pp 11-26 Taylor FJR (ed ) The bio logy of dino fl agellates. Blackwell Sci Turgeon J, Cembella AD. Therriault J-C. Beland P ( 1990) Spatial entific Publications, Londo n, pp 61 1-648 d i ~ tributi o n of resting cysts of Alexandrium spp., in ediments of Prakash A, Medcof JC, Tenant AD ( 197 1) Paralytic shell fish poison the lower St. Lawrence estuary and the Gaspe coast (Eastern Can ing in eastern Canada. Bull Fish Res Bd Can 177: 1-87 ada). In: Graneli E. Sundstrom B, Edler L, Anderson OM (eds) Sako Y. Kim C l-1 , Ishida Y ( 1992) Me ndelian inheritance of paralyt Toxic marine phytoplankto n. Elsevier. New Yo rk, pp 238-243 ic she llris h poisoning toxin in the marine di noflagellate Alexan von Stosch HA ( 1973) Observation on vegetative reproduction and drium catenella. Biosci Biotech Biochem 56:692-694 sexual life cycles of two freshwater dinoflagellate , Gyumodin Schmidt RJ , Loebli ch AR Ill (1979) A discussion of the ~ys t emat iwn pseudopalustre and Wolos~ynskia apiculalll p. nO\. Br. Phy ics of toxic Gonyaulax species containing para I ytic shell fish poi col J 8: I 05- 134 son. In: Taylor DL, Seliger HH (eds) Toxic dinoflagell ate blooms. White AW. assif J , Shumway SE, Whittaker OK ( 1993) Recent oc El evier, New York, pp 83-88 c urrence of paralytic shellfi h toxins in offshore shell fish in the Scholin CA, A nderson OM ( 1994) Identification of gro up-and train northeastern United States. In: Smayda T, Shimizu Y (ed ) Tox specific genetic markers for globally distributed Alexandriwn ic phytoplankton blooms in the sea. Elsevier, Amsterdam. (Dino phyceae) species: I. Restrictio n fragment length polymor pp 435-440 phism analysis of ~ mall subunit ribosomal RNA genes. J Phycol White DRL. White AW ( 1985) First report of paralytic shellfish poi (in press) soning in Newfoundland. In: Anderson OM, White A\1./. Baden Schrey SE, Carpenter EJ. Anderson OM ( 1984) The abundance and DG (eds) Toxic dinonage l late~. Elsevier. New York. pp 511 -5 16 distribution of the toxic dinoflagell ate. Gonyaulax tamarensis. in Wood AM ( 1987) Molecular biology. single cell a n a l y~is and quan Long Is land E tuaries. Estuaries 7:472-477 titative genetics: new evolutionary genetic approache in phyto Smith PC, Schwing FB ( 1991) Mean ci rc ul ation and variabil ity on plankton ecology. In: Mague FC, YenLsch C M (eds) lmmuno the eastern Canadian continental shelf. Conlin Shel f Res chemical approache~ to estuarine. coastal and ocenaographic 11 :977- 10 12 questions. Elsevier. New York. pp 4 1-73 Steidinger KA. Moewup 0 ( 1990) T he taxonomy of Gonyaulax, Py Wood AM, Leatham T ( 1992) The species concept in phytoplanKton rodinium, Alexandrium. Gessnerium. Protogonyaulax and Con- ecology. J Phycol 28:723-729 ,.