60 



last biota included meroplankton (orga- 

 tiisms spending part of their life cycle in the 

 water column), holoplankton (spending all 

 of their lives in the water column) , demer- 

 sal plankton (benthic species that vertically 

 migrate into the water) , and tychoplankton 

 (suspended benthic organisms) . Ballast wa- 

 ter therefore acts as a phyletically and 

 ecologically nonselective transport vector. 

 Certain taxa occurred in high densities: we 

 estimated copepod densities were greater 

 than 1.5 X 10' per cubic meter and spionid 

 polychaete larvae, barnacle nauplii, and 

 bivalve veligers greater than 2 X 10^ per 

 cubic meter {li). 



Despite the lack of selectivity, certain 

 taxa predominate. Five phyla accounted for 

 more than 80% of taxa recorded: crusta- 

 cearu (31% of all taxa present), polychaete 

 annelids (18%), turbellarian flatworms 

 (14%), cnidarians (11%), and mollusks 

 (8%). Taxa found in many or most vessels 

 included copepods (present in 99% of 

 ships), polychaetes (89%), barruicles 

 (83%), bivalve mollusks (71%), flatwomis 

 (65%) , diatoms (93%) , gastropod mollusks 

 (62%), decapod crustaceans (48%), and 

 chaetognaths (47%). 



For some taxa the number of released 

 individuab may vary greatly among ships, 

 whereas the frequency of release may be 

 high (Table I). Gastropods were abundant 

 in only 2.5% of ships but present in 62% of 

 ships sampled, decapods were abundant in 

 only 3.1% and present in 48% of ships, 

 spionids were abundant in 24% of ships and 

 present in 85% of ships, and nonharpacti- 

 coid copepods were abundant in 61% of 

 ships and present in 98% of ships. 



Behavioral and life history traits make 

 some taxa less prone to being transported 

 by ballast water. Taxa with both a strictly 

 benthic life-style and with brooded or 

 crawl-away young [for example, brooding 

 gastropods, bivalves, and anthozoans (14)1 

 would rarely be in the water column when 

 ballast is pumped on board. Similarly, 

 organisms with an extremely short plank- 

 tonic life (sponges, direct-developing bry- 

 ozoans, and ascidians) would rarely be 

 caught. Nektonic organisms (such as lish) 

 may be able to resist either the water 

 intake pressures of the ballast pump or may 

 be able to avoid the plankton net. How- 

 ever, any taxa likely to attach to algae 

 (15) could be taken up along with the drift 

 algae (16. 17). 



In the past 20 years, numerous aquatic 

 invasions have occurred (Table 2). Many of 

 these now appear to be related to ballast 

 water traruport (18). The taxa of these 

 documented invasions (Table 2) are all 

 represented (except comb jellies) in our 

 samples of ballast water (Table 1 ) . Howev- 

 er, some higher taxa frequently found in 

 ballast water have not been often reported 



as invasive species. Conversely, some high- 

 er taxa that are reported relatively frequent- 

 ly as invaders were not found frequently in 

 our samples. Although we recognize that 

 high frequency of release does not necessar- 

 ily lead to successfril invasions, we suggest 

 that there have been far more introductions 

 of polychaetes, flatworms, and diatoms 

 than have been reported. Invasions of in- 

 tensely studied larger-size animals (such as 

 fish, mollusks, and decapods) are more ap- 

 parent and thus more noticeable. We pre- 

 dict that more invasions of both large and 

 small organisms will be recognized as sus- 

 ceptible regions are investigated and that 

 new invasions will be discovered in well- 

 studied regions (19). 



Knowledge of species' natural geograph- 

 ic distributioru is of paramount importance 

 for interpreting patterns in ecology, evolu- 

 tion, and biogeography. Unfortunately, the 

 systeraatics of most marine taxa are far from 

 complete, and the discovery of previously 

 unrecognized species in regions impacted by 

 ballast water release (almost all coastal 

 zones of the world) must now be viewed 

 critically as potential invasioiis (20). Con- 

 versely, for easily identified species, unrec- 

 ognized historical transport may have led to 

 false conclusions of r\atural cosmopolitan- 

 ism. Thus, many introduced species may be 

 cryptic, having invaded and gone unrecog- 

 nized or been mistaken as native species. 

 Both these situations confound our under- 

 standing of historical patterns of dispersal, 

 gene flow, and speciation: geographic bar- 

 riers to dispersal and gene flow are readily 

 breached by ballast water transport. Simi- 

 larly, we must now recognize that the com- 

 position of aquatic conununities may be 

 influenced by both recognized and cryptic 

 invasions. 



Ships take up and release ballast water in 

 bays, estuaries, and inland waters and then 

 release this water into similar environments 

 around the world. Many of these bodies of 

 water are disturbed by the efiects of exten- 

 sive urbanization (21), rendering them es- 

 pecially susceptible to invasions (22) that 

 further alter community structure and func- 

 tion. The invasion of the Asian clam Pot- 

 amocorhula amurensis in San Francisco Bay 

 {2i), the zebra mussels Dreissena pdymorpha 

 and Dretsseru sp. in the Laurentian Gteat 

 Lakes (24, 25), and the comb jelly Mnenu- 

 opsis kidyi in the Black Sea (26) are dramat- 

 ic examples of the catastrophic impact of 

 ballast water introductions. The ecological 

 roles and impacts of invading species can 

 only be partially predicted from knowledge 

 of their biology and ecology in donor re- 

 giotu (2). For these reasoris, bays, estuaries, 

 and inland waters with deep water ports — 

 marine analogs of despoiled, highly invaded 

 oceanic islands — may be among the most 

 threatened ecosystems on the planet (27). 



SCIENCE • VOL. 261 • 2 JULY 1993 



REFERENCES AND NOTES 



1 C P Stone and D B Stone. Eds Corservalion 

 Biology in Hawaii (Univ of Hawaii Press, Honolu- 

 lu. 19891. W V Reid and K R Miller. Keeping 

 Options Alive The Scientific Basis for Conserving 

 Biodiversity (World Resources Institute. 1989). 

 lUCN/UNEP/WWF. Canng for the Earth A Strate- 

 gy for Sustainable Living (lUCN. Glarxj, Switzer- 

 land. 1991). M Lodge. Trends Ecol Evol 8. 

 133 (1993) 



2 F di Castri, A H Hansen, M Debussche. Eds 

 Bioiogical Invasions in Europe and the Mediterra- 

 nean Basin (Monograptiiae Biologicae 65. Kluwer 

 Academic, Leiden. Netherlands. 1990). J A 

 Drake e( a/ . Eds . Erx^ogy of Biological Invasions 

 A Global Perspective (SCOPE 37. Wiley. New 

 York. 1989). R H Groves and J J Burdon, Eds, 

 Ecology of Biological Invasions an Australian 

 Perspective (Cambridge Univ Press. London, 

 1986), R Hengeveld, Ed . Dyr^mics of Biological 

 Invasions (Chapman & Hall, London, 1989). H, 

 Komberg and M H Williamson, Eds , Quantitative 

 Aspects of the Ecology of Biological Invasions. 

 Philos Trans R Soc London Ser B 314, 501 

 (1986), H A Mooney and J A Drake. Eds., 

 Ecology of Biological Invasions of North America 

 and Hawaii (Spnnger-Verlag, New York, 1986) 



3 D Bramwell, Ed , Plants and Islands (Academic 

 Press, New York, 1979). C M King, Immigrant 

 Killers Introduced Predators and ttw Conserva- 

 tion of Birds in t^ew Zealand (Oxford Univ Press. 

 Auckland. New Zealand. 1984). P M Vitousak 

 and L R Walker. Ecol Monogr 59. 247 (1989) 



4 J T Carlton. Conserv Biol 3. 265 (1989). 



5 . Oceanogr Mar Bid Anna Rev 23. 313 



(1985). R J Williams. F B GrifTiths. E, J Van der 

 Wal. J Kelly. Esluanne Coastal Shelf Sci 26. 409 

 (1988). R P Baldwin. J R Soc N2. 22. 229 

 (1992), G M Hallegraeff. C J Boloh. J Bryan. B. 

 Koerbin. in Toxic Marine Ptr/toptanMon. E Graneli 

 etal. Eds (Elsevier. New York. 1990). pp 475- 

 480. G l»l Hallegraeff and C J Boloh. J Plankton 

 Res 14. 1067 (1992) See also {32. 57) 



6 Ballast water may be fresh, brackish, or manne, 

 depending upon the tiallasting site Sediments 

 may also be entrained, the release of such sedi- 

 ments tias tieen linked to the introduction of toxic 

 dinoflagellates (in their benthic cyst stage) in 

 Australia (G 1^ Hallegraeff and C J Bolch, Mar. 

 Poiiut Bull 22, 27 (1991), G M. Hallegraeff. 

 PhycologiaX. 79 (1993)1. (57) 



7 Ballast water may also t» taken atx)ard and 

 discharged in many other patterns. For example, 

 offshore water may be t>allasled in one ocean and 

 detMllasIed in another ocean, which would result 

 in the movement of oceanic taxa in addition to 

 neritic laxa 



8 Larger (>2 cm) organisms (such as fish) may fail 

 to pass through the intake grates or may be 

 destroyed tay tfie impeller pump tjlades 



9 ShipsweresampJedfrom 198610 1991 Fivetostx 

 vertical quantitative hauls were made in each 

 vessel using an 80-Mm mesh. 5-m-diameter 

 plankton net lowed at 5 m s"' in 10 to 20 meters 

 of tiallasi water of floodat>le cargo holds Samples 

 were examined alive under a stereomicroscope to 

 ensure the inclusion of fragile specimens All 

 samples were preserved and retained Speci- 

 mens of many taxa were cultured until they grew 

 to a size thai pemiilted identification Information 

 on the volume, source, and age of ballast water 

 was obtained Cargo holds sampled contained an 

 average of 1 09 k 10* (SO = 2 7 x 10^ metric 

 lons(=l 09 X 10' liters, SD ■■2.7 x 10°) of water, 

 total tjallast water on tx>ard averaged 2 01 x 10* 

 (SD = 6 4 X ltf>) metnc Ions (=2 01 x 10' liters, 

 SD = 6 4 X 10^ of water 



10 Two relatively common manne ptiyla not fourxj in 

 our samples are ttie Potifera (sponges) arxJ 

 C^enophora (comb jellies) The atisence of cteno- 

 phores may reflect a t)ias against extremely frag- 

 ile taxa Allematively, cterxsptiore disthtxition is 

 often temporally and spjatially uneven [K f^^ount- 

 ford. Estuarine Coastal Mar Sci 10. 393 (1980); 

 E Oeason. Estuanne Coastal Shelf Sci 15, 121 



•1 



