rects lateral aggregation ofchains duringfibrillogenesis (12). Glucose, a polyol,might interfere with such hydrophobicinteractions.

Since fibril formation is an early eventin collagen synthesis (13), we speculatethat the high glucose environment indiabetes may prevent fibril formationand the subsequent cross-linking reac-tion that occurs soon after synthesis ofcollagen. It has been shown that 30 per-cent of newly synthesized collagen isdegraded in the skin of streptozotocin-induced diabetic rats, while only 13 per-cent is degraded in control rats (3). Ourobservations may explain the enhancedcatabolism of newly synthesized colla-gen in diabetics. Collagen from diabetictissues has been suggested to be morehighly cross-linked because of its ob-served higher resistance to collagenaseand to extraction with acetic acid incomparison to normal tissue (14). Be-cause collagen becomes progressivelymore cross-linked with age, and becausenewly synthesized collagen is not cross-linked and is less stable than the preex-isting collagen in diabetic tissue, the se-lective depletion of newly synthesizedcollagen may leave tissues composedpredominantly of preexisting, more high-ly cross-linked collagen. The net resultthen would be an increase in the degreeof cross-linking and a decrease in thequantity of interstitial collagen in diabet-ic tissues.The specific effect of glucose on fibril

formation may also explain the tissue-specific changes of collagen in diabetics.A net loss of interstitial collagen massoccurs in intact skin (2), wounds (1), andbone (15), whereas a net accumulation ofbasement membrane collagen mass oc-curs in the intestine (2) and glomerulus(16). Interstitial collagen forms fibrils,whereas basement membrane collagendoes not (17). The accumulation of base-ment membrane collagen in diabetic tis-sue may be due to enhanced synthesis(17), decreased degradation (18), andsome additional mechanisms (19). Thus,glucose specifically stimulates the catab-olism of interstitial collagen-which re-quires cross-linking for persistence-anddoes not affect the catabolism of base-ment membrane collagen.

YEONG-HAU LIENROBERT STERN*

Department ofPathology, School ofMedicine, University of California atSan Francisco, San Francisco 94143

JOSEPH C. C. FuROBERT C. SIEGEL

Departments of Medicine andOrthopedic Surgery, School ofMedicine, University of California28 SEPTEMBER 1984

References and Notes1. W. H. Goodson III and T. K. Hunt, J. Surg.

Res. 22, 221 (1977).2. M. Schneir, J. Bowersox, N. Ramamurthy, J.

Yavelow, J. Murray, E. Edlin-Folz, L. Golub,Biochim. Biophys. Acta 583, 95 (1979).

3. M. Schneir, N. Ramamurthy, L. Golub, Diabe-tes 31, 426 (1982).

4. C. A. Vater, E. D. Harris, R. C. Siegel, Bio-chem. J. 181, 639 (1979).

5. R. C. Siegel, Proc. Natl. Acad. Sci. U.S.A. 71,4826 (1974).

6. T. Hayashi and Y. Nagai, J. Biochem (Tokyo)72, 749 (1972).

7. S. R. Pinnell and G. R. Martin, Proc. Natl.Acad. Sci. U.S.A. 61, 708 (1968); W. Traub andK. A. Piez, Adv. Protein Chem. 25, 243 (1971);M. L. Tanzer, Science 180, 561 (1973).

8. R. C. Siegel and J. C. C. Fu, J. Biol. Chem. 251,5779 (1976).

9. S. L. Schneider and R. R. Kohn, J. Clin. Invest.66, 1179 (1980); A. le Pape, J.-P. Muh, A. J.Bailey, Biochem. J. 197, 405 (1981).

10. M. P. Cohen, E. Urdanivia, M. Surma, V. Wu,Biochem. Biophys. Res. Commun. 95, 765(1980); A. le Pape, J.-D. Guitton, J.-P. Muh,ibid. 100, 1214 (1981).

11. P. J. Higgins and H. F. Bunn, J. Biol. Chem.256, 5204 (1981).

exotic clam.

The spread of the freshwater Asiaticclam Corbicula c.f. fluminea (Muller,1774) across North America since itsaccidental introduction about 50 yearsago (1) has led to controversy over themode of transport that could account forits invasiveness. Most often cited are

theories of human transport (2-4), inci-dental transport of small byssate juve-niles attached to bird feet or feathers orcarried within fish or bird gastrointesti-nal tracts (3, 5), and direct transport oflarvae or juveniles swept along streambottoms with water currents (2, 3, 6).Although each of these transportationmodes is feasible, there is little evidenceto support any one of them as a primarymeans of dispersal. For instance, theviability of Corbicula in bird gastrointes-tinal tracts during transport is question-able, and high clam mortality has beensuggested to occur (5, 7). Mackie (8),discussing possible methods of disper-sion for other sphaeriid bivalves, con-cluded that their transport within birdintestinal tracts is unlikely (again be-cause of mortality) but that externaltransport on bird feet or wings or oninsect legs are still viable options. Whilethis might be possible, although withprobably low frequency, it is still, aswith all the other proposed transportmodes, a passive activity. In each case

12. R. L. Treistad and K. Hayashi, Dev. Biol. 71,228 (1979).

13. M. J. Capaldi and J. A. Chapman, Biopolymers23, 313 (1984).

14. C. R. Hamlin, R. R. Kohn, J. H. Luschin,Diabetes 24, 902 (1975); L. M. Golub et al.,Biochem. Biophys. Acta 534, 73 (1978); K.Chang et al., Diabetes 29, 778 (1980).

15. M. E. Levin, V. C. Boisseau, L. V. Avioli, N.Engl. J. Med. 294, 241 (1976).

16. M. P. Cohen and A. Khalifa, Biochem. Biophys.Acta 500, 395 (1977); M. Brownlee and R. G.Spiro, Diabetes 28, 121 (1979).

17. P. Bornstein and H. Sage, Ann. Rev. Biochem.49, 957 (1980).

18. H. Fushimi and S. Tarui, J. Biochem. 79, 265(1976); A. Y. Chang, Biochem. Biophys. Acta522, 491 (1978); R. Weil, III, et al., Arch.Pathol. Lab. Med. 100, 37 (1976).

19. R. Vracko, Diabetes 23, 94 (1973).20. We thank D. Sansoe for technical assistance.

Supported by a grant from the Shriner's Hospi-tal for Crippled Children, San Francisco Unit(R.S.).

* To whom correspondence should be addressedin care of Klaus Weber, Max-Planck-Institutefor Biophysical Chemistry, Gottingen D-3400,Federal Republic ot Germany.

16 April 1984; accepted 14 June 1984

the transport is accidental, and the bi-valves play no active role in dispersion.We now report an unusual active partici-pation in transport by small adult Corbic-ula that may account for much of theirdownstream and perhaps interstream in-vasiveness.

Small adult Corbicula (shell length, 7to 14 mm) collected in March 1984 fromTallahala Creek near Runnelstown, Mis-sissippi, were found to be capable offloating after being exposed to gentlewater currents produced by an aquariumfiltration system (current speeds, 10 to20 cm/sec). The behavior of the clamswas consistent (Fig. 1). After initial ex-

posure to the water flow, the clamswould right themselves from the bareaquarium floor with their active andmuscular foot. The clams would come tolie with their valves perpendicular to thesubstratum and with siphons directedupward. In contrast to its normal pump-ing activities (as seen during feeding or

respiration), the exhalent siphon of eachclam was abnormally distended with awide lumenal space. After a period of 2to 8 minutes, during which the clamscontinued to pump actively, they wouldgently lift off the substratum, often withtheir foot extended, and drift into thewater column of the aquarium with thecurrent flow. The upright posture of the

1491

Flotation of the Bivalve Corbicula fluminea as aMeans of Dispersal

Abstract. Small specimens of the Asiatic bivalve Corbicula c.f. fluminea (Muller)secrete long mucous threads through their exhalent siphons that act as draglines tobuoy the animal into a water column. These mucous strands, secreted in response towater current stimuli, are produced by dense accumulations of ctenidial mucocytesand may help in the downstream or interstream dispersal of this rapidly spreading

on

July

13,

201

1w

ww

.sci

ence

mag

.org

Dow

nloa

ded

from

bivalve, usually elaborated by extensionof the foot, allowed water to flow be-neath as well as around it. This increasedthe lifting power of the water current.

Before they lifted off, the clams pro-

duced long, relatively viscous mucous

threads that extended upward into thewater flow from within the exhalent si-phon (Fig. 2). This produced sufficientdrag to lift the clams above the substra-tum. This kite-like effect lifts the smallclams into the water column in much thesame way as the silk threads of some

spiderlings act in aerial dispersal (9).While floating, the clams maintained an

abnormally expanded exhalent siphon,probably a reflection of the production ofthe mucous thread. Within the watercolumn the clams were relatively nonre-

sponsive to tactile stimuli that wouldordinarily cause adduction and with-drawal of siphon and foot. This insensi-tivity might be important in preventingearly adduction while floating.- Adduc-tion usually broke the connection be-tween the clam and mucous thread.Upon adduction and corresponding lossof mucus, the clams sank. The clamsalso gradually sank without adduction or

loss of the mucous thread when out ofthe influence of a water current of ade-quate velocity. Thus, in aquaria whenthe water flow was terminated the clamssank but at a rate two to four timesslower than an adducted clam withoutmucous threads (9-mm clams adductedwithout mucous sank 22 mm in less than3 seconds; clams of the same size ab-ducted with mucous threads sank 22 mmin 6 to 12 seconds).The mucus produced was transparent

and difficult to see unless it had attachedwater column debris. The clams were

able to produce large amounts of thisviscous substance. Using Zenker's fluidto fix specimens that were floating, we

were able to identify the ctenidium as theexact source of mucous threads. After

1492

Fig. 1. Sequential representa-tion of the lift-off and floatingbehavior of Corbicula. Fromleft to right the sequenceshows initial position on sub-stratum, protrusion of foot,righting to perpendicular posi-tion with siphons protruded,initial lift-off with foot protrud-ed and exaggerated exhalentsiphon, floating with foot re-tracted but continued siphonalactivity, and initial resettle-ment on substratum.

fixation, specimens were embedded inparaffin wax, cut into 7-,um sections, andstained with toluidine blue. Large muco-

cytes, which react beta-metachromati-cally, were found densely packed alongthe inner demibranchs of the ctenidia(Fig. 3). Apparently the water currentstimulus prompted the expansion of theexhalent siphon and corresponding ac-

tive pumping of water across the gills.The inner demibranch mucocytes secret-

Fig. 2. Carbon black-stained mucous thread

of recently floating Corbicula still attached to

clam that has adducted.

Fig. 3. Histological longitudinal sectionthrough posterior ctenidial inner demibranchshowing dense banks of large mucocytes.Sections through typical gill filaments areshown in the upper left. Zenker's fixative,toluidine-blue stain, and 7->Im section. Hori-zontal field width, 610 ,Lm.

ed large amounts of mucus that wastransferred, along with the pumped wa-ter, out the exhalent siphon. As long asthe valves remained abducted, the con-nection between gills and mucous threadremained intact. The delay in initial lift-off (2 to 8 minutes) reflected the timerequired for the clams to produce suffi-cient mucus for adequate increase infrictional drag. The buoyancy of theclams may be a reflection of the lowdensity of this hydrophilic substance.The widely abducted valves, maintainedduring floating, and the extended footmay also serve to increase the surfacearea exposed to water currents and thusto increase drag.

Flotation of small Corbicula was con-firmed in a field test in the Black Rivernear Brooklyn, Mississippi, in earlyApril 1984. Several small specimens (7 to14 mm long) were collected along a grav-el bank and placed in a shallow enamelpan along the river edge. Currents cross-ing and entering this pan were between35 and 70 cm/sec. Within 10 minutes,several of the clams produced long mu-cous threads which extended from theirexhalent siphons; these clams then float-ed out of the pan. The clams weretracked and seen to be transported only ashort distance (<2 m) before they wereentrapped along the river bottom byblocking gravel. While moving along thebottom they bounced on and above thegravel substratum until being wedgedwithin a crevice. We are uncertainwhether the clams repeated this type ofmovement, traveling downstream in apunctuated fashion. On smooth sandsubstrata the clams would be carriedfarther.Many species of gastropod molluscs

are known to raft or float on mucousbubbles [for example, Janthina (10),Helcion (11), and Hydrobia (12)].Stressed and moribund Corbicula arebelieved to crawl to the substratum sur-face and, because of the production ofgases during decomposition of their softparts, become buoyant (13). These clamsmay retain the gases between adductedvalves sealed with mucus (13). Ours ap-pears to be the first report of a healthyfreshwater bivalve using mucus as a wa-ter-column dragline for flotation. This isimportant for Corbicula in view of itsdispersal and invasive capacities.The dispersal of Corbicula larvae is

well known, and pediveligers are fre-quently taken in plankton tows (1). Lar-vae are produced in large numbers [forexample, 400 larvae per clam per day inLake Arlington, Texas (14)], but theselarvae and subsequent juveniles are sub-ject to high mortalities (1). The hermaph-

SCIENCE, VOL. 225

on

July

13,

201

1w

ww

.sci

ence

mag

.org

Dow

nloa

ded

from

roditic Corbicula attains sexual maturitywhen it reaches 7 to 10 mm in length (3).Histological sections of clams in thisstudy also revealed mature ova in clamsas small as 6 mm. That the clams mayreach sexual maturity at this size range,in conjunction with their ability to dis-perse by flotation, indicates that rapidpopulation blooms could occur in areaspreviously devoid of Corbicula. Also,because they reach maturity in 4 to 6months (1) and because of their broodingcapabilities, gravid or potentially brood-ing adults can quickly populate a newarea rather than there being a delay inpopulation growth resulting from larvalsettlement.There have been sporadic reports of

juvenile Corbiciula composing importantparts of the drift fauna of variousstreams. Numerous small Corbicula (3percent of the total drift fauna) wererecovered during a study of the lowerMississippi River (15). These clams wererecovered in particularly high concentra-tions in June and August, times that mayreflect reproductive peaks for this clam.Small adult or juvenile clams may settleout from a flotation mode in a relatively(or at least temporarily) static area,where they would fall out of the drift.Thus rapidly growing juveniles and ma-ture but small adults could populate arelatively calm area. With the high de-gree of fluctuation seen in southernNorth American streams, the clams maybe carried downstream from this staticarea in spurts and bounds.

Reports of population dynamics froma single area must be regarded with care,since the populations reported may notrepresent a single grouping but may in-stead be the result of multipopulationinvasions over relatively short periods oftime. This type of transport may bereflected in the variety of size classesseen in a given lotic population. In someMississippi streams, Corbicula popula-tions usually break down into three dis-cernible size classes. Typically, howev-er, there are anomalous intermediatesizes, which have frequently been allo-cated to individual growth variations,We cannot any longer assume that theseintermediates are part of a single popula-tion. Instead, they may reflect a growthstage from a different upstream popula-tion that recently drifted in. Also, smalladults may drift together and settle in analready populated environment, adding asecondary and artificial growth stage forthat given locality.

Britton and Morton (3) maintain that"the genus Corbicula contains speciesprimarily adapted to flowing water envi-ronments . . periodically subjected to

28 SEPTEMBER 1984

seasonal flooding." Flooding may allowfloating clams to transfer stream sys-tems. In view of this potential mode oftransport, we must add active dispersalbehaviors to the morphological andphysiological invasive advantages thatthis "weed" species has evolved.

ROBERT S. PREZANTKASHANE CHALERMWAT

Department of Biological Sciences,University of Southern Mississippi,Hattiesburg 39406-5018

References and Notes

1. J. C. Britton, in Proc eedings of the SymposiumofRec ent Benthologic al Investigations in Texasand Adjacent States (Texas Academy of Sci-ences, San Angelo, 1982), pp. 21-31.

2. R. M. Sinclair and B. G. Isom, Further Stuidieson the Introduceed Asiatic Clam (Corbicula) inTennessee (Tennessee Department of PublicHealth, Nashville, 1963).

3. J. C. Britton and B. Morton, in Proceedings,

Brain dysfunction or brain damage hasbeen observed with the use of neuropsy-chological and neuroradiological tech-niques in chronic alcoholics (1). Studiesof evoked brain potentials (EP's) havedemonstrated a number of functional ab-errations in chronic alcoholics (2). Sev-eral investigators have studied auditorybrain stem potentials in chronic alcohol-ics and have reported electrophysiologi-cal evidence of increased neural trans-mission time (3). Moreover, event-relat-ed potential (ERP) studies in chronicalcoholics have demonstrated deficits inthe P3 component with the use of infor-mation processing paradigms (4). Thepresence of these deficits in the centralnervous system has been presumed toreflect the consequence of chronic alco-hol abuse (toxic effects of alcohol on thebrain, nutritional deficits, or an interac-tion of alcohol and nutrition-related fac-tors). Although the neurophysiologicaldeficits observed in chronic alcoholicsare presumed to be alcohol-related ef-fects, it is possible that some of thesedeficits may be present in subjects athigh risk for alcoholism and thereforeantecede the onset of alcohol abuse.

Genetic factors may be involved in thedevelopment of alcoholism. Sons of al-coholic fathers represent a special groupat high risk for developing alcoholism (5)

First International Corbicula Symposiuim (Tex-as Christian University Research Foundation,Fort Worth, 1979), pp. 249-288.

4. Human transport includes spread by fishermenusing the clam as bait, as a food source by earlyChinese immigrants, and by pet shop dealers (2,3).

5. R. F. McMahon, Nautilus 96, 134 (1982).6. L. L. Eng, in Proceedings, First Iinteirnatiotial

Corbicula Svmposiufm (Texas Christian Univer-sity Research Foundation, Fort Worth, 1979),pp. 39-68.

7. G. M. Thompson and R. E. Sparks, Am. Midl.Nat. 98, 219 (1977).

8. G. L. Mackie, Bull. Am. Malac-ol. Union Inc.1979, 17 (1979).

9. N. I. Platnick, Svst. Zool. 25, 101 (1976).10. J. E. Morton, MIluscs (Hutchinson, London,

1979), p. 40.11. 0. Vahl, Sarsia 68, 209 (1983).12. R. Newell, Proc. Zool. Soc. Londotn 138, 49

(1962).13. R. F. McMahon, The Mollusca (Academic

Press, Orlando, Fla. 1983), pp. 505-561.14. D. W. Aldridge and R. F. McMahon, J. Mollus-

can Stud. 44, 49 (1978).15. A. Obi, thesis, Louisiana State University, Bat-

on Rouge (1978).

16 April 1984: 16 May 1984

even when they are separated from theirbiological parents soon after birth. Stud-ies of male adoptees indicate that thebiological rather than the adoptive par-ent is predictive of later drinking prob-lems (6). Further evidence for a geneticpredisposition comes from twin studiesindicating that the concordance rate foralcohol abuse among identical twins isalmost double the rate for fraternal twins(7); patterns of alcohol consumption arealso highly concordant among identicaltwins (8). This evidence suggests that agenetic factor may be involved in thepresence of neural pathophysiology as-sociated with alcohol abuse.The identification of a suitable biologi-

cal marker that is genetically transmittedis important in identifying individualsbefore the onset of the disease. More-over, biological markers can providefundamental data on the etiology of alco-holism. The search for such a markermust focus on a biological variableknown to be genetically determined andprevalent in abstinent chronic alcohol-ics. There is good evidence to indicatethat EP waveforms are genetically deter-mined. Monozygotic twins manifest EPwaveforms that are as concordant witheach other as those obtained from thesame individual tested twice (9).We now report the presence of P3

1493

Event-Related Brain Potentials in Boys at Risk for Alcoholism

Abstract. Recent neurophysiological findings have demonstrated that abstinentchronic alcoholics manifest deficits in event-related brain potentials. To explorepossible biological antecedents of alcoholism the present study examined boys athigh risk for alcoholism. Event-related brain potentials were recordedfrom biologi-cal sons of alcoholic fathers and matched control boys. Differences in the P3component of the potentials were obtained between the high-risk and controlsubjects.

on

July

13,

201

1w

ww

.sci

ence

mag

.org

Dow

nloa

ded

from