Historical and Ecological Factors in the Evolution, Adaptive

PDF document for free

1. PDF document for free

31496_722_2_375.pdf

AMER. ZOOL., 22:375-395 (1982)

Historica

l an d Ecological Factor s in the Evolution,

Adaptiv

e Radiation , an d Biogeograph y o f

Freshwate

r Mollusks 1 GEORG

E MORGAN DAVIS

Academy of Natural Sciences of Philadelphia, Philadelphia, Pennsylvania 19103

SYNOPSIS

. There are some 36 families that are wholly freshwater or with representativespecies i n freshwater. There are virtually no phylogenetic analyses for these families.Zoogeographic analyses o f freshwater molluscan faunas are hindered by a lack of signif- ican t systemati c studie s of these faunas. Such studies are essential if one hopes to develophypotheses about phylogeny o r biogeography. I t i s clear from a phylogenetic analysis of the Pomatiopsidae that phylogenetic, vicari-ance, dispersal, an d ecological factors all have significant effects on the patterns of dis-tribution o f this family. At one stage in history vicariance may be a dominant factor whileduring another stage o

f history, dispersal may be a dominant factor. At every stage,ecological considerations are necessary to understand the phenotypes seen and the spatialrelationships among taxa.

I n examining the distribution patterns of dominant freshwaterfamilies with regard t o their biological, ecological, and overall phylogenetic relationshipsit i s evident that ecology plays a major role along with dispersal and vicariance. Clearlya synthesis i s needed in biogeographical studies that incorporates vicariance dispersal,ecology, an d geology-paleontology.

INTRODUCTIO

N Fo r som e year s I have been attemptingto understand something about the origins,pathways o f evolution, modes and temposof evolution, adaptive radiations, an d dis-tributions o f freshwater molluscan faunas.In my pursuits i t has become increasinglyclear that i n beginning a biogeographicalanalysis involving organisms about whichlittle i s known, one does not start with a singl e hypothesi s (onl y vicariance ) or witha single analytic tool fo r th e systematicstudy (only multivariate phenetics b y

UPGMA)

. So diverse are the ways in whichorganisms ca n respond to environmentalchange an d s o diverse are the geological-ecological changes through time a s theyimpinge o n diverging grades and clades,that there i s n o a priori way of knowingthe initial boundaries within which t o as - ses s relationship s or origins. If one is in- tereste d in a group of taxa, for whateverreason, i n th e absence of a data base suit-able fo r phylogenetic analysis one must 1 Fro m the Symposium on Alternate Hypotheses in Biogeography presented at the Annual Meeting of theAmerican Society o f Zoologists and the Society of Sys- temati c Zoology , 27-3
0 Decembe r 1980
, at Seattle,Washington. begi n by collecting data on all their closerelatives i n faunas throughout the world. I n thi s pape r I use as my major examplethe analysis o f the origin and evolution of th e Pomatiopsida e wit h it s tw o subfamilies :

Triculina

e an d Pomatiopsinae (Davis,

1979a)

. In brief, this family is consideredto have a Gondwanaland origin with vicar-iance accounting fo r the relict elements in Sout h America , Sout h Africa , an d Austra - lia , th e deliver y vi a th e India n Plat e t o con - tinenta l Asia. 2 Additionall y ther e wa s lon grange dispersal through China t o Japan,the Philippines, t o th e United States (Fig. 2 Th e origi

n of continental S.E. Asia is complex.There is evidence that the Thai-Malay peninsular sec-tion separated from northern Gondwanaland

mid -

Palaeozoi

c an d tha t collisio n wit h mainlan d Asi a wa sin the late Triassic (Ridd, 1980). This early accretionof a "microcontinent" to mainland Asia does not alterthe scenario fo r th e introduction of the Pomatiopsi-dae t o Asia from the Indian Plate. The Indian Platehad a profound impact on Tibet, Burma and westernChina i n th e Tertiary at a time coincident with the availabilit y of freshwater snails of hydrobioid gradeorganization. Th e mid-Palaeozoic considerably pre- date s freshwate r hydrobioids . The Ridd (1980) re- por t provide s a strong reminder that with each newgeological find involving land mass movements an d formulations , biogeographi c hypothese s mus t b e re- evaluated . 37

5Downloaded from https://academic.oup.com/icb/article/22/2/375/2015964 by guest on 15 August 2023

376GEORGE MORGAN DAVIS

POMATIOPSINA

E

Blanfordia

. 1 Cecin a 2

Coxiell

a 3 Fukui a 4 On e

Pomatiopsis..

. 6

Tomichi

a 1

Aquidauama

- 8 FIG . 1. The current distribution of pomatiopsid species, the hypothesized area where the Pomatiopsidae flourished ove r 10 0 millio n year s ago , an d th e pathwa y o f distributio n followin g th e breaku p o f Gondwana - land . (Fro m Davis , 1979a
) 1)

. One subfamily, the Triculinae, under-went an explosive adaptive radiation in theMekong River, yielding three tribes, 11genera and at least 92 species within thepast 10 to 12 million years.

I n th e first hal f o f thi s pape r I as k gen

-eral questions pertaining to biogeograph-ical studies including my own approach,operational methods of systematic analysis,the relationship of systematic studies tobiogeographic hypotheses, and constraintsin structuring a phylogeny. I then provideexamples relevant to these general issuesfrom my studies of the Pomatiopsidae.The second half of the paper involves dis-persal, vicariance and ecological consider-ations.

M y genera l approac h t o thes e issue s an

dmy understanding of constraints follow: 1)Certainly evolutionary and biogeographi-cal hypotheses should be falsifiable if theyare to be valid (Popper, 1960, 1968a, b\Wilson, 1965; Platnick and Gaffnev, 1977).2) Sound biogeographic analyses are de-

penden t o n soun d systemati c analyses . To

-day systematic analyses should be donewithin the framework of population ecol-ogy, population genetics and modern eco-logical theory (Davis, 19796, p. 161). 3)Only after a thorough systematic study isit possible to attempt to assess phylogenyand in many, if not most cases there willbe insufficient data to make an

objective hypothesi s abou t phylogen y (Cai n an

dHarrison, 1960). 4) Recent literature con-cerned with methods in systematics andbiogeography is full of polemics and someread like religious dogma. To quote El-dredge (1979): "The dialogue has beenmarred at times by shouting (as a substi-tute for careful thought), confusion of ex-position (with consequent misunderstand-ing on the part of opponents), and atendency to characterize the opposition inmonolithic terms. There has been as muchtilting at windmills as reasoned discourse."Downloaded from https://academic.oup.com/icb/article/22/2/375/2015964 by guest on 15 August 2023

BlOGEOGRAPHY OF FRESHWATER MOLLUSKS377

Avoi d dogm a an d conside r carefull y th

evalue of each position. In establishing po-sitions recent authors have frequently setup straw men to bring greater attention totheir points and often ignored early pa-pers in which the same issues are discussed(often with greater clarity). With time andnew discoveries, older and mistaken con-cepts are rapidly replaced with the newand more sound ideas. In so doing the bestof the old should be incorporated with thenew. 5) Vicariance, dispersal, and ecologyare all important in understanding the dis-tribution of organisms throughout the bio-sphere. 6) It is most desirable to test thecongruence between hypotheses aboutphylogeny and hypotheses of biogeo-graphical patterns. Phenetic and cladisticanalyses should be used to complementeach other. Cladistic analyses can be donewith different operational methods. Nosingle method is the best or only method.7) One should be wary of extreme models.There is more to biogeographic analysisthan looking for congruence of Hennig's(1966) complete-rule cladistic analysis anda vicariant model of area cladograms. Oneshould indeed examine the congruence ofa cladistic analysis and the temporal as-pects of geologic change if the data permitit. However, one should use all possibleoperations to test the soundness of resultsbased on the initial method, and oneshould consider the impacts of ecology anddispersal on present and past distributions.

SYSTEMATI

C STUDIES AND

BIOGEOGRAPHI

C HYPOTHESE S Whic h come s first : systemati c analyse

swith development of hypotheses aboutphylogenies, or biogeographic models?While the pros and cons of one or the oth-er of these positions have been argued(Rosen, 1975; Ball, 1975; McDowall,

1978)

, the answer clearly depends on thekinds of data available. One could arguethat biogeographic models for groups offish, mammals, etc. could serve to predictaspects of phylogeny of a group of fresh-water snails (concept of Rosen, 1975). Thismight be the case; it equally could bewrong. Each group must be considered interms of its own history only after sound

systemati c analyse s hav e bee n done . Ther e ar e virtuall y n o objectiv e phylo

-genetic analyses for groups of freshwaterMollusca (Table 1). Accordingly, oneshould not score geographic regions forpresence, absence, and/or diversity of afamily of freshwater mollusks with a goalof preparing a sophisticated biogeographicmodel, until one can be sure 1) that eachgroup concerned is monophyletic, and 2)that the systematic relationships to sup-posedly related groups in other localitiesare worked out.

Th e valu e o f havin g a t leas t a three-tax

-on data set for initiating or testing hypoth-eses concerning biogeography has beendiscussed (Platnick and Nelson, 1978; Ro-sen, 1978). Data available from three ormore taxa of different geographic locali-ties may make possible an initial hypothesisthat leads to targeting new localities andfaunas for future studies. Subsequent datacollection and reformulated hypothesesmay go through several cycles to producea satisfactory and congruent set of hypoth-eses for phylogeny and biogeography. Anyhypotheses about biogeography or phylog-eny for most freshwater molluscan groupsprior to the above steps would have thehighest probability of being wrong.

CLADES

, CLADISTICS, AND PHYLOGENY A s pointe d ou t elegantl y b y Cai n an

dHarrison (1960) and Sokal and Camin(1965) there are three components to aphylogeny: phenetic, cladistic, chronistic.A clade is a monophyletic assemblage(Huxley, 1959) and a cladogram is a hy-pothesis about phylogeny by ordering thebranching sequences (i.e., sequence ofclades). By ordering the branches, one im-plies direction of change through time.

Th e nee d fo r technique s i n phylogeneti

csystematics for formal analysis that wouldenable one to test hypotheses, and estab-lishment of criteria and techniques weresubjects addressed by Hennig (1950, 1957,

1965
, 1966), Cain and Harrison (1958, 1960)

, Wilson (1965), Sokal and Camin(1965) and others. With the English edi-tion of Hennig's (1966) operational criteriaDownloaded from https://academic.oup.com/icb/article/22/2/375/2015964 by guest on 15 August 2023

378GEORGE MORGAN DAVIS

fo r phylogeneti c reasoning , cladis m be

-came enthusiastically endorsed by manyand the polemics began. Cladistic analysisis essential to formulate an hypothesisabout phylogeny. A cladistic analysis is notnecessarily equated with Hennig's (1966)cladism. Further, we see that many excel-lent systematic works, pre-cladism, usedset theory as a mental process and uniqueand unreversed characters to group taxaoften without formally stating their tech-niques or without making a formal analysisthat would make their hypotheses falsifia-

ble

. With appropriate operational criteria,cladistic analysis had come into its own, butit is not necessary to denigrate excellentcontributions of the past. The practitio-ners of cladistic analysis face all the con-straints and problems faced by pre-cladis-tic systematists (Cain and Harrison, 1960).

Th e ga p betwee n collectin g a thoroug

hmorphological data base for systematicanalyses, and producing a phylogeny isenormous. Cain and Harrison (1960) havefully discussed the problem involved. Toestablish a clade, one must establish raono-phyly. To consider a group monophyletic,one must eliminate cases of convergence.Convergence is probably the most under-estimated problem in systematics (see also

Davis , 1979a), is widespread in mollusks,and may not be detectable in poorly knowngroups. Havin g identifie d convergen t taxa , th

etask of establishing clades arises. Manywho engage in cladistic analysis insist onselecting primitive character states and de-rived character states for those suites ofcharacters that permit discriminationamong sets of taxa. The problems with se-lecting primitive character states are con-siderable (Cain and Harrison, 1960; Cain,

1964

; Davis, 1980a; Davis and Greer,1980) and summarized with regard to as-sessing molluscan phylogeny by Davis(1981a).

Ther e i s littl e fossi l recor d preservin

gthe most important suite of characters forassessing phylogeny within subclass or su-perfamily level taxa of the Mollusca. Thesecharacters are from internal organs, thesoft parts. The shell is of little value be-cause of convergence (Davis, 1979a). Manyof the most useful characters have severalunordered states (Davis, 1979a). Given nu-

merou s unordere d multistat e characters

,each choice of the most primitive state andthe most derived state increases consider-ably the probability of error.

Ther e is , o f course , unconteste d approv

-al among malacologists of the primitivestatus of many gastropod character states.The three-chambered heart is primitive,two-chambered derived; the rhipidoglos-sate radula is primitive, taenioglossate de-rived; the zygobranch gill primitive, pec-tinobranch gill derived; ctenidiumprimitive, lung derived, etc. These doserve at present to separate subclasses andsome orders. The problem is selectingprimitive characters at the superfamily orlesser taxonomic levels. Consider thefreshwater Rissoacea (Table 1) with its sev-eral families where the Hydrobiidae are,after removal of the Pomatiopsidae, stillprobably polyphyletic, and where eachfamily may have derived from a differentmarine rissoacean group or where one oranother freshwater family evolved from aspecies of an existing freshwater family.There are at least 19 families with a ris-soacean grade of organization (Taylor andSohl, 1962; Davis, 1979a) about which wehave very little morphological detail. Weare a long way from designating primitiveand derived character states for the Ris-soacea.

A

N EXAMPLE: THE POMATIOPSIDAE

A n aquati c snai l i n th e Mekon g Rive

rand an amphibious snail of China have animportant characteristic in common: bothtransmit a species of

Schistosoma infectingman. The former is a species of Tricula(Triculinae), the latter is a species of On-comelania (Pomatiopsinae). The Triculaspecies is part of a large endemic radiationin the Mekong River. Both genera hadonce been classified as Hydrobiidae. Tri-cula has been classified as HydrobiidaeLithoglyphinae (Brandt, 1974). Theseopinions raised a series of questions (Davis

et al., 1976). How were these taxa related?Were they indeed Hydrobiidae? What canone learn of the origin of these taxa? Wasthe Mekong River assemblage that Brandt(1974) classified as Lithoglyphinae mono-phyletic?Downloaded from https://academic.oup.com/icb/article/22/2/375/2015964 by guest on 15 August 2023

BlOGEOGRAPHY OF FRESHWATER MOLLUSKS379

Anatomica

l studie s o f al l orga n system

sof the Mekong River taxa yielded data thatwhen compared with similar data for Hy-drobiidae s.s. of Europe, made the follow-ing clear: The Meking River taxa in ques-tion are monophyletic: the Triculinae. TheTriculinae and Pomatiopsinae are mono-phyletic: the Pomatiopsidae. The familyHydrobiidae of Brandt (1974) and pre-vious authors is polyphyletic, a worldwideassemblage of taxa grouped on charactersof shell, radula, penis, and operculum(Davis, 1979a). So similar were shells andradula of

Lithoglyphus s.s. of Europe andpomatiopsids called Lithoglyphus or Litho- glyphopsis o f China , a s wel l a s Lacunopsis o f th e Mekon g River , tha t the y wer e al l con

-sidered Hydrobiidae Lithoglyphinae(Brandt, 1974). Ensuing studies (Davis,1979a) made clear that Hydrobiidae s.s.are primarily northern continental, espe-cially European and North American,while taxa of similar grade of organizationof the Yangtze River drainage and drain-ages to the south and west of it includingall rivers of Southeast Asia and India, arePomatiopsidae. Convergence in a fewcharacter states had masked the consider-able differences between these two fami-

lies . Any biogeographical analysis involv-ing the Hydrobiidae sensu lato would havebeen in error. Wit h th e Hydrobiida e exclude d fro

mconsideration of the origin of the AsianPomatiopsidae, attention was focused ontaxa of similar grade of organization fromthe southern continents. On the basis ofanatomy, Tomichia of South Africa has aclose phenetic relationship to

Oncomelania o f Asia . Wit h th e discover y tha t Coxiella ofAustralia is more phenetically similar to

Tomichia

tha n eithe r i s t o Oncomelania, a tentativ e hypothesi s involvin g historica

lbiogeography was formulated. The Po-matiopsidae are of Gondwanaland originand the direction of evolution was fromGondwanaland to western Asia to Japan.Research was intensified on potential po-matiopsid taxa of South America, SouthAfrica, Southeast Asia, and Asia from In-dia to Japan. It was found that pomatiop-sid genera of South America, South Afri-ca, and Australia differed little phenetically,that derived character states doubled in

TABL

E 1. Families of freshwater molluscs exclusive of

brackish-water taxa and Opisthobranchia, and references for those few families in which phylogenetic analysis has been extensive enough to permit biogeographical analyses involv- ing origins, vicanance and dispersal. Clas s Gastropod a

Subclas

s Prosobranchi aSuperfamily NeritaceaFamily Neritidae

Hydrocenida

eViviparacea

Ampullariida

eViviparidaeValvatacea

Valvatida

eCerithiaceaPleuroceridaeMelanopsidae

Syrnolopsida

e

Thiarida

eRissoaceaAssimineidaeBaicaliidaeBithyniidaeHydrobiidae

Lepyriida

e

Pomatiopsida

e (Davis , 1979
, 1980
; Davis and Greer,1980)

Pyrgulida

eStenothyridaeBuccinacea

Buccinida

eVolutacea

Marginellida

e

Subclas

s Pulmonat a Orde r Basommatophor aFamily Acroloxidae

Ancylida

e

Chilinida

eLatiidaeLymnaeidaePhysidae

Planorbida

e (Hubendick . 1956
; Meier-Brook,1982 Clas s Bivalvi aSuperfamily UnionaceaFamily Etheriidae

Hyriida

eMutelidaeUnionidae (Heard andGuckert, 1971; Davis andFuller, 1981)Corbiculacea

Corbiculida

ePisidiidaeArcacea

Arcida

eMytilacea

Mytilida

eSolenacea

Solenida

eDreissenacea

Dreissenida

eDownloaded from https://academic.oup.com/icb/article/22/2/375/2015964 by guest on 15 August 2023

380GEORGE MORGAN DAVIS

numbe r wit h Oncomelania (Pliocen e o

fBurma, living in western China to Japan)and they doubled in number yet again inthe pomatiopsine radiation that began inthe late Tertiary of Japan. Upper Creta-ceous fossils of South Africa and northernIndia are strikingly similar and of pomat-iopsid phenotype. Tricula lives todayacross northern India. These facts coupledwith the late Miocene-Pliocene beds con-taining clearly pomatiopsid taxa in north-ern Burma justify an hypothesis of intro-duction of the fauna to Asia via the IndianPlate (Davis, 1979a).

Th e Cai n an d Harriso n (1960 ) criteri

afor discussing phylogeny were met as re-gards the Pomatiopsidae (Davis, 1979a,1980a, b, 19816; Davis and Greer, 1980).Two of the three triculine tribes, with themost derived taxa, are endemic in the Me-kong and possibly, to a lesser extent, in theYangtze River drainages. Taxa of one tribewith the most generalized taxa are widelydistributed from India throughout South-east Asia. There are a sufficient number ofunique and unreversed characters to es-tablish discrete sets of taxa. As a result, aphenetic analysis without weighting char-acters coupled with a timetable of eventsmade clear by geological and paleontolog-ical data made possible a phyletic topology(Fig. 2), the main branches of which areconsistent with a cladistic analysis (Davis,1979a, 1980a). A Prim Network superim-posed after multidimensional scaling(multivariate analysis, NT-SYS: pheneticmethod), rooted as a tree by establishingTricula as the most generalized taxon,yields a branching pattern consistent witha cladistic analysis using Wilson's (1965)operational method (Davis, 1979a, 1980a).With the discovery of the triculine genusRobertsiella in the Malaysian peninsula,sufficient qualitative data were availableto attempt a cladistic analysis based onthe selection of primitive character states.The cladistic analysis was to assess therelationships among genera most closelyrelated to Tricula {i.e., genera of thetribes Triculini and Lacunopsini). Criti-cal to the analysis was the primitive po-sition of the seminal receptacle (refer toDavis and Greer, 1980, for details). Asthere is a directional sequence of moves of

th e semina l receptacl e fro m th e burs a cop

-ulatrix out along the common sperm ductto the sperm duct to the oviduct, the ques-tion is: Which end of the sequence is theoriginal position? Two cladograms weredeveloped (using Hennig's 1966 opera-tional criteria) based on the two possiblechoices (Fig. 3). The "s" on the branchesindicated the potential of taxa in that cladeto transmit

Schistosoma infecting man ormammals (snails resistance to or potentialfor infection with this parasite are undersnail oligogenic control [Davis, 1980a]).Then a cladogram was developed usingWilson's (1965) operational method inwhich one does not select primitive char-acter states (Fig. 4). The congruence be-tween cladograms A of Figure 3, and Fig-ure 4 coupled with the fact that lineageswith genes associated with the transmissionof

Schistosoma are unbroken in A indicatesthat A is probably correct while cladogramB has been falsified. In summary, a clado-gram based on Wilson's operational meth-od of weighting unique character stateshas falsified one of two cladograms basedon choosing primitive and derived char-acter states.

I n summary , an d wit h consideratio n o

fthe general remarks made earlier: 1) Anybiogeographic model involving Tricula orOncomelania classified as Hydrobiidaewould have been in serious error. Thefamily Hydrobiidae sensu lato is polyphy-letic and many relationships implied bysuch a classification are incorrect. 2) A bio-geographical model of the Pomatiopsidaecould only come after a thorough system-atic analysis. 3) By identifying cases of con-vergence and establishing monophyly(clades) it became possible to initiate hy-potheses about the histories of thesegroups. 4) Combinations of systematicmethodologies involving phenetics andcladistics were useful in establishing cladesand relationships among clades. 5) It wasclear that the weakest approach to devel-oping clades would have been the methodinvolving selection of primitive character

states

. 6) The distribution of pomatiopsidtaxa and the direction of evolution interms of derived character states is con-Downloaded from https://academic.oup.com/icb/article/22/2/375/2015964 by guest on 15 August 2023

BlOGEOGRAPHY OF FRESHWATER MOLLUSKS381

12 0

CRETACEOU

S140 Fic

. 2. Phylogenetic diagram in three dimensions, with time given in millions of years (on a log scale) fromthe Jurassic to the present. Branching Points: 1. Triculine and pomatiopsine lineages established in Gon-dwanaland prior to the breakup of the southern continents. 2. Divergence to form the Jullieniini (left group-ing) in the Miocene. 3. Radiation of specialized

Lacunopsis (Lacunopsini), which diverges from the Triculini.

Lacunopsis, on shell characters, resembles marine and freshwater Neritidae. Some species converge on Anculosa

(Pleuroceridae)

, Littonna (Littorinidae), or Calyptraea (Calyptraeidae). 4. Seven genera evolved in the Miocene,probably much at the same time,

Pachydrobiella (PA) converges on Pachydrobia (PAC) of the Triculini in shellshape and structure. 5. Anatomical and shell data clearly indicate that

Hydronssoia (HY) and Jullienia (JU)diverged from a common ancestor. 6. A late Miocene radiation took place in Japan, giving rise to endemicgenera

Blanfordia (B) and Fukuia (F), and Ceana (C). Cecina spread to western North America, while Pomatiopsis

(P ) occur s onl y i n th e USA . 7

. Blanfordia and Fukuia have either diverged from a common ancestor or are thesame genus. Data thus far available support the former interpretation (see Davis, 1979a).

A . Aquidauania, South America. B. Blanfordia, Japan. C. Cecina, Japan, Manchuria, USA. CO. Coxiella,

Australia

. F . Fukuia, Japan. H. Halevnsia, Mekong River. HU. Hubendickia, Mekong River. HY. Hydronssoia, Mekon g River . JU

. Jullienia, Mekong River. KA. Karelatnia, Mekong River. L. Lacunopsis, Mekong River,Yangtze River. O.

Oncomelania, China, Japan, Philippines, Sulawesi. P. Pomatiopsis, USA. PA. Pachydrobiella, Mekon g River . PAC . Pachydrobia, Mekong River. PAR. Paraprososthenia, China, Mekong River (Thailand, Laos)

. S. Saduniella, Mekong River. T. Tomichia, South Africa. TR. Tricula, India, Burma, China, Philippines,Mekong River. (From Davis, 1979a)Downloaded from https://academic.oup.com/icb/article/22/2/375/2015964 by guest on 15 August 2023

382GEORGE MORGAN DAVIS

B FIG . 3. Cladograms showing possible relationshipsamong taxa of Triculini and

Lacunopsis. The clado-grams are based on characters of the female repro-ductive system. A is based on the concept that theprimitive position of the seminal receptacle is on theoviduct; B is based on a primitive position of the sem-inal receptacle attached to the bursa copulatrix. SeeDavis and Greer (1980:268-270) for details. E, Ha-

lewisia; F , Pachydrobia; G , Robertsiella; H , Tricula; H, , T. burchi; H 2 , T. aperla;), T. bollingi; K , Lacunopsis. A , B , C, D, H, I, Y, Z are ancestral taxa; s indicates thepotential of taxa in that clade to transmit Schutosoma. (Fro m Davi s an d Greer , 1980
) ;eo - gruen t wit h th e timin g an d locatio n o f g< logical-paleontologica l events .

Vicariance

and dispersal biogeography I s i t tru e tha t "i n biogeography , th e con

-trast is between the concepts of chance dis-persal and vicariance" (Croizat, 1978)? Are"the differences between traditional andvicariance biogeography . . . analogous tothe differences between phenetic and cla-distic taxonomy" (Nelson and Platnick,1978)? Is there really the polarity that onone hand the traditionalists cling to "de-termining centers of origin and directionof dispersal" while on the other hand the

FIG

. 4. Set-theory solution for demonstrating rela-tionships among taxa E-H, J-K based on unique andunreversed characters

sensu Wilson (1965). Charac-ter-states upon which each set is based are given inDavis and Greer, 1980. E,

Halewisia; F, Pachydrobia; G , Robertsiella; H , Tricula; J , "Tricula" bollingi; K , La- cunopsis. (From Davis and Greer, 1980) "vicarianc e approach " attempt s t o unit

e"phylogenetic theory, concepts of distri-butional congruence, and theories of earthhistory" (Rosen, 1978)? Must a discus-sion of dispersal be tied to centers oforigin? Is it true that taxa are distributedwhere they are today "because their ances-tors originally occurred in the areas wherethey occur today, and the taxa now thereevolved in place" (Platnick and Nelson,1978)?

Th e answe r t o al l thes e question s i s no

.Given historical developments and massiveevidence for plate tectonics, the tradition-alist approach to historical "centers of cre-ation," and routes for dispersal from thesecenters had to undergo radical reapprais-

al

. It is to make a straw man out of "tra-ditionalist" to pit the 19th century againstthe syntheses possible given from late 20thcentury facts about plate tectonics. Theconcepts of vicariance in terms of the im-portance of barriers, plate movements,Downloaded from https://academic.oup.com/icb/article/22/2/375/2015964 by guest on 15 August 2023

BlOGEOGRAPHY OF FRESHWATER MoLLUSKS383

an d theorie s o f eart h histor y wer e bein

gconscientiously applied (Scientific Ameri-can, 1952-1972) before the so-called vicar-iance paradigm started its exponentialphase of development in the decade of

1970

. The straw man problem is com-pounded with vicariance enthusiasts re-defining dispersal to suit their own ends.

I us e th e ter m dispersa l a s i t ha s lon gbeen considered,

i.e., in terms of 1) a taxonincreasing its range without crossing bar-riers and without fragmenting; 2) a taxonincreasing its range without crossing bar-riers but with subsequent fragmentation;3) a taxon increasing its range by crossingbarriers. It is self serving to state thattype-2 dispersal is "vicariance in disguise"(Platnick and Nelson, 1978).

A s discusse d above , phenetic s i s no t in

-compatible with cladistics; likewise I main-tain that both vicariance and dispersal areimportant to both understanding the ori-gin of new taxa and the distribution ofworld biotas. It is irrelevant to ask whichis the more important aspect. It is relevantto question the relative contribution ofeach to the origin and distribution of thetaxa in question. It is important to realizethat at one stage in the history of a stock,vicariance may be a dominant factor whileat another stage in history, dispersal maybe the dominant factor. Further explana-tions about biogeographic patterns maychange as one varies the time frame andgeographic scale (Simberloff et al., 1981).

Th e origi n an d evolutio n o f th e Pomat

-iopsidae (Davis, 1979a) provide an exam-ple for discussing these issues as they relateto freshwater mollusks. From the UpperCretaceous to the present, vicariance anddispersal have played different roles at dif-ferent times with regard to the subfamiliesPomatiopsinae and Triculinae. From theUpper Cretaceous to the Miocene, vicari-ance was a dominant factor with faunalelements rafted to their current SouthAmerican, South African, and Australianlocalities and an additional segment car-ried to mainland Asia via the Indian Plate.With the initiation of the Himalayan orog-eny began the formation of those majorriver systems that provided highways ofdispersal. The amphibious pomatiopsine

Oncomelania dispersed along the Yangtze,throughout southern China to reach Japanand eventually North America. The aquat-ic Triculinae dispersed along three arcs;northern India through China to the Phil-ippines; northern India through northernBurma to western Yunnan, China; downalong the Mekong River; and from north-ern India through northern Burma downthrough western Thailand into Malyasia(Davis, 1980a; Davis and Greer, 1980).

Concentratin

g o n th e Triculinae , i t i

sreasonable to consider three centers ofevolution (not to be confused with histor-ical centers of creation or centers of ori-

gin)

, i.e., localized areas in which evolvedspecies-rich faunas. Centers of evolutionare recognized by zones of high speciesdiversity comprising one or more mono-phyletic groups. I do not consider a centerof evolution simply to be a locality wherethe first fossil species, possibly attributedto the group of interest, is found. Speciesmay disperse to new drainages from suchcenters. Lineages diverging through timemay swell into species-rich centers at dif-ferent places and in different times. Thosecenters increasing in numbers of speciesand diversity of species are centers of evo-lution as within them and from them comethe various morphological or physiologicalvariations that lead to new lineages andperhaps new species-rich centers later on.The first triculine center is recorded in thefossil record on the Indian Plate in theUpper Cretaceous; the second center is re-corded in the late Miocene-Pliocene fossilbeds and existing lake faunas of northernBurma and western Yunnan, China; thethird is seen in the living fauna concen-trated within 300 river miles of the Me-kong River from Khemarat, Thailand toKratie, Cambodia.

Considerin

g th e first discernibl e center

,these ancestral populations were not divid-ed by the appearance of a barrier. Theunderriding of the Asian mainland andinitiation of the Himalayan orogeny al-lowed for dispersal from the Indian Plateinto newly forming aquatic systems. Inrapid order micro-vicariances and dispers-al were associated with stream captures,lake formations, lake basin disruptions,Downloaded from https://academic.oup.com/icb/article/22/2/375/2015964 by guest on 15 August 2023

384GEORGE MORGAN DAVIS

etc

. From this first recorded center, thefauna spread into virgin freshwater envi-ronments, so that diversity increased innorthern Burma and western Yunnan.

Wit h th e Himalaya n orogen y an d th

einitiation of the Yangtze and Mekongrivers in the late Miocene, there openedvast new ecological space for the radiatingpomatiopsids. As the rivers evolved, habi-tats and tributaries suitable for triculinesopened up further and further down-stream. The tributaries became highwaysfor colonization. Further downriver theriver became wider and ecological spacemore complex.

Today , th e greates t diversit y i n th

eworld of a monophyletic assemblage offreshwater snails is found in the third cen-ter of evolution. This assemblage of some92 species now occurs on either side of thegreat falls of the Mekong River near theCambodian border. Considering the biol-ogy of these snails, speciation could readilyoccur without macrobarriers (of the vicar-iance paradigm) dividing the ancestral bio-ta. The snails are small, mostly <12 mmmaximum shell diminsion, and r-selected(Davis, 1979a) in that they live one year or

less

, putting relatively immense amountsof energy into a one-time reproductive ef-fort. They are opportunistic species livingin a river that goes through an annualcycle of flooding and low water associatedwith the monsoon cycle. The annual cata-strophic floods cause immense density-in-dependent mortality. There are continu-ous shifts of habitats as the river destroysthem and creates them. During variousstages of low water various species grow tomaturity and quickly reproduce at slightlydifferent times. Thus several differentspecies may occupy the same space but atdifferent times (Davis, 1979a). Newly cre-ated habitats are rapidly colonized due torapid dispersal mechanisms. These areideal conditions for parapatric speciation(Endler, 1977).

Th e distribution s o f specie s an d popu

-lations in this modern center of diversityappear continuous viewed in terms of mac-ro-zoogeograph\ but break up into nodesof diversity and patchiness of populationsas one increases magnification of space.Close examination reveals numbers of con-

tiguou s o r barel y disjunc t denies . Ther e i

sprobably little or no gene flow amongmany demes over a number of years wherehabitats are infrequently destroyed byflooding. I have sampled such pocketsalong the Mekong River over five years. Inother localities habitats are regularly cre-ated and destroyed with low probability ofstability over periods of time such as five

years

. Recruitment into these low-stabilityareas presumably involves individualsfrom numbers of demes. Considering thehabitat complexity and width of the 300mile stretch of river that also includes ma-jor tributaries, and the patchiness andsmall area of species and deme patchesthroughout this area, it is highly probablethat speciation has taken place in this cen-ter as habitats evolved and diversity in-creased. Speciation could still be occurringsince in this evolving river system there isno ecological steady-state. Considering ar-guments of Ehrlich and Raven (1969) andEndler

(1973

, 1977), micro-allopatric orparapatric speciation can take place in avery small geographic area where geo-graphic ranges of closely related speciesare nested.

Som e triculin e gener a endemi c i n th eMekong River have more than eight to tenspecies and there are races of species

. Theimpression gained is that demes divergedinto races, races to species until within theecological space, stabilizing selection re-stricted speciation.

Wha t dispersa l mechanism s ar e there

?Dispersal across barriers leading to the in-troduction of freshwater molluscan faunato new aquatic systems can occur in several

ways

. All of these mechanisms were un-doubtedly involved in the formation ofspecies-rich pomatiopsid centers of evolu-tion. Consider the invasion of brackish-water Stenothyridae up the Mekong Riversystem to reach headwaters that occupy anenormous area while the Triculinae weredispersing down evolving river systems.The stenothyrids, like the triculines, aremostly negative rheotropic and thus pushupstream. Over time, waterfalls and rapidsdo not seem to dampen chance dispersalinto tributaries and headwaters. LocalDownloaded from https://academic.oup.com/icb/article/22/2/375/2015964 by guest on 15 August 2023

BlOGEOGRAPHY OF FRESHWATER MoLLUSKS385

endemis m o f th e stenothyrid s an d certai

nTriculinae in tributaries and certain head-waters attest to highly local speciation andlack of gene flow, or survival.

1

. The foremost mechanism is streamcapture. There is extensive geological evi-dence of stream capture in the upper Me-kong-Yangtze River drainages. A commonsequence would be a lake being tilted, andthe contents subsequently flowing into anexisting but previously separate stream;two faunal elements are thus joined.

2

. Faunal elements are swept down-stream throughout the year. I have col-lected algal masses floating downstreamfilled with 200+ individuals of four differ-ent species. I have collected sticks floatingdownstream, covered with typical triculinesand encrusted egg capsules. I collectedone such branch just after it was sweptover the falls of the Mekong. Maintainedin an aquarium, the eggs hatched yieldingover 100 healthy young snails. It is easy tosee how new habitats are colonized andhow a species can increase its range fardownstream.

Whil e suc h downstrea m movemen

tcould restrict parapatric differentiationand speciation along certain naturallychanneled sections of the Mekong River,other vast sections of river are so complexdue to thousands of islands creating nat-ural backwaters and barriers to dispersalthat conditions for parapatric speciationare found.

3 . The distribution of freshwater mol- luscs

, particularly pulmonate gastropodsand bivalves, has, in great part, been ac-complished by passive dispersal (Russell-Hunter

et ai, 1964; Lassen, 1975; Russell-Hunter, 1979). In England, it has beencalculated that isolated ponds receive anaverage of one species every 9 years (Boy-cott, 1930, 1936), and that species diversityin these ponds relates to the age of theponds. Lassen (1975) documented the in-troduction of pulmonate species into onesmall pond where he estimated that thepond equilibrated at 4 species with immi-gration and extinction rates at equilibriumof 0.8 species per year.

Th e transpor t o f mollusc s b y variou s an - imals , particularly by birds and insects hasbeen well documented (Kew, 1893; Rees, 1965

; Fryer, 1974; Russell-Hunter, 1979),and some distribution patterns of fresh-water molluscs reflecting bird migrationroutes (Russell-Hunter and Warwick,

1957

; Russell-Hunter, 1979). Fryer (1974)noted that 20% of the corixid bugs of anEnglish pond had one or more specimensof the tiny bivalve

Pisidium clamped to anappendage, a situation previously report-ed (Fernando, 1954). Given that these in-sects do fly long distances, it is clear that

Pisidium. transport by bugs is much moreprobable than generally thought. Smallerspecies of operculate snails are particularlywell-suited for transport by birds, and au-thentic cases of passive aerial transporthave been documented (Russell-Hunter etai, 1964). Limpets are transported on theelytra of aquatic beetles and

Hemiptera, andlymnaeids, unionids, physids, and landsnails have been commonly found on thefeet and plumage of birds, including birdsin long distance migration (Rees, 1965).Egg masses of physids can pass throughthe digestive system of ducks and a feweggs can survive to hatching (Malone,

1965)
.

Unioni

d clams , fo r th e mos t part , dis

-perse with fish since the glochidial larvaeof the clams are parasitic on fish. However,clams do clamp onto the feet of migratorywater-fowl. Kew (1893) documented a casewhere a duck was shot on the wing with aclam clamped onto a foot. The leg was cutoff and after a day placed into a pan ofwater upon which the clam opened itsvalves and released the foot. Transport ofamphibious or aquatic snails and/or eggson mud stuck to hooves or birds' feet is toohighly probable to reject (Lassen, 1975). Inshort, stepping-stone dispersal (fromaquatic system to aquatic system) leads tolong range dispersal given sufficient time.

I hav e see n bird s alon g th e Mekon g Riv

-er fly off with small sticks from the riverand fly upstream a considerable distancewith them. Presumably some of these stickswill fall back into the water. Examining thearea from which these sticks were gath-ered, other similar sticks were frequentlycovered with eggs. It should be remarkedthat during low water in this river, snailsDownloaded from https://academic.oup.com/icb/article/22/2/375/2015964 by guest on 15 August 2023

386GEORGE MORGAN DAVIS

cove r al l soli d object s wit h egg s t o th e poin tthat space for eggs appears to be a limitingfactor. I n th e pomatiopsi d dispersa l t o Nort hAmerica via Japan (Davis, 1979a), of spe-cial note is the amphibious habit of Asian

Oncomelania and American Pomatiopsis, thehigh density populations in which they oc-cur, and their proven ability to withstandlong periods of desiccation - all importantattributes for passive dispersal. The Po-

matiopsis radiation in North America is anextension of Oncomelania

hupensis of Japan(Davis, 1979a). The Pliocene pomatiopsidradiation of Japan produced the marinegenus Cecina that probably dispersedalong Beringia to the northwestern UnitedStates where it is also found today (Davis,1979a). Precisely how pomatiopsidsreached the Philippines is unknown butthe arcs of distribution are consistent witha hypothesis of dispersal mixed with somevicariance events.

I n summary , moder n pomatiopsi d di

-versity today is not where the pomatiopsidancestors occurred 10 million years ago.Tracing backward in time, we can identifya series of centers of evolution. Pomatiop-sid history is best interpreted in terms ofboth vicariance and dispersal. The roles ofvicariance and dispersal varied significant-ly over time. Passive dispersal by birds, in-

sects

, and fish plays an important role inthe distribution and local species diversityof freshwater molluscs. The explanationsdo change as one varies the time frameand geographic scale. Rapid and episodicspeciation occurring in a relatively smallarea (by continent size standards) coupledwith geographic expansion can be per-ceived as a blur of vicariance and dispersal.

ECOLOGICA

L DETERMINIS M Ar e ecolog y an d biogeograph y (base

don sound systematics) mutually exclusiveas stated by Rosen (1978)? I maintain thatthey are not, and that what is needed is asynthesis that links ecology, systematics,geological history and biogeography. Eco-logical considerations are essential forsound systematic studies, assessing distri-butions of living biotas, and analyzingchanges in distribution through time. En-vironmental equilibria and steady statesare short term at best even over periods of

geologica l tim e relevan t t o speciatio nevents, i.e., 500, 1,000

, 10,000 years. Con-tinuous environmental changes drive evo-lution. Diversity of organisms arises fromcontinuous adaptation to shifting selectivepressures.

Ecologica

l principle s ar e importan t t

osystematic analyses that are the founda-tions for biogeographical models, and tounderstanding the distribution of taxathroughout the world. First consider theecology-systematic link. Ignorance of, orneglect of ecological principles still con-tributes to serious errors in systematics offreshwater mollusks. Consider the resultsof ignoring the impact of both environ-mental variables and species interactionson convergence, polymorphisms, colorpatterns, character displacement, siblingspecies, adaptive radiation. The foremostproblem is convergence (Davis, 1979a) asdiscussed earlier. Genera around theworld have been placed into higher taxaon the basis of greatly convergent charac-ter states. Systematists have failed to real-ize that species of different phylogeniesoccurring in different areas often con-verge phenotypically in adapting to similarecological conditions. In Figure 5 areshown species belonging to three differentsuperfamilies, yet on the basis of shell phe-notype, phylogenetic relationships aremasked by phenotypes adapted to similarecological conditions, i.e., living on solidobjects such as rocks in high energy envi-ronments (Davis, 1979a). Considering theprominent degree to which shell pheno-type has been used in molluscan system-

atics , no further comment should be nec-essary. Th e systematic s o f th e pomatiopsi d ge-nus

Tomichia simply cannot be done on thetraditional basis of comparative anatomy.While species of this genus have remainedmorphologically similar, they have radiat-ed ecologically with different physiologicalrequirements. An understanding of theclimatic changes in South Africa from theMiocene onward is essential to under-standing the physiological-ecological ra-diation. With the onset of aridity in west-Downloaded from https://academic.oup.com/icb/article/22/2/375/2015964 by guest on 15 August 2023

BlOGEOGRAPHY OF FRESHWATER MOLLUSKS387

• FIG

. 5. Shells from three superfamilies demonstrating convergence. Rissoacea, Pomatiopsidae: Triculinae:Lacunopsini of the Mekong River; a)

Lacunopsis globosa (length = 11.0 mm), c) L. fischerpietti (diameter = 16.3 mm)

, d) L. harmandi (d = 8.1 mm), b) Littorinacea, Littorinidae: Littorina obtusata (1 = 13.7 mm) marine,northeast USA. e) Ceritheacea:

Spehia

zonata (d = 11.0 mm) Africa, Lake Tanganyika. (From Davis, 1979B)Downloaded from https://academic.oup.com/icb/article/22/2/375/2015964 by guest on 15 August 2023

388GEORGE MORGAN DAVIS

er n Sout h Africa , ther e ha s bee n

acontinuing climatic change towards desertformation (Tankard and Rogers, 1978;Axelrod and Raven, 1978). With thedrying up of perennial freshwater, snailsadapted or became extinct. We see todayspecies adapted to cycles of freshwaterevaporating to hypersaline water, and toamphibious, terrestrial, as well as complete-ly freshwater aquatic conditions (Davis,

19806

, 19816). An understanding of chang-ing ecological conditions from the Mioceneto the present is essential to understandingmodern and past distributions of SouthAfrican Pomatiopsidae.

No w conside r ecolog y an d th e distribu

-tion worldwide of freshwater molluscantaxa. By comparison with terrestrial andoceanic biomes, freshwater habitats aretransient short-lived systems. The totalarea of freshwater systems is not large(l/70th ocean's area; Russell-Hunter, 1979)and is broken up into small discontinuousunits (excepting the few large river sys-

tems)

. Additionally, most lakes and pondsaccounting for about 80% of the world'sfreshwater area are <100 km

2 (Russell -Hunter, 1979). The life expectancy forsuch lakes is only about lO^lO 4 year s wit hmany existing only 10 2 (Russell-Hunter , 1979)
. I n suc h system s (<10 0 km 2 , <10 3 year

sduration) one finds a considerable amountof between-systems population diversitybut low speciation. It appears that there isa premium on adaptive plasticity, dispersi-bility, and dampened speciation (Russell-Hunter, 1979). High diversity and consid-erable speciation are found only in thosefew systems (lakes and rivers) exceeding

10 6 year s i n ag e and/o r i n ne w ecologica

lspace opened by tectonic activity (Davis,1979a; Boss, 1979). Considering the con-straints of the above, and with regard tofreshwater dispersal, overall deployment,diversity and adaptive radiation, it appearsthat the following factors are important: 1)reproductive strategy, 2) respiration, 3)biological-habitat constraints, 4) size, 5)physical-ecological parameters.

I t i s clea r fro m Tabl e 2 tha t whereve r alarge endemic radiation is described, it isdominated by cerithiaceans or hydro- bioids , no t b y bithyniids , viviparids , pul

-monates, etc. (Davis, 1979a). The cerithi-aceans or hydrobioids occur in lakes orrivers and there is no geographic exclusionof one type or another,

i.e., all cerithiaceanradiations are not restricted to Africa, etc.Where one group radiates, the other is ab-sent or minimally present. Clearly thesefacts result from ecology, not history.

I t furthe r seem s eviden t tha t ecology

,not history is involved in the ampullariidand Viviparidae Bellamyiinae radiations,and the fact that cerithiaceans dominate inAfrica, not hydrobioids. The ampullariidsand Bellamyiinae are clearly of Gondwan-aland origin. The ampullariids have mod-ern radiations confined to tropical areas inSouth America, South Africa, India-Cey-lon, and Southeast Asia (Pain, 1972). Theviviparids are fossil in South America (Par-odiz, 1969) with proven Bellamyiinae ra-diating in Africa, India, and SoutheastAsia. Thus the individual track of the Po-matiopsinae involving Africa, India, andSoutheast Asia is part of a generalizedtrack involving the Ampullariidae and Bel-lamyiinae (also certain Thiaridae and Plan-orbidae Bulininae). Given that these dis-tributions were most probably derived byvicariant historical events involving platetectonics, the Ampullariidae and Thiari-dae radiated extensively in Africa, but notin the hotbed of endemism in SoutheastAsia where the hydrobioid Triculinae andViviparidae now dominate (Table 3). Therelatively great amounts of slow moving,muddy-bottom and muddy-water rivers orstreams and expanses of swamp land oftropical Africa may account for the largerbithyniid and ampullariid radiations ofAfrica. One innovation in the Ampullari-

idae

, the sinistral genus Lanistes accountsfor 18 (69%) of the African ampullariidspecies. This genus is not found in South-east Asia. On the other hand, viviparidsradiated more extensively in SoutheastAsia. Hydrobioids do not do well in Africain spite of the fact that the first freshwaterfossil record of hydrobioids is from thePermian of Africa (Knight et al., 1960).This lack of hvdrobioid radiation and di-Downloaded from https://academic.oup.com/icb/article/22/2/375/2015964 by guest on 15 August 2023

BlOGEOGRAPHY OF FRESHWATER MOLLUSKS389

TABL E 2. World-wide distribution of large endemic fresh- water gastropod radiations with taxa that are dominant in them in numbers of species.

Cerithiacea

n TABL E 3. Numbers of species of prominent entirely fresh- water prosobranchs of Africa and continental Southeast Asia.

Rivers

Tennesse

e system , U.S.A .

Coosa*

. U.S.A . Lakes

Tanganyika

, Afric a

Sulawes

i (Celebes ) Biwa , Japa n

Hydrobioi

d Lakes and Springs Rivers

Baikal

, U.S.S.R . Mekong , Southeas t Asi

aCuatro Cienegas, Mexico La Plata, South AmericaOhrid, YugoslaviaTiticaca, South AmericaPeninsular Florida, U.S.A.

* Som e 1 2 specie s o f th e hydrobioi d gener a Somat-

ogyrus and Clappia were presumably endemic in theCoosa River (Walker, 1918) but these were overshad-owed by the numerous pleurocerid species of this riv-er. Most of this endemic fauna is now extinct due todamming of the river.

versit y i s apparentl y du e t o a paucit y o

flimestone-rich outcroppings and suitablesubstrates in perennial rivers of tropicalAfrica. Hydrobioids do not thrive in mud-dy-bottom rivers, rivers with a pH <7.0,or in aquatic systems that dry out regular-

Bithyniids

, viviparids , an d ampullariid s hav e no t radiate d t o th e exten t o f cerithi

-aceans or hydrobioid prosobranchs pri-marily for reasons related to feeding and/or habitat restrictions. Bithyniids and vi-viparids are adapted for filter feeding(Cook, 1949; Lilly, 1953; Taylor, 1966a;Owen, 1966; Monakov, 1972; Meier-Brookand Kim, 1977) and hence considerable eco-logical space is apparently closed to them.The ampullariids are macrophagous herbi-vores with some carnivorous tendencies(van Benthem Jutting, 1956; Demain andLutfy, 1965; Demain and Ibrahim, 1969)but are amphibious with both pulmonaryand ctenidial respiration. Ampullariids,the largest of all freshwater snails, canspend long periods of time on land andwithstand long periods of desiccationwhen their source of water evaporatesaway.

Tax a o f thes e thre e familie s inhabi tquiet, muddy ponds, lakes, canals, and forthe ampullariids, paddy fields and swamps.

Centra

l to South Afric a

Southeas

t Asi a (continental )

Bithyniida

eViviparidae

Ampullanida

e

Thiarida

eHydrobioids

Stenothyrida

eBuccinidae

References

:

Brandt

, 197
4 an d Davis ,29 1 9 2 6 9 5 t o 10 0 -20 * 0 0 8 ; Boss , 1979
; Davis , 1979a
; 1979
; Brown, 1980; Davis, 1 2 4 2 5 2 9100
2 0 ! t o 1 0

Hoaglan

d 1980a
, b. * Exac t numbe r o f purel y freshwate r hydrobioi dtaxa not precisely known; 8 of the 20 are South Af-rican Tomichia (Pomatiopsidae). I n m y observations , th e bithyniid s tolerat

ethe muddiest environments followed bythe ampullariids with viviparids requiringthe least muddy conditions.

Viviparid

s worldwid e hav e no t rivale

dceritheaceans in numbers or nominalsubgeneric taxa and in no large geograph-ical area do they come close to rivaling hy-drobioids. In continental Southeast Asia,however, they do slightly exceed freshwa-ter cerithiaceans (Table 3). Viviparids haveprobably had greater success than bithy-niids and ampullariids in Southeast Asiabecause they tolerate less muddy condi-

tions

, can live in many rivers and lakes notsuitable for habitation by bithyniids or am-pullariids (too strong a current, and/or toolittle mud), and because at least some vi-viparids have been shown to supplementtheir diets by scavenging (Allison, 1942;van der Schalie, 1965). The viviparid Me-kongia is particularly successful in conti-nental Southeast Asia with seven nominalspecies and an additional five nominal sub-species of which five taxa live in the Me-kong River (contrasted with one species ofcerithiacean).

A numbe r o f familie s wit h representa

-tives in freshwater are marine-brackishwater groups with a limited modern inva-sion into freshwater, e.g., families of theNeritacea, Buccinacea, Volutacea, Arca-cea, Mytilacea, Solenacea, and Dreissena-Downloaded from https://academic.oup.com/icb/article/22/2/375/2015964 by guest on 15 August 2023

390GEORGE MORGAN DAVIS

cea . Som e rissoacea n familie s likewis e ar eincluded here: Assimineidae, Stenothyri- dae

. There are ecological reasons for thesuccessful, relatively species-rich stenothyridand buccinid radiations in freshwateras well as the limited marginellid, mytelid,arcid, dreissenid, and solenid incursionsinto freshwater. The freshwater steno-thyrid, buccinid, mytilid and arcid radia-tions as well as the single marginellid oc-currence are uniquely Southeast Asian andexplained by the reported dynamic sealevel changes over the shallow Sunda

shelf . Th e repeate d transition s fro m marin e t

obrackish to freshwater to brackish to ma-rine as continental Southeast Asia gainedand lost connection with Indonesia andBorneo is sufficient to explain the adjust-ments of some taxa of these families tofreshwater where river beds extend far outinto the vast Sunda platform. I know of noother place, including Brazil, where sealevel fluctuations have affected such alarge area with major river systems to pro-duce the kind of result seen in SoutheastAsia.

Endemis

m an d adaptiv e radiation s ar eassociated with tectonic events (Taylor, 19666

; Davis, 1979a). Tectonic eventsopen new ecological space and variouslyaffect molluscan populations. Spectacularfreshwater molluscan endemism in South-east Asia is associated with the Himalayanorogeny,

2 Pleistocen e se a leve l change

sover the Sunda platform, and climaticallyfavorable conditions. Areas of little or notectonic activity from the early Tertiaryonward have slight faunal changes and lit-tle endemism. An example is the Blancanand recent faunas of the southern GreatPlains of the United States (Taylor, 19666).

Th e biolog y o f certai n o f th e group s dis -cussed restrict them to different habitat types

. These constraints are important inaccounting for some of the biogeographi-cal patterns involving cerithiaceans, hydro-bioids, ampullariids, bithyniids, and vivip-

arids

. Let us now examine aspects ofreproductive strategy, respiration, size,and the combination of these factors onspecies diversity, patterns of distribution,and passive dispersal.

Pulmonate

s hav e no t radiate d t o th esame extent as prosobranchs. It has beencalculated that prosobranchs are 1.5 to 10 time s mor e endemi c tha n aquati c pulmo

-nates (Boss, 1979). As the vast majority ofpulmonates must obtain atmospheric oxy-gen to fill their lungs, they must live in theshallows of lakes, ponds, standing water,or pools of quietly flowing streams. Theycannot survive the turbulence of large

rivers

. For example, while pulmonatesabound in Thailand and Laos, they are to-tally absent from the Mekong River. Be-cause of ctenidial respiration prosobranchscan remain submerged and exist at consid-erable depths. Depth zonation is an im-portant niche variable exploited by proso-branchs (Boss, 1979; Davis, 1979a). Onlysome minute planorbids and lymnaeidshave succeeded in adapting to continuousimmersion at some depth (Boss, 1979;Russell-Hunter, 1979).

Anothe

r facto r i s reproductiv e strategy

.A single hermaphrodite or brooding fe-male of a monoecious species can establisha new colony (Lassen, 1975). Of impor-tance, also, are size and weight of an in-dividual in addition to habitat position andpopulation density relative to passive dis-persal already discussed. Pulmonates arehermaphrodites and live at relatively highdensities in the shallows of freshwater lakesand ponds that are among the most tran-sient and ephemeral of

biomes

. Pulmonatespecies exhibit remarkable tolerances tovarying environmental factors, well suitedfor the necessities of dispersal and coloni-zation (Russell-Hunter, 1979). Besides eco-logical tolerances, the short-duration tran-sience of most freshwater habitats clearlylimits the number of generations per pop-ulation, per system, and dispersal capabil-ities apparently restrict speciation (Russell-Hunter, 1979). Given all factors, it is notsurprising that pulmonate species are themost widespread and least speciose offreshwater gastropods. Such taxa areseemingly less prone to extinction thanspecies with limited ranges. Local environ-mental changes causing extinction of anendemic species within a zone of environ-mental change may eliminate only somepopulations of a wide-ranging species. Onthe other hand, as mentioned above, dis-Downloaded from https://academic.oup.com/icb/article/22/2/375/2015964 by guest on 15 August 2023

BlOGEOGRAPHY OF FRESHWATER MoLLUSKS391

persa l capabilitie s associate d wit h wide -spread taxa apparently dampen speciationwith resultant low species diversity amongsuch taxa.

Generally

, sexuall y reproducin g pul

-monates that can self fertilize have distri-butions > parthenogenetic prosobranchsand brooding bivalve Corbiculidae andPisidiidae, 5= brooding Viviparidae andcerithiaceans, and > egg-laying dioeciousprosobranchs. The same hierarchy holdsfor spatial extent of species distributions.Consider the wide-ranging distribution ofthe sexually reproducing pulmonates

Radix auricularia an

d Indoplanorbis exustus. Theformer ranges over Europe, Asia, India,Southeast Asia, and northeast Africa. Thelatter is widespread throughout India andSoutheast Asia.

Th e cerithiacean s Thiara scabra (Miiller)and Melanoides

tuberculata (Miiller) are par-thenogenetic (or at least have numerousparthenogenetic races) and brood theiryoung as do all

Thiara, Melanoides, and Bro- tia. The y ar e Gondwanaia n wit h a present

-day distribution in Africa, India, andSoutheast Asia; they have been further dis-tributed by man. Many Corbiculidae andPisidiidae have widespread distributionsbecause they are small, brood young, andare well-known to clamp onto animalagents of dispersal as discussed above. Therapid spread into, and colonization ofrivers throughout the United States in re-cent years (following introduction into thewest coast) of

Corbicula

manilensis can beattributed to effective dispersal and wideecological tolerances of this small broodingclam (Britton, 1979). However, in thehotbed of endemism in Thailand, Laos,Burma, Cambodia, Malaysia, Indonesia,and Vietnam, there are 36 nominal speciesof the bivalve family Corbiculidae and Pi-sidiidae of which 31 (86%) are endemic tothe region (Brandt, 1974). The few non-endemic species are all tiny

Pisidium (<5.0mm length, 8 species in this region); onespecies, P. annandalei Prashad, rangesfrom Sicily, Greece, Israel to SoutheastAsia. This regional diversity rivals thatseen in North America, i.e., 34 nominalspecies of Corbiculidae (Herrington, 1962).Colonization ability of the larger

Corbicula (mostl y >1 2 m m length ; rangin g t o 4 6 m

mlength) has not dampened speciation inthis region any more than is seen in theViviparidae and cerithiaceans.

Larg e outbreedin g an d broodin g snail

s(>20 mm shell length) are next in orderof extensive species distributions. Exam-ples are the cerithiacean

Brotia

costula (Ra-finesque) and Semisulcospira libertina(Gould). The former extends from theHimalayas of India through SoutheastAsia to Indonesia. The latter occurs in Ja-pan, the Ryukyu Islands, Taiwan, andsouthern China.

Hydrobioid

s ar e smal l (<1 5 m m shel

llength; mostly <10 mm) primarily aquatic,predominately outbreeding snails (exceptthe parthenogenetic Potamopyrgus). NoAsian or African hydrobioid has the ex-tended range of

Brotia costula or Semisul- cospira libertina. I n summar y i t i s eviden t tha t th e genera

lecological requirements required of themorphological-physiological ground-plansof the various families of freshwater snailshave a great deal to do with their diversityand distributions after the major vicariantevent of Gondwanaland disruption. Pul-monates are excellent colonizers of tran-sient habitats and have broad niches withregard to physiological tolerances. Con-versely, small outbreeding hydrobioids ap-parently can specialize in relatively smallareas (Davis, 1979a), have narrow nichesand narrow ecological tolerances. Ecolog-ical constraints associated with mode ofrespiration, habitat, and reproduction im-pact on speciation, dispersal and adaptiveradiation. Small brooding bivalves aregood colonizers and readily dispersed byvarious animal agents; they are more cos-mopolitan than non-brooding outbreedingprosobranchs.

Th e relativ e role s o f histor y an d ecolog

yare clear in some cases and not in others.Given the disruption of Gondwanaland,the role of vicariance is reasonably clear inunderstanding the South American-Afri-can-India/Asian distribution of Pomatiop-sidae (as discussed) and Ampullariidae,and the African-Indian-Southeast Asiandistribution of the Viviparidae Bel

Biogeography Documents PDF, PPT , Doc

[PDF] biogeographic zones

1. Science

2. Earth Science

3. Biogeography

[PDF] biogeographic zones meaning

[PDF] biogeographic zones of india

[PDF] biogeographic zones of india map

[PDF] biogeographic zones of india pdf

[PDF] biogeographic zones of india ppt

[PDF] biogeographic zones of india upsc

[PDF] biogeographic zones of world

[PDF] biogeographical

[PDF] biogeography a study of plants in the ecosphere

12345 Next 200000 acticles