DECEMBER 1990 Mpnroarer, LscruRr

ARTICLES

CLASSICAL TAXONOMY OF MOSQUITOES-A MEMORIAL TOJOHN N. BELKIN 1,'

THOMAS J. ZAVORTINK

Department of Biology, Uniuersity of San Francisco, Ignatian Heights, San Francisco, CA 94117-1080

ABSTRACT. The major contributions of John N. Belkin to the field of mosquito systematics arenoted. Hallmarks of his publications are presented and his hypotheses about the evolution, speciation,competition, distribution and classification of mosquitoes are summarized. The current status of thefield of mosquito systematics is examined. It is argued that mosquito systematics is still largely at thealpha taxonomy level, that at best only 25-50% of mosquito species are known, and that we have noteven begun to develop a natural classification for the Culicidae. It is concluded that there will be littleimprovement in the status of mosquito taxonomy unless there is increased support for systematics.

INTRODUCTION

The scientist we honor this year is John N.Belkin, one of the finest systematists ever towork with mosquitoes. He was a painstakingworker, a perfectionist if you wish, whose pub-Iications set the standard for accuracy and com-pleteness that others must now strive to reach.He made major contributions to the field ofmosquito systematics, contributions that willinfluence the field for decades to come. Amongthese are the development of: field and labora-tory techniques for studying mosquitoes; astandard descriptive terminology for all stagesof mosquitoes; an explicitly stated taxonomicmethodology for mosquitoes; a classificationscheme for all culicids; and hypotheses to ex-plain the evolution, speciation and distributionsof mosquitoes. Belkin amassed the finest re-search quality collection of mosquitoes in exist-ence. If it is properly curated, it will be a price-less heritage for future generations. In recogni-tion of his outstanding contributions tomosquito systematics, the American MosquitoControl Association awarded Belkin the Medalof Honor-the highest honor given by the So-ciety-at the Annual Meeting in Salt Lake Cityin April 1980. Two weeks Iater he died.

I entered UCLA as a graduate student insystematic botany in 1961. As soon as the oldergraduate students in botany learned that I had

'Twelfth Annual AMCA Memorial Lecture deliv-ered at the 56th Annual Meeting of the AmericanMosquito Control Association, Lexington, Kentucky,April 2, 1990.

2 Acknowledgments are made to the Lowndes En-gineering Company for their participation in sponsor-ing the 1990 Memorial Lecture.

an interest in entomology also, they told meabout a fantastic systematist in the ZoologyDepartment, a man who studied mosquitoes.That man was, of course, John Belkin. Since hiscourse in "Principles of Systematic Zoology" wasrequired of all students in systematic botany,the older graduate students were very familiarwith him and his work. I met John the nextyear, and after discussing his research with him,I decided to pursue my doctoral studies underhis direction. I finished my master's in botanyand joined his research group shortly thereafter.John suggested several projects that I could un-dertake for my dissertation. Among these werea c5rtotaxonomic study of the Anopheles speciesin California, a revision ofthe crab hole breedinggenus Deinocerites, and a revision of the treehole breeding genus Orthopodomyia. AII of thesewere projects in which John was interested. Be-cause of my fascination with tree hole mosqui-toes, I chose to work on Orthopodomyia. I com-pleted my doctorate in 1967. Instead of leavingUCLA, I stayed and worked with John on his"Mosquitoes of Middle America" project. Be-cause of my Iong association with John, it isperhaps appropriate that I present this memo-rial to him.

JOHN BELKIN AND HIS SCIENTIFICCONTRIBUTIONS

On the personal side, John was a kind andgenerous man. He could, however, be a difficultman to work for. He demanded a lot of himself,and he accepted no less of others. On occasion,he would be brusque, irascible, obstinate. Hisdemeanor and his encyclopedic knowledge wereintimidating to many. But if a person wasn'tdeterred by John, and proved himself to be

596 JouRNar, oF THE AurRrcarv Mosqurro Cor.nnor. Assocrerron Vor-. 6,

we did not yet know that mechanism for mos-quitoes.

cations. Some of these ideas or hypotheses weredeveloped by Belkin himself; some were bor-rowed from other biological disciplines. Someare controversial views that have not gainedwidespread acceptance among culicidologists orzoologists in general. All these themes demon-strate, though, his overriding concern withanalysis and explanation. One hypothesis thatBelkin developed was that the intercontinentalregions in both the OId World and New Worldtropics were "Centers of Origin" where majornew adaptive types of mosquitoes originated. Hesaw the intercontinental regions as geologicallyunstable areas where land bridees were formedand fragmented repeatedly. Duiing fragmenta-tion, populations of mosquitoes would be re-duced in size and isolated, creating the kind ofsituation that could lead to the rapid fixation ofnew adaptive types. These would become betteradapted with time, and would disperse to adja-cent regions when Iand bridges reformed. Belkinalso saw evidence for other methods of specia-tion in mosquitoes. He proposed sympatric spe-ciation as the explanation for the existence offlocks of closely related species in groups utiliz-ing specialized, restricted breeding sites such asleaf axils and pitcher plants. Without doubt, thesingle most controversial proposition advancedby Belkin was that hybridization was a majorfactor in the speciation and evolution of mos-quitoes. He saw evidence for hybridization intaxa of all ranks-species, subgenera, genera,tribes. At the species level, he considered hy-bridization as the most parsimonious explana-tion for the combination ofblocks ofcharactersfrom 2 species in the different sexes or stages ofa third species. In the case of higher taxa, hebelieved that hybridization in the distant pastwas the explanation for the annectent gtoupsthat exist today. He reasoned that taxa that hadformed in isolation on fragments of former landbridges may have been morphologically distinctbut not reproductively isolated, and that theymay have hybridized when the fragments re-joined. Belkin's familiarity with the role of hy-bridization in the speciation of flowering plantsundoubtedly played a part in his willingness toembrace this concept for mosquitoes. More than40% of flowering plant species are of hybridorigin. They are fertile tetraploids formed bydoubling the chromosome number of partly orcompletely sterile hybrids between 2 diploid spe-cies. Belkin believed that he saw the same kindof morphological evidence for hybridization inmosquitoes that botanists saw in plants. He oncetold me that the only distinction between thesituation in plants and the situation in mosqui-toes was that botanists knew the mechanism bywhich hybrid species of plants formed, whereas

Belkin was strongly influenced by ecologicalthought, particularly that related to competi-tion. He did not believe that there could be 2 ormore species of the same species complex breed-ing in the same habitat in the same area. In thesection on methodology in his Mosquitoes of theSouth Parific, he stated that he used data frommorphology, bionomics and distribution. Thebionomical data to which he referred were thehabitats of the immatures. The first steps in histaxonomic procedure were to sort material togeographical source and habitat. His belief that2 or more species of the same complex could notcoexist in the same habitat went a long way inpreventing the description of intraspecific var-iants, particularly hairy larval and pupalmorphs, as distinct species. Belkin also saw com-petition as shaping entire mosquito faunas. Hiscomments about the tribes Aedini and Sabethiniare relevant in this regard. The greater numberof sabethine lineages in the New World tropicsrelative to the Old World tropics could be inter-preted as indicating that the tribe originated inthe New World and radiated there in the ab-sence of competition from aedines. Alterna-tively, if the sabethines originated in the OldWorld tropics, which was, by the way, whatBelkin believed, then they had been largely re-placed there by the more recently evolved plant-breeding aedine mosquitoes.

In interpreting distributional data, Belkinconsidered common, widespread mosquitoes tobe the most recently evolved species in theirgroup. He envisioned them as arising in a sourcearea, which, for the faunas he was interested in.was an intercontinental area in the tropics. thenspreading into adjacent areas when

-dispersal

routes became available and eliminating previ-ously evolved species. Species with limited dis-tributions were considered to be relicts, or pa-leoendemics, particularly if they occurred in un-usual habitats for their group or occurred nearthe limits of distribution for their group. Belkinbelieved most of the mosquitoes endemic to theislands of the South Pacific were older thantheir relatives in the source area to the west.They were relicts whose absence in the sourcearea was due to extinction. Incidentally, Belkinconsidered the mosquitoes of New Zealand to bean ancient continental fauna, and the species ofthis fauna to be the most primitive living rep-resentatives in their phyletic lines to be foundanywhere in the world. This belief profoundlyinfluenced his ideas about the evolution of theCulicidae.

In the classification of mosquitoes, Belkinplaced emphasis on characters of the immature

MBtr.roRrel Lncrune 597DECEMBER 1990

stages rather than on the general ornamentationof adults or the secondary sexual characters ofmales. He believed larvae preserved phyleticcharacters better than adults. He went so far asto say that we needed to study all instars ofIarvae in order to develop a natural classificationfor the family. In the true mosquitoes Belkinrecognized 12 tribes. This arrangement was amarked departure from recognizing only 3 co-ordinate groups, the anophelines, toxorhynchi-tines and culicines. It was Belkin's conviction,though, that the unusual features of anophelinesand toxorhynchitines had been overemphasized,and that these gtoups were no more distinctthan many of the lineages traditionally placedin the culicines. Furthermore, he did not believethat the culicines were a monophyletic group. Inhis Mosquitoes of the South Pacific, Belkin pre-sented a list of the 12 tribes, placing the moregeneralized tribes centrally and the more spe-cialized and annectent tribes at the ends of theIist. This can be interpreted as a "phylogenetic

bush," with the more specialized tribes radiatingout in several lines from the central core ofgeneralized forms. Groups considered to beprimitive in more traditional classifications, Iikethe anophelines and sabethines, were considereihighly specialized tribes by Belkin. Although hisclassification is radical in its arrangement ofgenera, it continues the conservative traditionof recognizing only a relatively few genera formosquitoes,

CURRENT STATUS OF MOSQUITOTAXONOMY

Where is the field of mosquito taxonomy to-day? The taxonomy of every group passesthrough 3 overlapping and intergrading Ievelscommonly called alpha, beta and gamma tax-onomy. At the alpha level of taxonomy, the mainconcerns are the discovery, characterization andnaming of species. These species are placed intolarge, comprehensive genera. At the beta levelof taxonomy, the primary concern is the devel-opment of a natural classification. The speciesare studied in greater detail and are reclassifiedinto smaller and more numerous genera thatindicate their genetic relationships more accu-rately. At the gamma level of taxonomy, theprimary concerns are various biological aspectsof organisms, such as the structure of naturalpopulations, the genetic basis of phenotypiccharacteristics, the nature of isolating mecha-nisms, and the rates of evolution and speciation.This Ievel of taxonomy emphasizes the interpre-tation ofbiological diversity. So, where does thefield of mosquito taxonomy stand in this classi-fication of taxonomies? Sadly, it is still by and

large at the alpha taxonomy level. Emphasis isstill on recognizing, characterizing and namingspecies, Even many who utilize nonmorpholog-ical techniques, like electrophoresis, are alphataxonomists whose primary concern is the rec-ogrrition and characterization of species. Howmany species of mosquitoes are there? Howmany more remain to be discovered? The rateat which species are discovered in any group oforganisms can be depicted as a sigmoid curve(Fig. 3). The rate is slow initially, increasesexponentially as interest in the group rises,starts to slow when more than half the speciesare known, and finally approaches asymptoti-cally the total number of species in the group.The graph showing the total number of pres-ently valid species of Aedes known at 2l-yearintervals from 1750 to 1975 shows the expectedupturn, but gives no hint of slowing (Fig' a)' Abar graph showing the number of presently validspecies of Aedes described in each 5-year periodfrom 1900 to 1980 shows more variation in therate of taxonomic work (Fig. 5). Those periodsof time in which relatively few species weredescribed generally coincide with times of war-fare or economic depression. However, theperiod in which the fewest species were de-scribed is 1975-80. Does this slowing down in-dicate that we have found more than half thespecies of Aedes, or does it indicate only a de-crease in taxonomic effort? I suspect it showsonly the latter. If the rate of description of Aedeshas not started to slow and if our knowledge ofthis genus is representative of the Culicidae ingeneral, then we know less than halfthe speciesin the family at this time. Belkin believed thatthe 155 named zoological species of true mos-quitoes that he treated in the Mosquitoes of the

Time

Fig. 3. Sigmoid curve showing the increase in thenumber of species known for any group of organisms.The broken line represents the total number ofspeciesin the group.

-o

.9o

o!

E

F

598 Jounrqnr, oF THE AueRrctN Mosqurro Cournor, AssocrATroN V o L . 6 , N o . 4

mru7U

j zzo

! 6 @

.3 6m

3 sa 1 g

: 4 e

€ 3 s

e mE 2 €€ r m

1 4

@

0

Fig. 4. Number of presently valid species of Aedesknown at 25-year intervals from 1Tb0 to 1g7b.

179 1V5 1M 1e5

1915 1925 1935 19451920 J930 1940

Date

about 3,146. The square root of this is 56. The-oretically, for mosquitoes there should be 56monotypic genera and the largest genus shouldhave 56 species. How different is the real situa-tion! The graph for mosquitoes is almost astraight line (Fig. 7). We recognize only 3 mon-otypic genera and no more than 2 genera of anyother particular size. At the rieht end of thegraph, 6 genera with over 100 to nearly 1,000species each are way off scale. The graph formosquitoes is clear evidence that we have noteven begun to develop a natural classificationfor the family. Incidentally, an estimate of thetotal number of genera that should, theoreti-cally, be recognized in a family the size of theCulicidae can be obtained by integration of theequation XY : 56. The number is 225. Werecognize just 37 at this time.

Not all taxa are studied at the same rate.Although study of mosquitoes is still at the alphataxonomic level for the most part, there are somegroups where study has reached the gamma levelof sophistication. Included among t[ese are sev-

Speci€s per gsnus

Fig. 6. Hollow curve showing the decrease in thenumber of genera as the number of species per genusincreases for any group of organisms with a sound,natural classification.

1& 1875 tS 1925

l img

1S 1975

IgEO

90

o 8 o.93 7 0o 6 0

- 5 0

E 4 0

2 z a

20

1 0

0

Fig. 5. Number of presently valid species of Aedesdescribed in each 5-year period from 1900 to 1980.

South Pacific represented less than half thefauna, because practically no collecting had beendone in the interior of any of the islands. Be-cause Belkin described as new or resurrectedfrom synonymy 77, or 50%, of these 155 species,the number of species known from the regionbefore his studies may have been as little as 25%ofthe actual fauna. I would estimate, then, thatwe know at best only somewhere between 25and 50% of mosquito species.

In those groups of organisms that haveachieved the beta level of taxonomic work. ithas been noted that a graph of the number ofgenera to the number of species per genus is ahollow curve that approaches both axes asymp-totically (Fig.6). At one extreme, there are manymonotypic or ditypic genera. At the other ex-treme, there are only one or 2 large genera. Inthe middle, there are relatively few genera withjust a few species. Mathematically, this curve isa hyperbola with the equation XY : K. For anygroup of organisms, K is the square root of thenumber of species in that group. For mosquitoes,the number of valid species presently known is

2 5 3 5 4 5 S

S p € o i e s p e r g 6 n u s

45

65 100-1000 species

Fig. 7. Number of mosquito genera with the numberof species indicated.

DEoEMBER 1990 MEMORIAL LECTURE

eral species groups in the genera Aedes, Anoph-eles and Culex. lntercstingly, most gamma leveltaxonomic work done in the United States isdone by individuals who do not consider them-selves to be taxonomists. They are instead vec-tor biologists, geneticists, evolutionists or ecol-ogists. I used the genus Aedes as an exampleearlier, so let me use it again by saying that thevarious studies of groups of Aedes that havebeen done in the laboratories of George B. Craigand Karamjit S. Rai at the University of NotreDame are among the finest gamma level taxo-nomic studies ever done on mosquitoes.

FUTURE OF MOSQUITO TAXONOMY

What is the future for mosquito taxonomy?Systematics is a much misunderstood and ma-ligned science. As noted by Simpson years ago(1945), systematics is both the most elementaryand most inclusive of the biological sciences. Itis the most elementary because organisms mustbe discovered and named before we can studythem and record information about them. It isthe most inclusive because everything that iseventually learned about organisms is utilizedin systematics to help recogrrize taxa and under-stand the relationships among them- Too manybiologists have equated systematics with onlythe description and naming of species' Theyhave failed to see the central role of systematicsin the interpretation of biological diversity. Asa result, systematics has been in long-term de-cline. The field is understaffed and underfunded,and relatively few students are being trained'The situation in mosquito systematics is essen-tially the same as that in systematics as a whole'The scientific community seems to be awaken'ing to the fact that we know only a very smallpercentage of the species on earth. Some biolo-gists estimate that the 1.5 million species oforganisms we know are only 5% of the total thatexist. The scientific community seems also to beawakening to the fact that the number of sys-tematists is inadequate to perform the work thatis stil needed. In recent years a number ofscientists have called for national or global bio-logical surveys and have argued for gteater sup-port of systematics through the creation of po-sitions, the funding of research and the trainingof students. Many biologists attach great ur-gency to these matters because the growing hu-man population is accelerating the rate at whichhabitats are being destroyed and species aregoing extinct. Some estimate that 25 to 50% of.the species on earth will be extinct within thenext 30 years. Some even see this mass extinc-tion of biological diversity as a threat to thesurvival of human civilization.

The future will be, I believe, exactly what the

larger community of biologists wants it to be. Ifno value is given to systematic studies and thefield of systematics continues to be understaffedand underfunded, then there will be little pro-gress in mosquito taxonomy' There will be someievisionary work, but no major syntheses, nodevelopment of a natural classification and nomajor faunal studies. Gamma level taxonomicstudies will continue, but a point will be reachedwhere such studies will be hindered by the lackof a modern classifrcation for mosquitoes. If' onthe other hand, support for mosquito systemat-ics increases, then the kinds of studies that willlead to the development of a natural classifica-tion of the family will be possible. Faunal studieslike the "Mosquitoes of Middle America" andthe "southeast Asia Mosquito Project" shouldbe resurrected and continued. It is only throughmajor projects like these that the diversity ofmosquitoes becomes known. Taxonomic revi-sions produced by projects like these provide thedata that will ultimately lead to the developmentof a natural classification. That classificationwill, in turn, suggest innumerable hypothesesthat can be tested by studies at the gamma levelof taxonomy, and may provide some basis forunderstanding the interrelationships betweenmosquitoes and the disease-causing organismsthey vector.

CONCLUSION

In conclusion. let me note that we all standon the shoulders of our predecessors. The fieldof mosquito systematics is fortunate to have hada man with the intellect and motivation of JohnBelkin. He showed us the kind of studies thatcould be done and left us explicit instructionson how to do them. AII that remains is to com-plete the work that Belkin had hoped to dohimself.

ACKNOWLEDGMENTS

I thank Bruce F. Eldridge and John D. Edmanfor their suggestions in preparing the MemorialLecture, Sharon L. Belkin for providing photo-graphs, and Sandra S. Shanks and Mary AnnTenorio for preparation ofthe graphs.

REFERENCES CITED

Heinemann, S. J. and J. H. Belkin. 1978. Collectionrecords of the project "Mosquitoes of Middle Amer-ica." 10. Panama, including Canal Zone (PA,GG).Mosq. Syst. 10:119-196.

Simpson, G. G. 1945. The principles of classificationand a classification of mammals. Bull. Am. Mus.Nat. Hist. 85:1-350.