se a wedi otin Petont AD eatin aRGePEmife® Mh e Ree tR DAD otgatinnn eM 246 >The APTI a Rin. wR unRiahte bom Oo PE robe NAR REAL Ate SENNA Bh MARR Ba PRR OMG EME Pathe Baste mana FatsiAsin named AnPRLATASNAM A tonbalaacers\eTime A206 RRA SA MN STISA TS OI ce Pata tenn eter re re Eee aaa iad shesteenstn st eoamatut et cena ae ~ ede aenie eerie eeamerinir Sores ert aetna De ccherts eee ene ethrahe he ane Monae nerncht SEP NPC re aaeta name eee : en a ee en \elonniehsRAalinann anit shia menontindnsBakin eae f malign ett pA Ales Bin ieMintnaiatem vay Atn el and ontom Op Rn HH Nake G matter hy Rem vaneme mhiine! ans pM eRer te) aehfh ah ROA ARO nn sete Biceyinninek WA Aelia ®t Bnet arA te bad othe haw” AR EOE NA ae Emini an RAMEN EM mee REET RR. AE ni ate oP tne mk 8 e Cater Ran Ae nated le Ni eat a Ae an RAN. tates anime one Peediebo nine pm ma man taen NN RO RAR INR ag me Bema Oe NA ene R Ae nem eRl AN Tae at ee mats ne Ae Re aaNet eH eam AANA PR nn at nd A nian atumnble Pinar @dtnh, Sa ee eee eee ee eee Te ee ee ian ariesiom ete eslnl al itu tncetinznement-aVinciet pon ‘nnaett oninmeriaer pts at, Cai iets” Rant chA elie) 400 Mans aA 0 aetna o eFien er mE H meth geen nein aes tem ine nl ams oie nar ere eee eee et Pe ey eee ee " Ss want Speen ai aehiedntimiknauntn Catt aati en aS eee ts MES “ ee MER el ADR Ks Ee - \4, SBR >? M7! ey S wasn Oo is NS S Nase eins tino w3S7 Zz ’ a BRARIES SMITHSONIAN INSTITUTION NOILNLILSNI NVINOSHLINS S3ZIYVYUSIT LIB! NO] LIB NOI. LIB = o 2 g a = Zz = > ed = 2 2D rE fE — o os E = m = ; o = = w OILALILSNI NVINOSHLIWS SA!IYVYGIT LIBRARIES SMITHSONIAN INSTITUTION NOI NVINOSHLINS S31Y¥vVyusIt n” me = 22) & z rE WX < = = a 2 a = STA ere 49 Oo ze, DO We ®& Oo \ Tey 4,Jf SY WE ENQY 83 4 = > = Bea a a 7) . = ae BRARIES SMITHSONIAN INSTITUTION NOILNLILSNI NVINOSHLINS S3IYVYSIT LIB ” = 7) = a 7) = ul .o w a Qe m = m I om ry \ YY tc =, = = = = NN = 4 = ine $ oe S yILALILSNI” NVINOSHLIWS S3IYvVHAIT LIBRARIES SMITHSONIAN INSTITUTION NOIL LoS z — ie a z a = Me as = =) > A 5 z Ss = = es E a E cid fe ' ca m m m w : = a z 7) < BRARIES SMITHSONIAN INSTITUTION NOILNLILSNI NVINOSHLINS S3luvugi7a_LIBF . ~ SY » . ‘ “S NYINOSHLIWS NVINOSHLIWS SMITHSONIAN SMITHSONIAN iy nt NVINOSHLIWS NVINOSHLINS sa1uvuai LIBRARIES SMITHSONIAN INSTITUTION NOIL LIBRARIES, SMITHSONIAN — w — ~ (ep) Z rs = ul uJ n ee ” Wy Yn = al = | Ym, ~i a = cad 4 Y bes 4 . = > = > 4 7p) z w” Fad Tp) Fes BRARIES SMITHSONIAN INSTITUTION NOILNLILSNI NVINOSHLINS SaIYVYSIT LIBR wW NS FQ ~ a JTION NOILNLILSNI 4 ; = = i Lp LIBRARIES NOILNLILSNI MLALILSNI SMITHSONIAN StiTy wn QB) 4 Vv SAIuYVYUSIT LIBRARIES INSTITUTION NOILI a Sear oe SAV @8ll_liBRARIES ITION NOILNLILSNI Yd!i1 LIBRARIES i z 19 TION 4) Qo NO = x= a z J HITHSONIAN INSTITUTION NOILOALILSNI NVINOSHLINS S3IYVYUEGIT LIBRARIES. SMIT!I ~ = a z i Zz + f° e = ey; i o) 0] ke >] Vag Lif 2 > i e ee ip * - : ; z = & % D = es . 2 eB Z G INOSHLINS S3IYVNAIT LIBRARIES SMITHSONIAN INSTITUTION NOILMLILSNI NVING NVINOSHLIWS NS WRN SOSA AN SMITHSONIAN NVINOSHLIWS NVINOSHLIWS SMITHSONIAN * ITHSONIAN INSTITUTION NOILNLILSNI NYINOSHLINS S3IYVYAIT LIBRARIES SMITH Zz < z= Oo wn ae i = WY — (Zp) — (ep) —— w z z mS z aM a w RNS or wn “ za ox a NY SS oc et a +7Z = x =) NG SA Zz 4 aA G oc SEA SAN w e aw Yl: Zi fe) = ro) ES 5 Be 73 =! 2 = =J Zz INOSHLINS |S31YVYGIT LIBRARIES SMITHSONIAN INSTITUTION NOILALILSNI_NVIN > Pus _ 9S a ° Y, 2 o S Ne 5 © Gy © : aie A FE : © Lil z SS AS" 5 2 = A ae = = oe S\S on n°? 7 = o . SS z z be z oe ITHSONIAN INSTITUTION NOILMLILSNI NVINOSHLINS Saiyvugi7 LIBRARIES SMITI renames Ww = n ras aL & X = < = rs = XQ = Sj | rT WOW G apes O SESS \\ Se) E SO 2, Ee = EF wer 2 See ne 5 z 5 he INOSHLINS S31YVYGIT LIBRARIES SMITHSONIAN INSTITUTION NOILNLILSNI_ NVIN . \ ‘ LIBRARIES NOILNLILSNI NOILALILSNI LIBRARIES NOILNLILSNI AITHSONIAN INSTITUTION NOILALILSNI NVINOSHLINS S3IYVYEIT LIBRARIES SMIT bass x INSTITUTION saluvugiy INSTITUTION saluvuait INOSHLINS S3IYVYUAIT LIBRARIES SMITHSONIAN INSTITUTION NOILALILSNI NVIN NVINOSHLINS S31¥VvVual LIBRARIES INSTITUTION w za ae = ae Z = Z Pe = : by 2 D i if MY ® = = E 2 “iy = NTHSONIAN _INSTITUTION NOILNLILSNI NVINOSHIIWS SSUVE RPTL LIBRAR IES soMit z > : 0) Z “ = _ =r XE — + Tr ie, + SF oe _ QO Ss a _ ao “UZ S oc c aN ac Cc oc Gy = fre = NS oO = mo 4, * Zz i 2 eg = F INOSHLIWS _S4 1YVvuag ns LIBRARI Eee MS e ee rien NOILALILSNI _NWIN > Fe ‘> ; 1@) Gull = oO os 5 Sahin . » “ y od S ae i ani, Le ve ” lien sits i Peon 4 yt Pi | z tn i, J ry ae oe J or ’ mt | OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES No. 134, 175 pages, frontis., 91 figures, 26 tables December 22, 1979 The Biology and Physiology of the Living Coelacanth Edited By John E. McCosker and Michael D. Lagios California Academy of Sciences, San Francisco, California 94118 DE. Mg CPLIFOR SIQNAICP o S Npep w© SAN FRANCISCO PUBLISHED BY THE ACADEMY ne The Biology and Physiology of the Living Coelacanth ees hi ane ‘eid reo —S a CN JN As on /3ke PES L lis ABT ata. ——— \ Yili BS \\ > f ff ~~ SS FT Sw ae —_ ty "Because thts ts where the action ts aoing to be, Baby." Drawing by Robt. Day: © 1966 The New Yorker Magazine, Inc. OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES No. 134, 175 pages, frontis., 91 figures, 26 tables December 22, 1979 The Biology and Physiology of the Living Coelacanth Edited By John E. McCosker and Michael D. Lagios California Academy of Sciences, San Francisco, California 94118 CADEMy. gy * gx 2 E563 5 co oO O GP ONpED SAN FRANCISCO PUBLISHED BY THE ACADEMY COMMITTEE ON PUBLICATIONS Laurence C. Binford, Chairman Tomio Iwamoto, Editor Frank Almeda, Jr. William N. Eschmeyer George E. Lindsay (US ISSN 0068-5461) The California Academy of Sciences Golden Gate Park San Francisco, California 94118 PRINTED IN THE UNITED STATES OF AMERICA BY ALLEN PRESS, INC., LAWRENCE, KANSAS CONTENTS Introduction. By MICHAEL D. LAGIOS ‘and JOHN E. McCOSKBR .........4..46- NUMBER Ihe 2 “-. My story of the first Coelacanthh By M. COURTENAY-LATIMER ................ The influence of the Coelacanth on African ichthyology. Biya Vin G ATE MESS MID sh ook G/acoks va 805 se airs a ee eee . Inferred natural history of the living Coelacanth. By ORINVE SMCCOSIGERS 2:..sy.0 fos8s Sees soa he 6 he wee Hote ae nite See . The Coelacanth and the Chondrichthyes as sister groups: A review of shared apo- morph characters and a cladistic analysis and reinterpretation. ByaVilGHAre-D. DAGIOS SOR «tse ae ee ee Ss eee, er rae eee Coelacanths: Shark relatives or bony fishes? By LEONARD J. V. COMPAGNO, Wt MPAene Duy OY Mile EVAR ID ICAGIOS a... 0s © oo rs eaten rice aio ene Ee ee ee Ventral gill arch muscles and the phylogenetic relationships of Latimeria. JB Bek OR VA UBS BS eater ae toe kl eae Ogre Ie Mr RE RE SPREE ROAR dear GY AA Gar ob 018.60 Observations on the structure of mineralized tissues of the Coelacanth, including the:scalessand, their associated odontodes,” By W= A. MILLER ~.525.2225- 202-2 eee Mechanisms of osmoregulation in the Coelacanth: Evolutionary implications. ByAROBERT We GRIPPITH andsPETERK: TiePANG®: 22555 s220.62 bee een nee Some biochemical parameters in the Coelacanth: Ventricular and notochordal fluids. Bye Ost RAS MIUSSE'NE infos oe oe a Sale a Dis Oho cree he ee ee eee Chordate cytogenetic studies: An analysis of their phylogenetic implications with particular reference to fishes and the living Coelacanth. By GUIDO DINGERKUS... . Immunochemical and biological studies with growth hormone in extracts of pituitaries imoMlexistine primitive fishes. By TETSUO HAYASHIDA. .o. oe oe eee . Evolution of the creatine kinase isozyme system in the primitive vertebrates. By SUZANNE JE. EISHER andiGREGORY S. WEUIAM, «i: 2.5 235 scr atone eee . Amino acids and taurine in intracellular osmoregulation in marine animals. By J. B. LOMBARDINI, PETER K. T. PANG and ROBERT W. GRIFFITH ........ . Recapitulation of the discussion at the close of the 15 June 1977 symposium on the relationships of primitive fishes, with particular reference to the Coelacanth. ranscribediby SUSAN BROWN and'GEORGE BROWN ..%. - 23.9332 e502 2s oes Pages DEDICATION This volume is dedicated to the memory of Faith Atkins Breeden, a patient and persistent supporter of scientific endeavors. ACKNOWLEDGMENT Publication of this volume was made possible by a grant from the Charline Breeden Foundation. OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES The Biology and Physiology of the Living Coelacanth No. 134, 5 pages December 22, 1979 INTRODUCTION By Michael D. Lagios and John E. McCosker California Academy of Sciences, San Francisco, California 94118 Forty-one years ago today a most remarkable event occurred off the coast of South Africa. The discovery by a young woman scientist of a steely-grey, heavily-scaled, curious living fossil fish thought extinct for 70,000,000 years was the stuff of dreams. It was immediately heralded as the biological find of the century by adventurer and ichthyologist James Leonard Brierly Smith. Perhaps no single living organism has so dra- matically affected the public consciousness and scientific imagination as Latimeria chalumnae. For scientists it provided a unique opportunity to gaze backward at evolution, through its living tissues, at a lineage dating to the Early Devo- nian. For the rest of humankind, the celebration of the coelacanth in irreverent prose by Ogden Nash, in song by Charles Rand, through the vi- sual arts by New Yorker cartoonist Robert Day, and cinematographically through the *‘Monster From the Black Lagoon,’ bears witness to the appeal of this unicorn reborn. In addition to the excitement generated by the presumed antiquity of the living coelacanth was the immediate disposition of the scientific com- munity and the lay press to consider Latimeria the nexus to tetrapod evolution. In 1954 Science Newsletter wrote that “‘the coelacanth is be- lieved to have given rise to amphibians and to be the ‘missing link’ in man’s eventual evolution from sea creatures.”’ It is fitting that two score years since its celebrated discovery the coel- acanth undergo reappraisal. The present collec- tion of papers, the majority of which were pre- sented in an AAAS symposium in June 1977, provides an interpretive spectrum which will tantalize, though not completely satisfy, devout coelacanthophiles. Some of these works will re- affirm the established systematic position of the coelacanth while others are very divergent from the prevailing interpretation, suggesting a rela- tionship with a hypothetical paleozoic pre- chondrichthyean group. A wide spectrum of re- search methods and materials was used by the various contributors in their evaluation of the coelacanth. These range from novel re-evalua- tions of more conventional hard calcified tissues such as gill arches and teeth to newer ap- proaches using soft tissue morphology, compar- isons of the antigenic structure of peptide hor- mones and the definition of creatine kinase isoenzymes. This volume is based on tissues and data col- lected by the Expedition of the Royal Society (1969) to Aldabra and Cosmoledo and the French/British/American (1972) and the Califor- nia Academy of Sciences (1975) expeditions to Grande Comore. Popular accounts of the 1972 Expedition by Griffith (1973), Thomson (1973) nN OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, No. 134 FiGure |. Fishermen’s pirogues at rest in Iconi, Grande Comore. 5° hn a = J ; 7 i) * tee +s afta ; ’ eke ‘ = <¢_ _ alte a inte DOR WE es ees. FIGURE ; John E. McCosker and Michael D. Lagios dissecting a preserved coelacanth at the Veterinary Laboratory, Moroni, Grande Comore. Photo courtesy of Al Giddings. LAGIOS & McCOSKER: INTRODUCTION 3 eer wen FiGure 3. Members of the 1975 California Academy of Sciences Expedition (Daniel Robineau was not present). Top row: John McCosker, David Powell, Chuck Nicklin, Sandy McCosker, and Lester Gunther. Bottom row: Al Giddings, Sylvia Earle, John Breeden, and Michael Lagios. = COM RES é * , Pe . Ce. (OTS EXPRDITHONCDELACANTE FiGURE 4. Postage stamp issued in honor of the 1975 Expedition. 4 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, No. 134 FIGURE 5. examine the juvenile 1974 specimen and the eggs of the 1972 ovigerous female specimen. Photo courtesy of Al Giddings. and Anthony (1976) and of the 1975 Expedition by McCosker and McCosker (1976) provide in- sights into the difficulties encountered by each group such that the elusive goal of capturing, studying and returning with a living coelacanth remains unsatisfied. The California Academy of Sciences Expedition, supported primarily by the Charline Breeden Foundation, was successful in returning with two specimens which had been frozen soon after capture. Tissues from those specimens were distributed to thirty laboratories with the assistance of the Society for the Pro- tection Of Old Fishes. Given the present politi- cal situation in the Comores, it seems unlikely that such research materials will soon again be- come available, and for that reason we are even more grateful for the support provided by gen- erous benefactors and agencies. {t would be asking too much from such a di- verse field of approaches and interpretations to Daniel Robineau and Jean Anthony of the Paris Museum and Sylvia Earle of the California Academy of Sciences reach a consensus regarding the systematic po- sition of the coelacanth. What we have attempt- ed to provide for the readers is a unique oppor- tunity to field the question for themselves. We also feel that this volume will contribute signif- icantly through its historical articles by Mmes. Courtenay-Latimer and Smith and through the comprehensive listing of direct and indirect lit- erature citations which deal with coelacanths. In summation, we take pride in the knowledge that two score years thence, a more lucid view of Latimeria is surfacing. ‘‘Coelacanth’’ has found its place in public parlance, as fondly cited by the late Professor JLB Smith in telling the story of ‘‘a prominent member of the British Parliament who, in attacking an opponent, called him a ‘Coelacanth’ on the grounds that from his long silence in that august assembly it was a surprise to find him still alive,’ or as uti- lized by Time Magazine in describing Richard LAGIOS & McCOSKER: INTRODUCTION M. Nixon as a “‘Coelacanth of American anti- Communism.’ Yet the most astute observer of the enigmatic Latimeria was the late Ogden Nash. We are pleased to introduce this volume with his quatrain: It jeers at fish unfossilized As intellectual snobs elite; Old Coelacanth, so unrevised, It doesn't know it’s obsolete. LITERATURE CITED ANTHONY, J. 1976. Opération coelacanthe. Vivre et revivre Paventure 7. Collection ed. by A. and J. Arthaud. Librarie Arthaud, Paris. 201 pp. GRIFFITH, R. W. 1973. A live coelacanth in the Comoro Is- lands. Discovery 9:27-33. McCosker, S., AND J. E. McCosker. 1976. To the islands of the moon. Pac. Discovery 29:19-32. THOMSON, K. S. 1973. Secrets of the coelacanth. Nat. Hist. 82:58-65. OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES The Biology and Physiology of the Living Coelacanth No. 134, 5 pages December 22, 1979 MY STORY OF THE FIRST COELACANTH By M. Courtenay-Latimer Tsitzikama, South Africa Most ichthyologists believe that the modern coelacanth story began in 1938 with the discov- ery of a curious mauvy blue fish on the deck of Captain Goosen’s boat. Yet for me, the coel- acanth story actually started as far back as 1911, and were it not for an island with a strange at- traction and several men I met along the way, the coelacanth story might not have happened. At the age of four years I went to visit my moth- ers maternal home on the Addo heights from where at a certain time of the evenings the rays of the Bird Island Lighthouse could be seen re- flecting on the upstairs bedroom windows. Being nervous of the rays of light, my mother drew the curtains, and said “‘It is only the Lighthouse rays guiding the sailors and all people sailing on the dark seas.’’ This story of the light so im- pressed me, that the wonder grew. As I grew older I again visited the farm for school vacations and always my interest in Bird Island and the lighthouse stood out as a most intriguing feature, more especially as the young folk made special evening visits to a point to get a clearer view of the island and lighthouse. But children had to be seen and not heard and it was not until the age of 16 that I was allowed the privilege of joining the young people and really watching the lighthouse rays. Now an urge sprung up within me to visit the island. I felt I must get to the island. But why? No lady went to an isolated place like that! Six years later I became the appointed curator of the East London Museum at the princely sum of £2 per month. I was in my seventh heaven. I had always dreamed of working in a museum but here too, ladies were not accepted. However the Museum had been built for East London and I was appointed and expected to do the best for the official opening by the Cape Provincial Administration on 24 September 1930. The Museum collection was very sad; all kinds of odd specimens and six birds sadly eaten up by dermestid beetles. However I sorted and cleaned and then from home brought in the Lati- mer collections (for as a family we had amassed some good treasures). Now as a young and very enthusiastic worker my life was thrown into building and collecting for the Museum. Week- ends, public holidays and personal vacations were put to the benefit of the Museum. I inter- viewed fishing clubs and the manager of the fish- ing trawlers to save material for me, and in no time beautiful marine material was brought in. At this stage I first met Dr. JLB Smith. He vis- ited the Museum and was most intrigued with the mounts I had made of the small fish which I could handle. He as always was most encour- aging and gave me greater inspiration to do my very best in this work. During the following five years I worked un- der the guidance of Sir Henry Meir, Mr. D. Boyce, and Dr. Ernest Gill, learning about var- COURTENAY-LATIMER: FIRST COELACANTH ious aspects of museum work. My desire to visit Bird Island remained undaunted, until 1935 when I met a Mr. Patterson, Director in Chief of Islands off South Africa. No words could ex- press my excitement. But he was adamant and could not give me permission to go alone to Bird Island, BUT PROMISED if I got one other older lady to go with me he would grant me permission to visit Bird Island for at least a six month stay! On my return to East London I lost no time in persuading my mother to accompany me. But my father refused to hear my plea. After my perpetual begging, he finally but very reluctantly gave his consent. I wired Mr. Patterson and on the Ist Novem- ber 1936 my mother and I were ready to sail on the packet boat at 5 a.m. Then a wire arrived to say my father was joining us. The panic that followed was unbelievable but in the end Port Captain Tarpey had pity on me and organized that my father, mother and I would set sail on the packet boat for Bird Island. No one could believe the joy—AT LAST! After one and a half hours sailing the Captain of the packet boat said he was afraid the surf was too rough and we may not be able to land. After heaving to for an hour or two the wind dropped, and we finally landed on the west side of the island. A dream come true. To mom and dad proba- bly a very bleak outlook. I was enthralled the gannets were nesting and their constant calls carra! carra! carra! were almost deafening. I was enthralled by the smell of the guano and the joy of the colonies of Jackass Penguins that lined the route to our residence for our six weeks stay. Once we settled in, no time was lost in inves- tigating the life on the island, the reefs, and the fishing off the shores of the island. Many after- noons found me way out at sea fishing and col- lecting sea birds and thus it was we met Captain Goosen of the trawler Nerine. He was interested in my collections and finally came ashore to see what I had collected. Thereafter I used to go out to the trawler and collect material from their nets. The time dawned for my return to East Lon- don. I had now collected 15 huge packing cases of sea birds and marine material but how to get them back? I appealed to the trawler Captain who was only too happy to help me. Thus the contact was started and after my return to East London Capt. Goosen and a Mr. Peacock be- came firm Museum friends, and continued col- lecting and bringing in material which I could cope with in my small Museum though not too many fish were handled during this time as our space was very limited and I had no way of keeping the specimens fresh until I had time to mount them. 22 December 1938 dawned a hot, shimmering summers day. At 10:30 my newly installed phone rang to say the trawler Nerine had docked and had a number of specimens for me. I was busy completing the creating of a fossil reptile in a case, and at first thought “‘what shall I do with fish now? So near Christmas?’ Then I con- sidered I should go down and wish the men on the trawler a ‘‘Happy Christmas.’’ So I rang for a taxi and went down to the fishing wharf. It was now 11:45 and all the men had left leaving an old Scotsman who said *‘Lass they have all gone but I will show you the specimens set aside for you by Capt. Goosen.”’ I went onto the deck of the trawler Nerine and there I found a pile of small sharks, spiny dogfish, rays, starfish and rat tail fish. I said to the old gentleman ‘‘They all look much the same, perhaps I won't bother with these today’’; then, as I moved them, I saw a blue fin and pushing off the fish, the most beau- tiful fish I had ever seen was revealed. It was 5 feet long and a pale mauvy blue with irridescent silver markings. *‘What is this?’’ I asked the old gentleman. “‘Well lass,’ said he, ‘‘this fish snapped at the Captain’s fingers as he looked at it in the trawler net. It was trawled with a ton and a half of fish plus all these dogfish and oth- ers.’ ‘Oh’ I said, “‘this I will definitely take to the Museum, but I shall not worry with the rest.”’ I called Enoch, my native boy, and with a bag we had brought down he and I placed it in the bag and carried it to the taxi. Here, to my amazement the taxi man said ‘‘No stinking fish in my taxi!’ I said ‘‘Well you can go, the fish is not stinking—I will call another taxi.” With that, he allowed us to put it into the boot of the taxi. On arrival at the East London Museum I dumped it onto the table vacated by the fossil I had just assembled. And then began a library search of our meagre literature on fish. I searched the library. At the back of my mind I had written lines whilst at school on a ganoid fish—this kept ticking in my mind. Was it? No it could not be! Was it a lungfish gone balmy? No it could not be. 8 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, No. 134 FIGURE |. Then I decided the only person who would help me without a laugh would be Dr. JLB Smith of Rhodes University, Grahamstown. But to preserve it? What must I do. Now 3 p.m., the blue had faded, becoming a dirty grey. I phoned one of my Trustees, Mr. W. Sargent, who said he would lend me a hand truck. I sent Enoch to fetch it and we then took it through the streets of East London to 19 Nahoon View Road, where a friend, Mr. R. Center, practised taxidermy. He was not good at fish mounting, but I reckoned he would be able to advise and help me for prior to this I had asked my chairman Dr. Bruce Bays to place it in the mortuary at the hospital. I ap- proached the cold storage. But no one was anx- ious to have a stinking fish. It was NOT stinking; it was fresh and beautiful, but it was a blazing hot day so I had to work fast. Mr. Center was kindness itself. We got it onto a table and placed bands of cloth soaked in formalin and securely wrapped newspaper over it. I wrote Dr. Smith and sent a sketch, hoping to hear from him in a day or two. But no news. 24th still no news. Christmas day and no reply. At 5 p.m. on the 26th, I visited Mr. Center. In great fear, we opened the casing and found the formalin had not penetrated through the heavy scales. We sat and considered what should be ‘Almost armour like”’ scales of the coelacanth, described by Miss Courtenay-Latimer in her letter of 23 December 1938 to JLB Smith. Photo courtesy of Al Giddings. done. Mr. Center said the outer casing was hard. He considered he could skin it and commenced straight away. The flesh was bluish green and very oily. He cut the tongue away and said I best keep it until he got a form made. I returned to my home, very disappointed and worried, for I felt the fish was an important discovery and needed proper care. On 3 January 1939, at 10 a.m. a wire from Dr. Smith arrived: ““MOST IMPORTANT PRE- SERVE SKELETON AND GILLS = FISH DESCRIBED,”’ I trotted off to Mr. Center but alas the inners had been dumped. We went through the refuse dumps but no luck. Mr. Cen- ter had saved what he could of the fish which was the entire outer casing and the hard boney tongue. Letters and telephone messages went back and forth. A Mr. Kirsten who did a great deal of photography for the Museum had taken pho- tographs on the morning of the 22 December for me; but when I contacted him he said in taking the spool from his camera it slipped and the en- tire spool was spoilt. My Board of Trustees were not interested. It was 16 February 1939. We had been having terrific rains, and at 9 a.m. Dr. JLB Smith phoned the day before he came to the East Lon- COURTENAY-LATIMER: FIRST COELACANTH FIGURE 2. don Museum. As he walked into my small office Where I had the now mounted fish he said, “I always KNEW somewhere, or somehow a prim- itive fish of this nature would appear.” A reporter from the Daily Dispatch came in with a photographer. At first he, Dr. Smith gave them no satisfaction. Then he sat down at my desk and told them about the wonder of this find. But said he: “‘No photographs.”? ‘‘Oh”’ pleaded I, ‘‘all the photographs taken after being The author and Professor Smith viewing the unpacking of the second coelacanth at Grahamstown, South Africa, 31 December 1952. Photo courtesy of JLB Smith Institute of Ichthyology. brought to the Museum on 22 December have been lost, please could they take one shot.” Reluctantly he agreed, and how wise he had been, for that photographer sold those photo- graphs to every newspaper in the world plus the Illustrated News, and when we, the East Lon- don Museum, wanted a copy, he made us pay Gon The news of the discovery was now world wide and phone calls were received from all over 10 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 134 the world. One museum phoned to state that I was to admit that the fish had been dragged up from a muddy grave! When I refused to admit to this they said ‘‘What proof have you?”’ I stat- ed ‘‘When I took the fish off the trawler it was mauvish blue in colour but by 3 p.m. had faded to a dirty grey, which proved the pigments had died.” A public viewing for one day was arranged and people flew from all over in a single day and evening we had over 20,000 visitors to see this wonderful discovery. My Board of Trustees now became extremely interested and were anxious to get a sale for the specimen. This shocked me. So once again I appealed to my friend and sup- porter Dr. JLB Smith who now took the entire burden of the fish and its discovery onto his shoulders. The specimen was packed and sent to his special care at his private residence, where we felt that he should be in sole charge and have the honour to be able to work on this discovery. The public of East London, now aroused, plied the Board of Trustees with many whys and wherefores and stirred up great irritation. On the return of the specimen the coelacanth was placed on an all-day view, where the public streamed to view it. Thereafter it was taken to the South African Museum to be properly mounted by Mr. Drury, the taxidermist at the South African Museum, Cape Town. It was not until January 1940 when the re- mounted specimen together with a coloured cast returned to the East London Museum, where it remains today as one of the Museum’s chief ex- hibits. This story is one of the most astounding rec- ords of a woman’s intuition, for: had I never gone to Bird Island; had I never met Dr. JLB Smith who, of all the scientists I met as a young girl struggling with meagre funds in a small Museum, al- ways gave encouragement and never criti- cism; and had I not gone to the wharf to wish the men a Happy Christmas, there never would have been a coelacanth dis- covery in South Africa, on 22 December 1938. OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES The Biology and Physiology of the Living Coelacanth No. 134, 6 pages December 22, 1979 THE INFLUENCE OF THE COELACANTH ON AFRICAN ICHTHYOLOGY By Margaret M. Smith J.L. B. Smith Institute of Ichthyology, Grahamstown, South Africa When the Coelacanth was pulled up onto the deck of a trawler on December 22nd 1938, not only did it rock the zoological world but it also had a profound effect on South African ichthy- ology, mainly through the man who was pre- pared to make this staggering announcement when all his senses reeled at the impossible come true. Before the appearance of the living coel- acanth, the study of fishes in South Africa had begun with Andrew Smith during the first half of the nineteenth century. At the beginning of the twentieth century that remarkable and in- defatigable zoologist J. D. Gilchrist described numerous deepsea forms, and together with W. W. Thompson published a number of excellent checklists of South African fishes with extensive synonymy and bibliographic references. A fateful faux pas then changed the direction of ichthyological research when Gilchrist forgot a dinner engagement with the then Director of the South African Museum. The latter was so incensed that he brought Keppel H. Barnard out from England and directed him to work on the South African marine fishes! To this end he bought every book available on sea fish so that the South African Museum obtained almost every important treatise ever produced on ma- rine fish. When Barnard completed his mono- graph in 1927 he was, to quote his own words, ‘sick of fishes.’’ He then turned his fine intellect to elucidate the problems of the South African crustacea and mollusca, leaving the icuinyuiog- ical field wide open. Into this, timidly publishing his first fish paper in 1931, came a chemist and dedicated research worker, James Leonard Brierly Smith. Living in Grahamstown, he soon came to serve the four nearby museums: Port Elizabeth, Grahams- town, Kingwilliamstown and East London (as honorary curator of fishes), and by 1938 had es- tablished himself overseas as an up-and-coming ichthyologist. When, therefore, JLB announced the capture of a living coelacanthid fish, a fish believed ex- tinct for over 50 million years, it was not sur- prising that while his statement was immediately accepted overseas, at home it was met with complete disbelief! I remember so well our return to Grahams- town. The announcement with the photograph had already appeared in the East London Daily Dispatch. We went straight to the Albany Mu- seum to lay this incredible news before the man who encouraged JLB to work on fishes. We were met with a stoney face—how could JLB 12 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, No. 134 FiGuReE 1. have made such a terrible mistake. ‘‘Have you seen its picture?’” asked my husband. ‘Yes, of course’’ was the reply—‘‘that’s just a Kob (Ar- gyrosomus hololepidotus, family Sciaenidae) with a regenerated caudal.’’ JLB assured him that it was no such thing. He’d caught and han- died thousands of Kob during his lifetime and this fish was far removed from being a Kob! Nothing would convince him and we left. The next day JLB met Dr. Liebenberg, a bot- anist who worked at the Albany Museum and an old friend. He was most upset, placed his hands on my husband’s shoulders and said ‘‘What on earth made you do this dreadful thing?’’ ‘‘What dreadful thing, Lieb?’ ‘‘This coelacanth non- sense’ was the reply. ‘‘You’ll never again be able to hold up your head in any scientific com- munity.’ “‘But it is a Coelacanth!’’ he insisted “No, man, it can’t be. Old H— says it isn’t, and if he says it isn’t then it can’t possibly be one.”’ He was inconsolable, and it wasn’t JLB who convinced him it was a Coelacanth. However, in America, the ichthyologists were JLB Smith holding a venomous stonefish during field collections at Pinda. a lot less skeptical. “If JLB Smith says it is a Coelacanth, then it must be. I know his work well and he'd never give a wrong diagnosis on a matter like this.”’ One of the immediate results of the capture was a flood of correspondence. The most fas- cinating came from palaeontologists from all over the world. Not only did letters from ich- thyologists pour in, but they also sent their re- prints with queries or nomenclatorial advice. The palaeontological literature we received at that time still forms an important part of the In- stitute’s library and as such is available to workers all over South Africa. JLB decided to follow no particular school as far as the nomenclature of the bones was con- cerned. He considered it right and proper that a non-palaeontologist, and therefore someone completely outside the “‘battle of the names,” should describe the fish. So he gave the bones numbers, enabling subsequent workers to de- cide exactly what names they should bear. Many years later when the Danish coelacanth SMITH: INFLUENCE ON AFRICAN ICHTHYOLOGY PREMIO £100 RECOMPENSE Examine este peixe com cuidado, Talvez lhe dé sorte. Repare nos dois rabos que possui e nas suas estranhas barbatanas. O unico exemplar que a ciéncia encontrou tinha, de comprimento, 160 centimetros. Mas ja houve quem visse outros, Se tiver a sorte de apanhar ou encontrar algum NAO O CORTE NEM O LIMPE DE QUALQUER MODO — conduza-o imediatamente, inteiro, a um frigorifico ou peca a pessoa competente que dele se ocupe. Solicite, ao mesmo tempo, a essa pessoa, que avise imediatamente, por meio de telgrama, o profes- sor J. L. B, Smith, da Rhodes University, Grahamstown, Uniao Sul-Africana. Os dois primeiros especimes serao pagos a razao de 10.000$, cada, sendo o pagamenty garantido pela Rhodes University e pelo South African Council for Scientific and Industrial Research. Se conseguir obter mais de dois, conserve-os todos, visto terem grande valor, para fins cientificos, « as suas canseiras serao bem recompensadas. REWARD COELACANTH Look carefuliy at this fish. It may bring you good fortune. Note tk and the fins. The only one ever saved for science was 5 ft (160 cm.) long. Others you have the good fortune to catch or find one DO NOT CUT OR CLEAN IT ANY WA* : get 1t whole at once to a cold storage or to some responsible official who can care for it, and ask bk o notify Professor J. L. B. Smith of Rhodes University Grahamstown, Union of S. A., immediately by tel aph. For the first 2 specimens £100 (10.000 Esc.) each will be paid, guaranteed by Rhodes University ar hy the South Afri- ean Council for Scientific and Industrial Research. If vou get more than 2, sav: the._ cil, as every one is valuable for scientific purposes and you will be well paid. uliar double tail, > been seen. Tf Veuillez remarquer avec attention ce poisson. I] pourra vous apporter bonne chance, peut étre. Regardez les deux queuex qu'il posséde et ses étranges nageoires. Le seu] exemplaire que la science a trouvé avait, de longueur, 160 centimétres. Cependant d’autrés ont trouvés quelques exemplaires en plus. Si jamais vous avez la chance d’en trouver un NE LE DECOUPEZ PAS NI NE LE NETTOYEZ D'AUCUNE FACON, conduisez-le immediatement, tout entier, a un frigorifique ou glaciére en Gemandui a une personne competante de s’en occuper. Simultanement veuillez prier a cette personne de faire part telegraphi- quement a Mr. le Professeus J. L. B. Smith, de la Rhodes University, Grahamstown, Union Sud-Africaine, Le deux premiers exemplaires seront payés a la raison de £100 chaque dont le payment est ga- ranti par la Rhodes University et par le South African Council for Scientific and Industrial Research. Si, jamais il vous est possible d’en obtenir plus de deux, nous vous serions trés grés de les con- server vu qu’ils sont d’une trés grande valeur pour fins scientifiques, et, neanmoins les fatigues pour obtan- tion seront bien recompensées. FiGure 2. Reward poster circulated along the east African coast which resulted in the identification of the 1952 coelacanth by Capt. E. E. Hunt. palaeontologist, Jgrgen Nielsen came to South Africa in 1953 to see the second coelacanth, we suggested he should visit East London, just over 100 miles away to see the first coelacanth. ‘‘Quite unnecessary” he said, “‘the descriptions and photographs of the various structures in your monograph on the first coelacanth are more than adequate. I have no need to see them for myself.” South Africa had been rocked by the discov- ery of Latimeria. Anglers and trawlers alike were looking out for strange sea creatures. Specimens flowed into museums. Everyone was aware of Smith of Grahamstown. One of the best material rewards was his being made an Hon- orary Foreign Member of the American Society of Ichthyologists and Herpetologists, and as such received the journal Copeia. American ich- 14 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, No. 134 FIGURE 3. thyologists not only sent him rare books but helped fill in the earlier parts of Copeia. The re- sult was that for many years we had the only set of Copeia in Africa. So herpetology also bene- fitted! This, the true turning point in his career, was obscured by the second world war. For the next eight years, until the war ended, he continued to teach chemistry, do research work in chem- istry and produced two chemistry text books during that time. In 1946 the South African Council for Scien- tific and Industrial Research came into being, and JLB Smith was made one of its first three senior bursars. The coelacanth had done its work well, and the University released Smith to concentrate on his ichthyological work, to pro- duce THE SEA FISHES OF SOUTHERN AF- RICA and to hunt for the home of the living coelacanths. It is of interest to note that Lie- benberg, the botanist who was so upset about JLB’s announcement of the capture of the first coelacanth was an important link in the chain us Captain E. E. Hunt, the discoverer of the second coelacanth, aboard his vessel at Dzaoudzi, Isle Mayotte. Two weeks later this craft was destroyed by a cyclone. ~ & ey that eventually invited JLB to undertake the writing of THE SEA FISHES OF SOUTHERN AFRICA. He knew JLB had at one time started writing a book to aid anglers to identify the fish- es they caught. I wonder how much the East London coelacanth added to this. By the end of the war there was little doubt that the East London coelacanth was a stray. Most of the “‘wise’’ scientists concluded that it had come from the great depths, the Danes even mounted a ‘‘deep sea’? round the world expe- dition to look for coelacanths. They amassed a wealth of wonderful material, but no coel- acanths. Even though they passed close to the Comoro Islands, they probably fished in waters too deep for them! Smith, however, maintained that the coelacanths lived among rocks, at depths able to be reached by hook and line and in areas remote from biologists. As an angler he was even able to predict how a coelacanth would play when once hooked. He deduced that the East London specimen had drifted down on the SMITH: INFLUENCE ON AFRICAN ICHTHYOLOGY wt —_/ im a* a‘ Meare FiGure 4. JLB Smith and ‘‘Malania anjouanae,” the second coelacanth specimen, 29 December 1952. Captain Hunt is at JLB’s right. Monsieur P. Coudert, Governor of the Comores, is at his left. southward flowing Mozambique—Agulhas Cur- rent. He proposed that it must have come from somewhere south of Cape Delgado where the great south equatorial current divides to wash the coast of east Africa, where one branch swings northwards up the Kenya coast and the other southwards to South Africa. When the mammoth work THE SEA FISHES OF SOUTHERN AFRICA was completed in 1949, JLB and I set off on a series of expeditions up the east coast of Africa to seek the home of the coelacanths. We also were interested in studying the tropical fishes, for the major re- cruitment of the South African fauna to this day comes down the Mozambique—Agulhas Current. These expeditions were wonderful training times for me. In 1952 I started using diving techniques to collect in subtidal areas. One who has never dived cannot appreciate the magic of tropical reefs; the coral with all its attendant animals bathed in clear warm blue water is a marvelous sight. By the time we tracked the coelacanth home to the Comoro Islands, I knew most of the shal- low-water fishes that flitted in and out of the coral gardens. We collected, photographed, painted and preserved fish specimens as we slowly explored the virgin reefs north of Beira to Kenya. This resulted in a steady stream of publica- tions from JLB’s pen. Soon ichthyologists the world over began to realise the wealth of marine life along the east coast of Africa. Small feeble fishes from as far away as the Philippines were found in east African waters. Closely related species to those from Japan were discovered in the Mozambique channel. Our publications eventually caused an American ichthyologist to write that it would seem the great Indo-Pacific was just one little puddle after all! 16 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, No. 134 FiGure 5. The author scrubbing the second coelacanth outside of the Department of Ichthyology, Rhodes University. The eventual discovery of the Comoro Islands as the home of the coelacanths did not imme- diately stop all expeditions. By now the need to collect tropical marine species even further afield spurred us on to work at the islands north of Madagascar—from the Seychelles down to the Aldabras—and twice again along the Mo- zambique coast before JLB Smith called a halt to devote the remainder of his life to publishing monographs and smaller papers on the material. In 1968 at the age of 70, ‘‘coelacanth’’ Smith died. During his lifetime the coelacanth had riv- etted the attention of the zoological world on South Africa. It had aided the establishment of the Department of Ichthyology at Rhodes Uni- versity so that JLB Smith could devote all his time to his beloved fishes. Ichthyological and palaeontological books poured into his library now safely incorporated in the Institute’s li- brary. The search for the coelacanth brought both of us collecting among reefs never before visited by any scientists, and a knowledge of western Indian Ocean fishes that will stand me in good stead for the rest of my life. The finding of the home of the coelacanths caught the imagination of the world and added considerably to Smith’s reputation, so firmly es- tablished by the identification of the East Lon- don coelacanth. The foyer of the new building that houses the JLB Smith Institute of Ichthy- ology has been specially designed to house an extensive and well documented permanent ex- hibition of the story of the coelacanth. It starts 400 million years ago, and proceeds first to a full scale model of the East London coelacanth and finally to a special glass tank displaying the ““second’’ coelacanth. It is fitting that so many visitors to the quiet little city of Grahamstown find their way to pay homage to, and to stand gazing in awe at, the story of the fish that has done so much for South African ichthyology. OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES The Biology and Physiology of the Living Coelacanth No. 134, 8 pages December 22, 1979 INFERRED NATURAL HISTORY OF THE LIVING COELACANTH By John E. McCosker California Academy of Sciences, San Francisco, California 94118 INTRODUCTION It is patently presumptuous to describe the natural history and behavior of a living fish which, though from a lineage traceable to the Upper Devonian, has never been observed in a healthy living condition. To date more than 88 specimens of the living coelacanth, Latimeria chalumnae Smith, have been captured, but with the exception of brief observations of dying specimens made by Millot (1955), Locket and Griffith (1972), and a British Broadcasting Com- pany photographer (D. Attenborough, pers. comm.), most of the knowledge concerning the behavior of this species is based upon inferential evidence. To that body of knowledge I will herein contribute additional data gathered dur- ing the 1975 California Academy of Sciences Coelacanth Expedition and the subsequent study of recent and older museum specimens of Latimeria. Other data collected by researchers working at Grande Comore and previously un- published are included here and gratefully ac- knowledged. With the exception of the first known speci- men of Latimeria, all subsequent captures have been made off the islands of Grande Comore and Anjouan in the Comoran archipelago. The Com- oro islands have undergone dramatic political upheaval since our 1975 expedition. The over- throws of the French colonial government, the Abdallah government, the Soilih government, and the Denard government have closed the country to western expeditionary groups. Dur- ing the Soilih government all official records were destroyed (Anon. 1978) including, as I have ascertained, all recent coelacanth catch data. I am therefore including catch record data which were provided me by the Department of Production in Moroni in 1975, through inter- views with Comoran fishermen and their chief Mohammed Ali Chabane, and through subse- quent correspondence with Said Bacar Moussa of Mitsamouli, Grande Comore. These data (Appendix 1), supplement the coelacanth catch records tabulated by Millot et al. (1972). OCEANOGRAPHIC SETTING OF COELACANTH CAPTURES The restriction of coelacanth captures since 1938 is probably a result of oceanographic and sociological conditions unique to Grande Co- more and Anjouan. Grande Comore and An- jouan are high volcanic islands which lack a fish- able fringing reef and inner lagoon; Grande Comore and Anjouan are rare environments in that the submarine topography of most other 18 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES. No. 134 Itsandra . Maogueni _N'Gouni Volcano M' Bachili *_i:Moindzaza Volcano } Mitzoudje FiGuRE 1. acanth capture cites identified. Soundings are from Forster (1974). Bathymetric map of Grande Comore with coel- nearby volcanic coastlines has been smoothed by sedimentation and/or coral growth along their slopes (Forster 1974). Whereas the shallow la- goons of other Indian Ocean islands are pre- ferred fishing grounds, their absence at Grande Comore and Anjouan has necessitated that fish- ing be done from the deep reef edge at depths of 20-200 m. Another contributory factor is the Comoran fishery for the deep-reef associated oilfish, Ruvettus pretiosus, a fish of special sig- nificance to Islamic culture because of its pur- ported antimalarial properties when used as a salve. The submarine topography of Grande Co- more is particularly sheer along the southwest coast except for that along the village of Iconi where a small fan spreads seaward from the N’Gouni volcanic rim (Fig. 1). This has been the primary site for coelacanth capture (16 of 86 specimens), probably a result of the proximity to Iconi, the major fishing village of Grande Co- more, as well as the favorable oceanographic conditions which exist there. A popular misconception exists concerning the depth of capture of Latimeria. Coelacanths are not bathypelagic as is often suggested, and have been captured between depths of 70 and 20, 15} Latimeria caught 10: 5 O 100 200 300 400 500 600 depth (m) FiGURE 2. Reported depth of capture of 65 coelacanths. 600 m. It is quite likely that Latimeria prefers a habitat of 200 m or more but occasionally makes feeding forays into shallower water, par- ticularly during and about the dark of the new moon phase. It is my impression that the depth of capture recorded by Millot et al. (1972) for many coelacanths is in error and exaggerated toward greater depths. Supporting evidence is based on interviews made by the author and Sandra McCosker (McCosker and McCosker 1976) of Comoran fishermen and of M. Francis Debuissy, then Chief Veterinarian of the De- partment of Production of the Comores. Native fishermen employ a hand-line fishing method which, due to currents, suspends the line obliquely in the water column. The actual length of line is by necessity considerably greater than the actual depth of capture. This was verified by 1975 Expedition members by accompanying Iconi coelacanth fishermen onto the fishing grounds and comparing their opinion of depth with the actual depth as measured by scuba div- ing. Francis Debuissy informed us that Comoran fishermen often erroneously report their dis- tance from shore as the depth of capture, such that a fisherman fishing one-half kilometer from shore will report that as a depth of 500 m. Presented in Figure 2 are combined data from Millot et al. (1972: fig. 4) and 15 subsequent cap- tures. The majority of fishes are purportedly captured between 150-300 m, which probably reflects the apparent depth of the oilfish-catching effort. The 100-300 m bathymetric horizon forms a narrow band | km or less in width around Grande Comore. TEMPERATURE AT DEPTH OF CAPTURE It has been suggested that Latimeria inhabits submarine caves wherein freshwater aquifers McCOSKER: COELACANTH NATURAL HISTORY FIGURE 3. seep (J. Anthony, pers. comm.; Forster 1974). Submarine recesses were observed by expedi- tion members who, while scuba diving, discov- ered cool, freshwater aquifers between depths of 30 and 70 m. Forster (1974), citing the geol- ogist B. G. J. Upton, described the geologic con- dition responsible for this phenomenon: Although the basaltic rock is mainly hard and impervious there are porous clinkery horizons and longitudinal holes within the lava flows, both of these systems would make good aquifers. It is therefore likely that the considerable rainfall on the high ground of M’Kartala, which has no per- manent surface streams, could provide a supply with sufficient pressure to flow out even at 200 m depth. It is important to note that the coelacanth catch is increased during January, February, and March (see Fig. 5) in spite of the reduced fishing effort due to the monsoon rains. A fairly good Coelacanth fisherman with handline. temporal correlation exists between the rainfall at Maoueni, Grande Comore, and the incidence of coelacanth capture. Maoueni, is located on a broad lava fan which flows onto the SE coast. This correlation is in agreement with the drought of 1975/1976 and the unfortunate lack of coel- acanth captures during January—March 1975, a period of normal monsoon rainfall when the fish- ing effort was increased several fold on behalf of our expedition (McCosker and McCosker 1976). Fewer coelacanths than normal were cap- tured throughout Comores during the drought condition which existed across the SE African subcontinent in 1975-1976. The temperature at the depth of coelacanth capture appears to be slightly below surface temperatures. The majority of coelacanth cap- tures have been near or above the deep ther- mocline which flanks the steep island slope. Anjouan data from Domoni and Mutsamudu (Menaché 1954) identified thermoclines between 150-200 m, with temperatures between 23—26°C 20 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, No. 134 FIGURE 4. in the upper 100 m and 15°C below 200 m. Un- published bathythermograph data collected by the Vancouver Aquarium at Iconi, Grande Co- more, convey a similar deep and stable ther- mocline at 115-125 m, and temperatures of 15— 17°C below 200 m. It is important to note that the temperature at depth, measured by a suspended bathythermo- graph, is not necessarily the temperature regime inhabited by Latimeria. The freshwater aquifers are generally cooler than the surrounding salt- water and might, for that reason, be selected for by Latimeria to permit them to extend their range of thermal tolerance into the relatively richer shallow reef area for feeding. COELACANTH PREY ITEMS The dentition, jaw structure, and reduced gut length of Latimeria are adaptations for a pred- atory feeding behavior. Published records of gut contents are limited to entire cuttlefish (Mc- Allister 1971) and the lanternfish Diaphus me- topoclampus (Millot and Anthony 1958). The examination of recent specimens has resulted in Coelacanth fishermen fishing along the dropoff for oilfish and coelacanths near Itsandra, Grande Comore. additional observations of prey species, allowing further insight into coelacanth behavior. The large female Latimeria at the United States National Museum of Natural History (C49, 165 cm, USNM 205871) contained the gill arches, hyoid, dentary, cleithrum and several vertebrae of a Stout Beardfish, Polymixia noblis (family Polymixiidae). A comparison of the re- mains with the jaws of intact P. noblis (USNM 202120) indicated that the prey specimen was ca. 17-18 cm in standard length. Wheeler (1975) states that P. noblis is a near-bottom dweller, widely distributed in all tropical oceans at depths of 183-640 m. The male Latimeria at the California Acad- emy of Sciencesa(@79; 108 ‘em, 1CASP 33111) contained two intact, but partially digested, specimens (7-8 cm) of deepwater snappers, Symphysanodon sp. (family Lutjanidae). I am advised by G. D. Johnson that they are the sec- ond known Indian Ocean specimens and prob- ably represent a new species. Elsewhere, species of Symphysanodon are from moderate depths (119-476 m) around islands (Anderson 1970) and live on or just above the bottom. McCOSKER: COELACANTH NATURAL HISTORY heenad Latimeria caught S ON OD J F month FIGURE 5. MAM JS J A Correlation of mean monthly rainfall at Maoueni, Grande Comore, during 1961-1970 with the monthly capture of coelacanths (82 specimens) in the archipelago. Figure by K. O'Farrell. No identifiable remains were discovered with- in the stomach of the small, male Latimeria at the Scripps Institution of Oceanography (C78, 103 cm, SIO 75-347). Small, apparent fish frag- ments, including otoliths and scales, were found in the spiral valve by C. L. Hubbs. The adult, male Latimeria at the California Academy of Sciences (C59, 122 cm, CAS 24862) contained two specimens of deepwater cardi- nalfishes (family Apogonidae), Coranthus poly- acanthus. The specimens, 12—14 cm in standard length, were partially digested, but conclusively identified by G. D. Johnson on the basis of their peculiar osteology. Coranthus is monotypic and was previously known only from two deepwater (150 m) specimens from Réunion Island (Fraser 1972). The above-listed prey species are typical in- habitants of insular deep reefs, associated with but living slightly above the substrate. All prey species appeared to have been swallowed entire, presumably the result of a grouper-like rapid-in- halation feeding method. It is important to note that open water (non-reef associated) bathype- lagic or mesopelagic prey items have not been observed among coelacanth gut contents. CONCLUSIONS An hypothetical life history of Latimeria cha- lumnae can be constructed on the basis of its anatomy, diet, catch records, and Comoran oceanographic and meteorological data. On that basis, it appears that Latimeria behaves like a large, reef-associated piscivorous grouper. The ovoviviparous reproductive system is a specialization related to the retention of urea, not unlike that of the ovovivaparous elasmo- branchs. The large 163 cm female of 5 January 1972 contained 20 eggs, 8.5—9 cm in diameter and 300-344 g in weight (Millot and Anthony 1974). Although it is possible that some may have aborted during capture, it appears that they represent the normal complement of a pregnant female. The only known embryos were near 22 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, No. 134 term and indicate that length at birth is more than 32 cm (Smith et al. 1975). The estimated gestation period is approximately 13 months (not unlike many elasmobranchs) with births pre- sumably occurring in February (Smith et al. 1975; Hureau and Ozouf 1977). The juvenile Latimeria are probably predatory as indicated by their dentition and jaw structure. The allo- metric growth of the upper jaw and head length (McAllister and Smith 1978) would assist the predatory behavior of young fish. The smallest Latimeria captured on hook and line is 42.5 cm long and weighs 800 g. Its age, first estimated to be 3% years by Anthony and Robineau (1976), has been reinterpreted to be one year or less (Hureau and Ozouf 1977). Age and growth es- timates made from scale analysis suggest that the largest females (180 cm) are nearly 11 years old. It appears that females attain a larger length and weight than males (McAllister and Smith 1976; and my data), a condition shared by cer- tain sharks and probably related to the repro- ductive commitment to ovoviviparity and preg- nancy. The limited geographic distribution of Lati- meria must in some way be related to the re- duced vagility which accompanies its live-bear- ing reproductive mode. The lower depth limit of Latimeria distribution has not been delimited, but it is quite likely that the seamount chain be- tween the Comores and the African coast is traversable by juveniles and/or adults. The chance discovery of a Latimeria off South Af- rica in 1938 has not been repeated, in spite of an actively continuing and reasonably informed fishery. It is likely that coelacanths exist uncom- monly along the offshore seamounts and banks of the western Indian Ocean, but this has not been confirmed due to the capture difficulties and the lack of a prolonged exploratory fishing effort. Their presence might most effectively be explored using shallow depth submersibles. The seasonality and lunar periodicity of coel- acanth capture indicates that their behavior and/ or presence in shallow water fluctuates. Presum- ing that Latimeria presence in shallow water is affected by rainfall-fed submarine aquifers, it seems likely that the fat-investment of the swim- bladder is an adaptation which assists vertical migration. Dead, intact specimens are slightly denser than seawater. The high, extracellular lipid and wax ester content of the muscles com- pensates somewhat for the lack of swimbladder function (Nevenzel et al. 1966). The fish is there- fore slightly negatively ‘“‘buoyant’’ (not unlike large groupers, cirrhitids, and blennioids), al- lowing it to perch on a reef platform and lunge short distances to capture prey items. Its body shape and fin size and location are adapted for such a feeding method. The intercranial mobil- ity, subcephalic musculature, and jaw angle (Thomson 1966, 1970, 1973; Alexander 1973) also contribute to rapid prey capture and en- gulfment. The coelacanth eye is adapted to moderate depths in clear, tropical waters. The retina pos- sesses numerous, densely-packed rods; cones (single type) are very rare and possess a single oil droplet (Ali and Anctil 1976). The visual pig- ment maximum absorbance is at 473 nm (Dart- nall 1972). The large, nearly color-blind eye is therefore adapted to low light levels (indicative of a primarily nocturnal activity pattern?) and similar in habit and structure to elasmobranchs which occupy a similar habitat (Millot and Ca- rasso 1955). The relatively large eye of the em- bryos and juvenile specimen evidence an allo- metric growth (McAllister and Smith 1978) which presumably would allow young fish to occupy the same photic horizon as adults. If the natural history of the living coelacanth is as I have inferred, then it is quite likely that Latimeria is amenable to aquarium captivity. Its lack of a functional swimbladder would allow its existence at ambient pressure and a minimal re- duction in temperature and light level might be the only modifications necessary. I look forward to the solution of the political problems within the Comores so that an expedition may be mounted and a coelacanth might finally be cap- tured in a living, healthy condition. At that time, Latimeria will be given an opportunity to agree or disagree with the life history portrait that its students have hypothesized. ACKNOWLEDGMENTS Many individuals have assisted in this study. I am particularly grateful to: Jean Anthony and Daniel Robineau, Paris Museum of Natural His- tory, for allowing me to examine coelacanth specimens; Francis Debuissy, then of the Com- oran Department of Production, for assisting me during the 1975 Expedition and providing infor- mation about coelacanth fishermen; the Vancou- McCOSKER: COELACANTH NATURAL HISTORY ver Aquarium, for permission to cite their un- published temperature data; G. David Johnson, then of the Scripps Institution of Oceanography, for identifying coelacanth prey items; William N. Eschmeyer, California Academy of Sciences, and the staff of the U.S. National Museum of Natural History, for assistance with specimens in their fish collections; and many citizens of Grande Comore for their hospitality, informa- tion, and assistance during and since the 1975 Expedition. Particular thanks are due President Ahmed Abdallah, Prince Nacr-Ed-Dine, Chief Mohammed Ali Chabane, Abdallah Said Omar, and Said Bacar Moussa. LITERATURE CITED ALEXANDER, R. M. 1973. Jaw mechanisms of the coelacanth Latimeria. Copeia 1973(1):156—-158. Ati, M. A., AND M. ANcTIL. 1976. Retinas of fishes. Spring- er-Verlag, Berlin. 284 pp. ANDERSON, W. D. 1970. Revision of the genus Symphysan- odon (Pisces: Lutjanidae) with descriptions of four new species. Fish. Bull. 68:325—346. ANONYMOUS. 1978. A man and his dog. Time Magazine, 21 August, p. 37. ANTHONY, J., AND D. ROBINEAU. acteres juveniles de Latimeria chalumnae. C. R. Acad. Sci. 283: 1739-1742. DaRTNALL, H. J. A. 1972. Visual pigment of the coelacanth. Nature 239:341-342. Forster, G. R. 1974. The ecology of Latimeria chalumnae Smith: results of field studies from Grande Comore. Proc. R. Soc. London, Ser. B 186:291—296. Fraser, T. H. 1972. Comparative osteology of the shallow water cardinal fishes (Perciformes: Apogonidae) with ref- erence to the systematics and evolution of the family. Ich- thyol. Bull., J. L. B. Smith Inst. Ichthyol., No. 34. 105 pp. HurReEAU, J.C., AND C. OzouF. 1977. Determination de l’age 1976. Sur quelques car- et croissance du coelacanthe Latimeria chalumnae. Cy- bium, 3° sér. 2:129-137. Locket, N. A., AND R. W. GRIFFITH. of a living coelacanth. Nature 237:175. MCALLIsTeER, D. E. 1971. Old Fourlegs, a ‘‘living fossil.’’ National Museums of Canada Odyssey Series. 25 pp. , AND C. L. SmitH. 1978. Mensurations morpholo- giques, dénobrements meristiques et taxonomie du coel- acanthe, Latimeria chalumnae. Nat. Can. 105:63—76. McCosker, S., AND J. E. McCosker. 1976. To the islands of the moon. Pac. Discovery 29:19-32. MENACHE, M. 1955. Etude hydrologique sommaire de la ré- gion d’Anjouan en rapport avec la peéche de trois coel- acanthes. Mem. Inst. Sci. Madagascar, Sér. A 9:151-185. MILLor, J. 1955. First observations on a living coelacanth. Nature 175:362-363. , AND J. ANTHONY. 1958. Latimeria chalumnae, der- nier des crossoptérygiens. Pages 2553-2597 in P. P. Grassé, ed., Traite de Zoologie, Vol. 13. Masson, Paris. , AND 1974. Les oeufs du coelacanthe. Sci. Nat. Paris, No. 121, pp. 3-4. 5 , AND D. RoBINEAU. 1972. Etat commenté des captures de Latimeria chalumnae ... effectuées jusqu’au mois d’octobre 1971. Bull. Mus. Nat. Hist. Nat. 39:533- 548. MILLoT, J., AND N. Carasso. 1955. Note preliminaire sur Voeil de Latimeria (Coelacanthidae). C. R. Acad. Sci. 241:576-577. NEVENZEL, J. C., W. RoDEGKER, J. F. MEAD, AND M. S. GORDON. 1966. Lipids of the living coelacanth, Latimeria chalumnae. Science 152:1753—1755. SmitH, C. L., C. S. RAND, B. SCHAEFFER, AND J. W. Atz. 1975. Latimeria, the living coelacanth, is Ovoviviparous. Science 190:1105—1106. THOMSON, K. S. 1966. Intercranial mobility in the coel- acanth. Science 153:999-1000. 1970. Intercranial movement in the coelacanth Lat- imeria chalumnae Smith (Osteichthyes, Crossopterygii). Postilla 149:1-12. . 1973. New observations on the coelacanth fish, Lar- imeria chalumnae. Copeia 1973(4):813-814. WHEELER, A. 1975. Fishes of the world. MacMillan Pub. Co., New York. 366 pp. 1972. Observations APPENDIX 1: COELACANTH CAPTURE LIST This list is a continuation of the list compiled by Millot et al. (1972). These data were provided by the Grande Comore Direction de la Production and several informants from Iconi, Moroni, and Mitsamiouli.. As mentioned in the text, the pur- ported depth of capture and distance from shore are fisher- mans’ estimates; they should not be considered to be accu- OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, No. 134 rate. There is a discrepancy of three specimens between this list and that in Millot et al., such that C67 here would be C70 by their system. This is due to two specimens which they note in a footnote (p. 547) and a taxidermied adult (ca. 100 cm length) captured near Mitsamiouli and attached to the wall of the bar in the Maloudja Hotel. Abbreviations are: A, Anjouan; GC, Grande Comore. Distance Speci- from shore Depth Weight Length men Date Time Location (m) (m) (kg) (cm) Sex C70 5172 0100 A: Domoni 2,000 400 78 163 2 C71 22 Ill 72 0200 GC: Iconi 1,000 300 10 90 fe) C72 12 V 72 2300 GC: Iconi 800 90 40 140 9? C73 12 VIII 72 0300 CG: Iconi 400 100 — 90 ? C74 16 X 72 — A: Dzindni 1,000 350 30 120 3? C75 6 VII 73 — GC: Mitzoudje 500 — 35 132 3 C76 27 VII 73 — GC: Iconi 100 100 10 86 fo) C77 6 XI 73 0200 A: Vouani — 175 32 120 3? C78 22 XI 73 2100 GC: Iconi 800 180 24 103 3 C79 27 XI 73 0330 GC: M’Bachili 400 225 30 110 3 C80 14 II 74 0100 A: Miroutsy — 220 40 139 3 C81 17 V 74 2100 GC: Vanamboini 300 150 40 139 ? C82 17 VII 74 1645 GC: Iconi — 180 0.8 42.5 & C83 18 X 74 —_ GC: Iconi — 240 45 180 g C84 9 XI 74 = GC: Iconi 2,000 250 37 145 2? C85 22175 — A: Mromouhouli 3,000 300 —_ 165 Q C86 27176 —_— GC: Iconi — — = = = C87 IX 76 — A: Domoni — — 11 90 — C88 III 77 — GC: Iconi — = = os = OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES The Biology and Physiology of the Living Coelacanth No. 134, 20 pages December 22, 1979 THE COELACANTH AND THE CHONDRICHTHYES AS SISTER GROUPS: A REVIEW OF SHARED APOMORPH CHARACTERS AND A CLADISTIC ANALYSIS AND REINTERPRETATION By Michael D. Lagios Children’s Hospital, San Francisco, California 94118, and Steinhart Aquarium, California Academy of Sciences, San Francisco, California 94118 CHAPTER I The principal expectation generated by Ms. Latimer’s finding of that singular blue fish in a discarded trawler haul was that the soft parts, biochemistry and physiology of the living coel- acanth might mirror that of its long extinct but morphologically similar brethren. The signifi- cance of this lay not in studying these features in a living fossil alone but rather in what these studies might imply for its extinct sister group, the Rhipidistia, and the immediate preterrestrial condition. Coelacanths and Rhipidistia are convention- ally classified as sister taxa and share numerous seemingly homologous features including lobed fins, scales, lungs and most importantly a pe- culiar intracranial joint (Romer 1966, 1968; Thomson 1969). This conventional classifica- tion of the coelacanth as a Crossopterygian re- flects a selection bias of calcified tissues and ex- ternal form, since prior to 1938 that was all the evidence available for review, and because sys- tematists have been more accustomed to count- ing vertebrae than determining vascularization, biochemistry, and protein structure. It was with this bias that the coelacanth has been studied by numerous independent investigators hoping to look backward into the Devonian through a beast little changed since they and our preter- restrial ancestors presumably diverged. Yet these expectations reflect the romance of the coelacanth more than the reality of the dead oily fish putrifying in the hot summer sun of East London where Marjorie Courtney Latimer dis- covered it in 1938. This paper will present a reappraisal of the systematic position of the coelacanth which sug- gests that this fish is in no way related to the Rhipidistia or their tetrapod descendents and in fact probably represents an archaic sister group to the Chondrichthyes—or briefly, a palaeozoic shark-like fish. This reappraisal is predominantly based on three complex characters: 1) The pituitary gland: pars distalis structure and organization; 2) Cation excreting tissue: the presence of a ho- mologous rectal gland; and 3) Osmotic adaption: specifically urea and TMAO retention in the marine environment. 26 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, No. 134 In a cladistic analysis traits may appear to be similar because: 1) they are common primitive or plesiomorphic features of the group, 2) be- cause of independent convergent evolution, and 3) finally, because they are homologous special- ized characters or synapomorphic. Only the lat- ter are useful in evaluating systematic relations. My thesis is that the coelacanth and Chondrich- thyes share a number of independent synapo- morphic characters which make an association with the Rhipidistia and their tetrapod descen- dents untenable. In order to support this con- clusion, this presentation will establish that the similarities in the several characters I present are: 1) truly specialized, that is apomorphic and not primitive or baseline features of an ancestral group; 2) exclusive to the proposed sister groups, that is the Chondrichthyes and the Coelacanthi- ni; and 3) most likely homologous rather than examples of convergent evolution. Pituitary Complex The pituitary complex, derived from an evag- ination of the roof of the stomodeum and a cor- ollary growth of the overlying hypothalamic floor, is a constant and diagnostic feature of the Vertebrata. Both its morphology and the peptide sequences of its various hormones indicate a very conservative rate of evolution. Amino acid sequence analysis of the neurohypophysial hor- mones, arginine vasotocin and subsequent sister derivatives (Acher et al. 1968; Sawyer 1969, 1977) and more recent analysis of the pars dis- talis glycoproteins, TSH, LH (Fontaine and Burzawa-Gerard 1977) and to some extent growth hormone (Hayashida et al. 1975; Hay- asnhida 1977; Wallis 1975), have been utilized by comparative biochemists and systematists to es- tablish protein phylogenies and to aid systematic group analysis. Analysis of the morphology of the pituitary, its innervation, vascularization and distribution of cell types, has generated an enormous literature in comparative endocrinol- ogy over the last fifty years and has provided many signal discoveries in biology. Despite this intense interest in pituitary struc- ture, however, only a few basic patterns of or- ganization, i.e., the way in which the neurohy- pophysis and adenohypophysis relate to each other in terms of innervation, vascularization and distribution of specific cell types, have been documented. These patterns are summarized in the accompanying table and schematic diagram (Fig. 1). Some of the variations are rather minor in character, and all save the tetrapod pattern, are limited to fishes. Each pattern is so distinc- tive that it is diagnostic for the group which pos- sesses it. The extant petromyzonts and myxinoids dem- onstrate pituitary organization in rather simple form. The adenohypophysis is a relatively flat, compact discoid mass of cells which shows close approximation to the neurohypophysis poste- riorly to form a neurointermediate lobe, a char- acteristic feature of all aquatic vertebrate groups. There is no direct innervation of the an- terior adenohypophysis or equivalent pars dis- talis. In the petromyzonts numbers of vessels occur in the thin fibrous tissue separating the pars distalis from the overlying hypothalamic floor and although no hypothalamo-hypophysial portal system occurs, perivascular connective tissue may serve as a diffusion gradient for neu- rotransmittor substances (Gorbman 1965; Tsu- neki and Gorbman 1975; Goossens et al. 1977). In the myxinoids the vascular arrangements are less well developed but similar (Henderson 1972; Kobayashi and Uemura 1972). All gnathostomes, fishes and tetrapods have a well developed hypothalamo-hypophysial por- tal vascular system through which specific neu- rotransmittors, identified only recently in mam- mals, control the secretory function of specific adenohypophysial cell types. The majority, in- cluding relic actinopterygians, dipnoans and tet- rapods, retain a compact pars distalis with only modest evidence of regional differences in dis- tribution of specific cell types within the gland, most clearly evident in birds (Wingstrand 1966; Hayashida and Lagios 1969; Lagios 1968a, 1968b, 1970; Hansen 1971; Polenov 1968). In contrast to both the agnathans and the fore- going group of gnathostomes, which have re- tained a number of plesiomorph features in pi- tuitary organization, are the teleosts and Chondrichthyes. Teleosts are unique among vertebrates for the number of radical innovations in pituitary or- ganization which they manifest. Among these are direct synaptic innervation of specific pars distalis cell types with a corollary suppression of the hypothalamo-hypophysial portal system, an extreme segregation of cell types such that the gland becomes compartmentalized, and an interdigitation of the neurohypophysis with the pars distalis (Vollrath 1967; Lagios 1970; Zam- LAGIOS: COELACANTHS AND CHONDRICHTHYES AS SISTER GROUPS 27 ACTINOPTERYGII TELEOST TETRAPOD FIGURE 1. PETROMYZON + PRO Patterns of vertebrate pituitary organization. Schematic diagrams of representative vertebrate types in midsagittal section. Anterior, left. Neurointermediate lobe (pars intermedia and associated posterior peptidergic neurohypophysis) shaded. Vascular systems and separate blood supplies of neurointermediate lobe (Dipnoi) and pars nervosa (Tetrapod) indicated as schematic vessels. Specific hormone-producing cell types of pars distalis indicated by key. Chondrichthyes: Note the separate ventral lobe with gonadotrope and thyrotrope sequestration and its separate vascular supply. brano 1971). These features occur among phy- letically more primitive teleost groups, e.g., the clupeomorphs, elopomorphs and ostariophy- sines, but are best developed in the more ad- vanced recent groups, e.g., the perciforms. Te- leost pituitary organization is decidedly apomorphic, probably more so than any other vertebrate group. This would appear to be of very recent origin. Although more remote in origin, the Chon- drichthyes similarly exhibit unique features in pituitary organization as compared to a more conservative baseline pituitary structure. These unique features are exemplified by the elasmo- branchs in which the pars distalis is peculiarly divided into a number of separate cell masses or lobes, each of which sequesters one or more functional cell types and shows specific special- izations in vascularization. Although both the rostral and proximal partes distalis, the more conventional portions of the elasmobranch pi- tuitary, show vascular specializations—with each area receiving portal venous drainage from a specific region of the hypothalamic floor (Mel- linger 1960; Meurling 1967a)—it is in the devel- opment of a ventral lobe that this tendency oc- curs in its most extreme form. The ventral lobe is a physically separated ventral portion of the pars distalis which may be connected by a persistent tubular remnant of the hypophysial cavity (= Rathke’s pouch equiva- lent) with the more dorsally located median lobe of the pars distalis (Norris 1941; Alluchon-Ger- ard 1971; Knowles and Vollrath 1975). The lobe itself is closely associated with and often adher- ent to the carotid anastomosis, a characteristic feature of elasmobranchs (Norris 1941), and fre- quently lies in arecess within the chondrocranial floor (Figs. 2 and 3). The ventral lobe has a char- acteristic branching tubular structure which contrasts with the more compact arrangement of follicles and cords of the remainder of the 28 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, No. 134 Ficures 2 and 3. Fig. 2. Notorynchus maculatus (Seven-gill shark), midsagittal diagram. Anterior, right. Pituitary: dark shading. Note the ventral lobe separated from the remainder of the adenohypophysis (A) in a recess within the floor of the chondrocranium (light shading). Portion of hypophysial duct connecting ventral lobe with remainder of pars distalis is visible. Vessels in the recess represent carotid anastomosis. OX, optic chiasm. Scale 5mm. Fig. 3. Heterodontus francisci (Hornshark), midsagittal diagram. Anterior, right. Ventral lobe lies in a shallow recess within chondrocranium. Note associated carotid anastomosis and hypophysial veins. Scale 5 mm. LAGIOS: COELACANTHS AND CHONDRICHTHYES AS SISTER GROUPS 29 pituitary gland. Uniquely among vertebrates, the ventral lobe receives a direct arterial supply from the internal carotids and is independent of the venous effluent of the hypothalamo-hypo- physial portal system (Mellinger 1960; Meurling 1967a). This physically separated, independent- ly vascularized portion of the pars distalis se- questers in large part, if not exclusively, the gonadotropic activities of the pituitary in elas- mobranchs. Evidence for this has been estab- lished from extirpation experiments (Dodd 1975; Dobson and Dodd 1977) bioassays using chick testis (Scanes et al. 1972) and Xenopus oocyte systems (Firth and Vollrath 1973), and activa- tion of specific ventral lobe adenyl cyclase using mammalian LH-FSH-RF (Deery and Jones 1975). These studies corroborate immunohisto- chemical localization (Mellinger and Dubois 1973) based on a specific reaction with anti- ovine LH limited to the elasmobranch ventral lobe. Firth and Vollrath (1973) did identify some LH-like activity in the median lobe of Scylio- rhinus, but the mean value was only one tenth that present in ventral lobe extracts. Such hormonal sequestration is characteristic of the other lobes of the elasmobranch pars dis- talis. Evidence indicates that TSH is restricted to the ventral lobe along with LH-like activity, while ACTH activity is localized in the rostral lobe (Jackson and Sage 1973; Mellinger and Du- bois 1973). For a long while considerable speculation was generated concerning the functional significance of a vascular arrangement which would appear to exclude gonadotropic function from central nervous control in the elasmobranchs. Seasonal studies however, have clearly shown that go- nadotropic activity must be correlated with CNS control in this group. More recently, elegant ex- periments have shown specific activation of ven- tral lobe adenyl cyclase with elasmobranch hy- pothalamic homogenates (Deery and Jones 1975). This evidence suggests that elasmo- branchs regulate gonadotropic function by a specific neurotransmittor as in other verte- brates, and that the peculiar morphology of the system is a ‘nonsense’ feature in terms of en- docrine control and function, as witness the re- productive success of elasmobranchs over the last 300 million years. Since the survival value of a “‘nonsense’’ feature is null, and by defini- tion it is outside the influence of selection pres- sure, its potential value in a cladistic analysis is increased. Among the Holocephali, the sister group of the extant elasmobranchs, pituitary organization is similarly unique although more bizarre (Fig. 4). The equivalent ventral portion of the pars distalis remains in close association with the homologue of the internal carotids, but both the lobe and the homologous vessels remain outside the chondrocranium in the roof of the oral cavity (Fahrenholz 1828; Sathyanesan 1965). The in- ternal carotid homologues form an anastomosis in some forms (Allis 1912; Meurling 1967b; Jasinski and Gorbman 1966) and in others show ipsilateral development, but do supply the ex- tracranial portion of the pars distalis, or Rach- endach-hypophysis. Embryologically the Rach- endach-hypophysis shows a tenuous tubular connection through a defect in the chondrocran- ium with the persistent hypophysial cavity of the more conventional intracranial pars distalis (Honma 1967), a condition entirely comparable to that of the sharks Notorynchus and Hetero- dontus. Like the elasmobranch ventral lobe the holocephalan Rachendach-hypophysis appears to sequester gonadotropic function, based on seasonal studies of gonad maturation, and shows a histochemical reaction consistent with a di- sulfide containing glycoprotein secretion com- parable to that of the known gonadotropes of the elasmobranch ventral lobe (Sathyanesan 1965). Recently Dodd and Dodd (1979) have confirmed high levels of gonadotropic activity in the Rachendach-hypophysis of Hydrolagus ina radio-immunoassay using an antibody to a par- tially purified elasmobranch gonadotropin. Apart from prior brief general descriptions (Millot and Anthony 1955, 1958, 1965) studies of the coelacanth pituitary are limited to a few in- dividual specimens, only two of which received proper histologic fixation (Lagios 1975; van Ka- menade and Kremers 1975). These latter two studies represent two specimens, a small im- mature female, 85 mm total length, and a large ovigerous mature female, 167 cm. Despite the fact that van Kamenade and Kremers (1975) were limited by their material to the more prox- imal portion of the pituitary, their independent observations and those of the present author (Lagios 1972, 1975) are remarkably similar. The most striking feature of the coelacanth pituitary complex is the great anterior elongation 30 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, No. 134 Ficures 4 and 5. Fig. 4. Hydrolagus colliei (Ratfish), midsagittal diagram. Anterior to right. Separated portion of the pars distalis, the Rachendach-hypophyse (R) lies outside the chondrocranium surrounded by lymphoid tissue, L, beneath the oral mucosa, M. More conventional intracranial hypophysis, A, is depicted in relation to the optic chiasm, OX, and saccus vas- culosus, SV. Scale 10 mm. Fig. 5. Latimeria chalumnae (Coelacanth). Sagittal diagram. Anterior, left. Rostral extension (pars buccalis) of the pars distalis, PD, greatly attenuated and 83 mm long in this 140 cm adult male. Internal carotids, IC, converge to anastomose in a deep recess in the basisphenoid, BS. Lateral vein, V. Notochord, N. PI, pars intermedia (neurointermediate lobe); PA, preafferent arteriole of portal system; OX, optic chiasm. Scale 10 mm. LAGIOS: COELACANTHS AND CHONDRICHTHYES AS SISTER GROUPS 31 of the pars distalis (Fig. 5) (Millot and Anthony 1955). This elongate extension, variously termed pars buccalis, or rostral division, consists of an attenuated hypophysial cavity from which branch numerous tubular structures whose ar- chitecture is reminiscent of the holocephalan Rachendach-hypophyse. The intermittent islets of endocrine tissue noted previously (Millot and Anthony 1955, 1958; Lagios 1972, 1975; van Kamenade and Kremers 1975) represent vari- able discontinuous clusters of branching tubules arising from a continuous duct-like hypophysial cavity (Fig. 6). Aside from the tubular hypophy- sial cavity, this lobe, like the ventral lobe of elasmobranchs, is physically separated from the remainder of the pars distalis. At its rostral ex- tremity it is intimately associated with an anas- tomosis or convergence of the internal carotids (Lagios 1972). Pars distalis follicles lie within the adventitia of the carotids prior to their anasto- mosis in the chondrocranial floor. Clearly evi- dent in serial reconstruction of this region are direct arterial vessels which arise obliquely from the carotids and run retrograde before bifurcat- ing within the lobe. Histochemical studies (La- gios 1975; van Kamenade and Kremers 1975) have shown the presence of disulfide-containing glycoprotein secretion granules within the cells of this lobe. Only thyrotropes and gonadotropes among definitively identified functional pituitary cell types show this histochemical reaction. Cor- roboration from endocrine ablation experi- ments, bioassay, radio-immunoassay and im- munofluorescent reactions to specific hormones have not yet been performed in the coelacanth. The cytology of the small disulfide containing cells in the immature female (Lagios 1975) con- trasts markedly with large vacuolated cells of this same lobe in the sexually mature ovigerous female reported by van Kamenade and Kremers (1975) and noted by M. Olivereau (pers. comm.) who studied the bulk of the pars buccalis of the same 5 January 1972 specimen. Both investi- gators note the similarity of the cytology of these cells to known gonadotropes. Their change with sexual maturation and their histochemistry are both consistent with this interpretation. CHAPTER II Rectal Gland Cells specialized for the active transport of excess cation are a common feature of marine FiGure 6. Latimeria chalumnae (Coelacanth). Schema- tized frontal section through conventional proximal pituitary complex in Latimeria. Level is ventral to median eminence. Residual hypophysial cavity (C) extends as a tubular process through pars distalis (PD) and continues rostrad. Note inter- mittent islet of rostral extension (= ventral lobe homologue) arising from tubular cavity. V, ventricle of neurohypophysis, NH; SV, saccus vasculosus; PI, pars intermedia. Dark shad- ing: connective tissue. Scale, 1 mm. vertebrates. The most widely distributed and best known are the so-called *‘chloride cells’ in the gills of fishes. Although originally described in teleosts, “‘chloride cells’*> occur among lam- preys, elasmobranchs and holocephalans as well as bony fishes. Such a wide distribution would suggest that this is a very plesiomorphic feature of vertebrates. Materials The present account is based on rectal gland gross anatomy, histology and ultrastructure of the following sharks: Squalus acanthias, Par- maturus xaniurus, Mustelus henlei, Charchar- odon charcharias, Carcharhinus melanopterus, C. amblyrhynchus, the holocephalan Hydrola- 32 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES. No. 134 FIGURE 7. Hydrolagus colliei (Ratfish). Midsagittal diagram of distal (postvalvular) hindgut. SV, spiral valve, RV, rectal valve (= distalmost plica of spiral valve), RG, intramural cation-excreting glands, homologue of those of elasmobranchs. Scale 0.5 mm. gus colliei, and the coelacanth Latimeria cha- lumnae. The sharks were made available to me courtesy of John E. McCosker, Steinhart Aquarium, California Academy of Sciences and John Stephens, Occidental College, Los Ange- les; Hydrolagus courtesy of Professor Stephens and George and Susan Brown, Department of Fisheries, The University of Washington, Seat- tle; and coelacanth material, courtesy of the late Earl S. Herald and the living John E. McCosker, Steinhart Aquarium; Robert Griffith, then of the Peabody Museum of Natural History, Yale Col- lege; James Atz and C. Lavett Smith, American Museum of Natural History; Charles Rand, Long Island University, New York; and Robert Lavenberg, Los Angeles County Museum of Natural History. The Chondrichthyes have an additional spe- cialized cation-excreting epithelium exclusive to the group, arranged as a gland derived from the terminal (post-valvular) hindgut. Among sharks this structure was recognized as the rectal, dig- itiform or cecal gland long before its function was realized. The gland exists in its simplest form among the Holocephali where it occurs as a circumferentially disposed group of separate racemose glands, each with its own duct, within the submucosa of the terminal hindgut. This por- tion of the gut is supplied by a discrete rectal artery which enters the dorsal aspect of the gut. In Hydrolagus colliei the majority of the sepa- rate submucosal glands lie in the dorsal portion of the gut wall, while in Chimaera monstrosa they are restricted to this site (Crofts 1925; Fange and Fugelli 1963; Lagios and Stasko-Con- cannon 1979). In Hydrolagus colliei a bluish discoloration of the serosa of the terminal hindgut, which ap- pears related to vascularization, belies the pres- ence of the intramural rectal glands in a nar- rowed post-valvular gut segment. Their location can be determined with certainty in the living specimen by noting the point of entry of the dor- sal rectal artery. Sections of the hindgut reveals discrete 1-2 mm yellow-beige submucosal glands in the bowel wall between the rectal valve proximally (the terminal plica of the spiral valve) and the thin, muscular segment of hindgut distally. The latter exits the body directly (Fig. 7); there is no cloaca. es) ies) LAGIOS: COELACANTHS AND CHONDRICHTHYES AS SISTER GROUPS FiGures 8 and 9. Fig. 8. Parmaturus xaniurus (Filetail Cat Shark). Schematic drawing of rectal gland relationships. Left lateral exposure, body wall excised, rectum (R) and liver lobe (L) retracted ventral. Rectal gland (RG) consists of a single fusiform gland supported by a dorsal mesorectum (M). Dorsal rectal artery enters gland at cephalic pole, then courses with duct of rectal gland to reach dorsal surface of rectum. Kidney (K). Scale 10 mm. Fig. 9. Latimeria chalumnae (Coelacanth). Sagittal diagram. 140 cm adult male. Rectum (R) is partially cut away to reveal position of rectal gland, RG, in relation to more distal cloaca, C. Bladder (B) is bilobed. K, Kidney; L, Liver; T, Testis and meso-orchium: LG, fat-filled vestigial lung. 34 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, No. 134 ces i‘ IP « hamagtens yatta in pa easels Pipaenmhel MAS G aie! yer oes, aon == ar ait aes iphiad ore se an 9S (10) ~. aii Ficures 10 and 11. Fig. 10. Latimeria chalumnae. Right lateral diagrammatic views of rectal gland in an adult male (LACM 6824-1). Distal rectum (R), rectal gland (RG), bladder (B) and ventral floor of abdomen have been hemisected to expose duct of rectal glance urse of right urethra into posterior rectal wall, and entrance of right ureter (arrow) in bladder. Right testis (T) and ductus deferens course ventral far posterior to rectum but exit was not traced in this specimen. Fig. 11. Latimeria LAGIOS: COELACANTHS AND CHONDRICHTHYES AS SISTER GROUPS 35 Among elasmobranchs the rectal cation-ex- creting tissue becomes a single, compact, extra- mural compound gland, lying in the dorsal me- sorectum and emptying by a single duct into the dorsal wall of the post-valvular gut (Fig. 8). The relationship of the gland and dorsal rectal artery, evident in the Holocephali, becomes clearer, since it directly supplies the gland. Despite the common appellation ‘‘digitiform™’ the majority of elasmobranch rectal glands are fusiform and appendix-like in shape. The extant coelacanth, Latimeria, has a sim- ilar gland, termed “‘post-anal’’ by Miullot and Anthony (1972, 1973a). The present description which differs from that of the latter authors, is based on the examination of three adult speci- mens: two large mature males (C59, 122 cm total length, California Academy of Sciences CAS 24862, and 121 cm total length, Los Angeles Museum of Natural History LACM 6824-1), and a mature ovoviviparous female (C26, 160 cm to- tal length, American Museum of Natural History AMNH 32949) and one of her contained em- bryos. Histologic material was available from both adult males and a second coelacanth em- bryo loaned by Charles S. Rand, M.D. Ultra- structural studies were limited to C68, the same specimen which provided the material for pub- lished ATPase studies (Griffith and Burdick 1976). The coelacanth rectal (“‘post-anal’’) gland lies within the dorsal mesorectum, close to the dor- sal wall of the posterior-most portion of the rec- tum, just as it angles steeply through the ventral body wall to exit. In the mature male, the an- terior face of the gland is adherent to the surface of the rectum itself (Fig. 9). A single duct leads from the ventral portion of the gland, paralleling the rectal contour, before entering the dorso- posterior wall of same, 3 cm proximal to the rectal orifice in C26 (Fig. 10). The anatomical relationships of the coel- acanth rectal (‘‘post-anal’’) gland are virtually indistinguishable from those of the rectal glands of sharks, although Millot and Anthony (1972) claim significant differences between the topog- raphy, anatomy and histology of the gland of Latimeria and those of elasmobranchs. They have interpreted the rectal (‘‘post-anal”’) gland of the coelacanth as emptying into a cloaca which receives a common urogenital papilla more distally on its posterior wall. The coel- acanth rectal gland is therefore not comparable to elasmobranchs in this view since it is not strictly rectal, that is proximal to the anal ori- fice. However, dissection of the LACM speci- men clearly shows paired urethrae entering the posterior wall of the rectum but separate from the ductus deferens (archinephric ducts) which course far caudad of the rectum. Dingerkus et al. (1978) have clearly shown a separate exit of the conjoined ductus deferens posterior to the rectum. Thus the male “cloaca” in Latimeria could not be confirmed. In the ovoviviparous AMNH female, the rectum, urethrae and ovi- duct exit separately into a shallow scale-bearing depression (Fig. 11) and the rectal gland clearly empties by a single median duct into the pos- terior wall of the rectum. Crofts (1925) has shown a similar close prox- imity of the rectal gland duct, ureters and ovi- ducts in Chlamydoselache and in Heptanchus embryos. In these sharks the rectal gland duct enters the rectum so posteriorly that its actual exit can be described as anal, i.e. the point of demarcation between the cloaca and terminal hindgut. Elasmobranchs show considerable variation in the relationships of the duct to the rectal wall, and this apparently relates to the length of the postvalvular segment (Crofts 1925) rather than differences in embryologic origin. Given this variation in elasmobranchs, and the greater differences in the homologous intramural glands of the Holocephali, the small differences between the gland in Latimeria and the predom- inant situation in sharks falls within the degree of variation evident in extant elasmobranchs, and well within the variation in Chondrichthyes. Millot and Anthony also cite the absence of a well developed central lumen in the coelacanth gland as evidence indicating non-homology with those of elasmobranchs, in which the lumen is — chalumnae. Left lateral diagrammatic view of rectal gland, R in ovoviviporous (AMNH) female specimen. M, mesorectum: B, left bladder; K, kidney; OV, dominant right oviduct viewed through rent in mesorectum. Inset, ventral view of vent, rostrad (superior in figure) is the large separate rectal opening, caudad, a funnel-shaped common oviduct and to either side the small paired urethral openings. OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, No. 134 Ficure 12. Coelacanth rectal gland EM. A. A group of tubules are seen in cross section. Note the numerous mitochondria (arrow). Nucleated red blood cells, R, lie in capillary between adjacent tubule profiles. N, nucleus of tubular cell, L, lumen. B. Basal cisternae of a rectal gland cell. Note deep infoldings of the basal cell membrane (arrow), producing the basal cisternae between which lic mitochondria, M. The ultrastructure is identical to other epithelia engaged in active ion transport and ts indistinguishable from rectal glands of the Chondrichthyes. LAGIOS: COELACANTHS AND CHONDRICHTHYES AS SISTER GROUPS 37 FiGure 13. Comparable magnifications of Ratfish (Hydrolagus colliei). Labels as in Fig. 12. 38 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, No. 134 FiGurReE 14. present. The latter would appear to be an em- bryologic remnant of the enteric diverticulum which produces the gland Anlage. However, among the Holocephali, the separate submuco- sal homologous rectal glands lack such a central gut lumen. The rectal glands of the Chondrichthyes and that of the extant coelacanth are virtually iden- tical in their histology and ultrastructure. The terminal, functional portions of the gland are composed of a single layered columnar epithe- lium which displays the characteristic arrange- ment of cell surface, organelles, and histoen- zymology of active cation transport. A complex system of basal and lateral cisternae, represent- ing deep plate-like parallel invaginations of the basal and lateral cell membranes, enormously increase the surface area in contact with the ex- tracellular compartment. These cisternae are as- sociated with numerous mitochondria, which generate the energy required to drive the active transport of cations. Junctional complexes, comprising multiple maculae adhaerentes and zonulae adhaerentes, bind the luminal aspects of the cell surfaces (Figs. 12A, B, 13A, B and Seven-gill Shark (Notorynchus maculatus) right, rectal gland. Labels as in Fig. 12. 14). In sharks the role of these glands in cation excretion has been confirmed by physiologic ex- periments. Similar experiments have not been performed on the Holocephali or the coelacanth, but both demonstrate similar sodium-potassium activated ATPase within the gland, comparable to that of the better known elasmobranch struc- ture (Griffith and Burdick 1976; Lagios and Stas- ko-Concannon 1979). CHAPTER III One of the first biochemical surprises to emerge from studies of the coelacanth was a high level of serum urea and hepatic-urea cycle enzymes comparable to those of elasmobranchs. At the time of this discovery (Brown and Brown 1967; Pickford and Grant 1967) however, it was interpreted as further evidence for an associa- tion of the coelacanth with the lungfish, Protop- terus, in which retention of urea is an important adaptation to estivation. Due to the prevailing bias which saw the coelacanth as “Old Four Legs,’’ this biochemical evidence tended to con- firm the systematic position of the coelacanth as LAGIOS: COELACANTHS AND CHONDRICHTHYES AS SISTER GROUPS 39 a member of the Sarcopterygil, a taxon created by the late Alfred Sherwood Romer. There are three basic methods of water and electrolyte regulation observed in living marine organisms. Simplest and most pervasive of these is the ionic regulation seen in all invertebrates and in the hagfishes—but not in lampreys. Plas- ma is iso-osmolar to seawater and contains very similar ionic concentrations. Divalent cations, such as calcium, magnesium, etc. are regulated as is sodium and potassium in the hag. Since this particular mechanism is so widely distributed among invertebrates, including the primitive chordates, and also present in the hags, it prob- ably represents an unspecialized, primitive or plesiomorphic condition. The second and next most pervasive mecha- nism of marine adaptation is characteristic of most gnathostomes (including modern bony fishes) and present in the more ancient lamprey. Tetrapods retain a similar osmolality and elec- trolyte composition. In this group, exemplified by modern marine teleosts, serum is maintained hypo-osmolar to the environmental salinity. Body water is obtained by drinking seawater and excreting large amounts of cations through a branchial excretory mechanism, the chloride cells of the gills. In euryhaline forms, both the total osmolality and the ionic composition are rigorously maintained independent of environ- mental salinity. The third, least common and the most restrict- ed in its distribution of the adaptive mechanisms is urea retention. This adaptation is seen only among chondrichthyian fishes—sharks, rays and chimaeras—and in the extant coelacanth among fishes with a long palaeontologic history of ma- rine residence. Large quantities of urea are re- tained in the blood to achieve an osmolality vir- tually identical to seawater. Excess salt is excreted through a rectal gland, although a bran- chial salt excreting apparatus and chloride cells are present in Chondrichthyes. The presence of similar cells in the coelacanth gill is highly prob- able but has not been confirmed. On the basis of its very limited distribution in marine vertebrates, virtually confined to a single group, the Chondrichthyes, and its absence in marine invertebrates, hagfish and lampreys, it would be parsimonious to accept urea retention as a specialized or apomorphic condition. Some, however, have argued that since other forms have similarly developed urea retention, specif- ically the marine and estuarian amphibians Bufo marinus and Rana cancrivora, that this feature is without systematic value. I disagree. Urea re- tention in two species of amphibians which can adapt to limited marine habitats is obviously a specialized secondary feature in that group and can best be interpreted as isolated convergence. The ability to synthesize urea is a primitive fea- ture in vertebrates; it is the retention of urea in marine environment which is apomorphic. Similarly, urea retention in the burrowing toad Scaphiopus and the modern estivating lungfish- es are specializations. The contrary argument that urea retention as a marine adaptation is primitive for sharks and coelacanths ignores the limited distribution of this biochemical mecha- nism and its absence in relic agnathans. Fur- thermore, there is evidence that protein function (specifically hemoglobin in sharks) is dependent on the maintenance of high levels of urea (Bon- aventura et al. 1974) which may account for the retention of serum urea of 170 mM/I in the freshwater Lake Nicaragua Bull Shark Carchar- hinus leucas (Thorson et al. 1973). In addition to retention of large amounts of urea, the chondrichthyian fishes and the coel- acanth share one other biochemical peculiarity not seen in either marine or estivating amphibia. This is a high level of trimethylamine oxide (TMAO) in both groups. Among sharks and the coelacanth the levels of TMAO contribute very significantly to the total serum osmolality. Con- version by liver explants of C'* labelled choline and TMA to TMAO in both shark and coel- acanth occur at rates sufficient to account for the blood levels. Although TMAO levels may also be substantial in marine teleosts, evidence from hatchery raised salmonids and cyprinids suggests that TMAO levels in these fishes may be entirely of dietary origin. Biochemical evi- dence of endogenous synthesis of TMAO in te- leosts 1s contradictory. Griffith and Burdick (1976) and Pang et al. (1977) have argued that despite the superficial similarities of urea retention among elasmo- branchs and the coelacanth, two significant dif- ferences exist: the coelacanth is slightly hypo- tonic (ca. 100 mOsm) to seawater sampled at the capture site; and the urea concentration in blad- der urine and serum were virtually identical. Pang et al. (1977) interpreted the relative serum hypotonicity of the specimen relative to sea- water as an indication of a fundamentally differ- 40 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, No. 134 ent method of adaptation. However their results are perhaps an artifact in that the specimen, a small, 85 cm immature female, remained in the surf at a 1 M depth in a chicken wire cage for 6-8 hours prior to serum sampling. The speci- men was caught during the monsoon season when large freshwater aquifers drain into the lit- toral, reducing the salinity considerably. Al- though these were not sampled by Griffith, I have encountered these same freshwater aqui- fers in the littoral and in deeper water of the reef face at 30-40 M depth (see McCosker, this vol- ume). Native Comorans employ this tidepool freshwater runoff to wash themselves and their clothing. The 100 mOsm difference between the serum of the damaged and dying specimen and that of the offshore water at its capture site may simply represent transient cation loss in hypo- tonic media. Similarly, much has been made of the high bladder urea level. This has been inter- preted to mean that coelacanths lack an active urea resorption. More probably, lack of urea re- sorption reflects the moribund condition of the specimen rather than fundamental differences in the physiology or urea excretion in sharks and coelacanths. CHAPTER IV Epple and Brinn (1975) in an extensive study of the relationships of islets of Langerhans among vertebrates point out another peculiar character shared by the Chondrichthyes and the coelacanth. In both groups the endocrine islet tissue occurs as a system of tubular structures within a compact extra-intestinal pancreas, an arrangement not identified in any other agnathan or gnathostome. Although the authors designate this arrangement as a ‘‘primitive gnathostome type,” it differs significantly from those seen in the extant Agnatha which shows islet tissue sep- arate from the exocrine pancreas, occurring within the intestinal mucosa in Myxinoidea and intramurally in the Petromyzontoidea. An un- expected result of their study was the parallel- ism between the groupings based on pituitary and islet organization. The same vertebrate groups which exhibit a specific organization of islet tissue exhibit a distinctive pattern of pitu- itary Organization. CONCLUSION shares with the Chondrich- thyes several unrelated, apomorphic and prob- haracters which are difficult 1© COCiaCanin to reconcile with the conventional systematic interpretation of the Actinistia as a sister group to any of the extant Osteichthyes, let alone the Rhipidistia. These characters are reflected in pituitary and endocrine islet organization, the development of a rectal cation excreting gland and urea retention in the marine environment. The pituitary complex is specialized in a man- ner entirely comparable to the Chondrichthyes, particularly the more plesiomorphic hexanchid sharks. Both groups share an independently vas- cularized, separate portion of the pars distalis which sequesters gonadotrope function and is closely associated with a carotid anastomosis. Further, both show a similar compartmentali- zation of the remainder of the distalis in refer- ence to acidophil cell types. The pituitary com- plex alone thus encompasses several independent apomorphic features. In a peculiar, parallel manner (Epple and Brinn 1975), both the extant coelacanth and the Chondrichthyes show a comparable and distinc- tive organization of the endocrine islet tissue, with duct-like islets distributed within a compact extra-intestinal pancreas. Evidence accumulat- ed from both living agnathan subgroups indicate that the coelacanth/chondrichthyian pattern of islet tissue is derived and distinct from that of other gnathostome groups. Both the coelacanth and the Chondrichthyes share the development of an accessory cation- excreting tissue derived from the rectum which augments salt-excretion by the more plesio- morphic branchial *‘chloride cells.’’ Both groups show a specialized urea retention in the marine environment, an adaptation not utilized by any other piscine vertebrate group albeit indepen- dently developed by two singular euryhaline, tropical amphibian species. Although multiple independent convergences may account for these shared derived characters, it is far easier to suggest that they represent synapomorphies. On the other hand, the two groups, coelacanth and Chondrichthyes, obviously differ in many respects and in numerous superficial features the coelacanth resembles the primitive Osteichthy- es. Striking among the latter are the presence of bone, scales, lobed fins, an opercular apparatus and a vestigial lung in the coelacanth and their absence in the Chondrichthyes. A number of these similar characters are decidedly plesio- morphic, however, and cannot contribute to a cladistic analysis. LAGIOS: COELACANTHS AND CHONDRICHTHYES AS SISTER GROUPS 4] Bone Bone is the oldest documented vertebrate tis- sue, and forms the only Ordovician evidence of the group. The absence of bone in the extant Chondrichthyes, notwithstanding Moss (1970) who argues for the presence of cellular mem- branous bone about teeth in Squalus acanthias, is a secondary or derived feature of the group. Jarvik (1977) has recently reappraised the Acan- thodii and has concluded that they and the ex- tant sharks are related and should be included in Chondrichthyes (Elasmobranchii). From the viewpoint of this analysis, early Chondrich- thyes, i.e., Acanthodii, had both dermal and en- dochondral bone documenting the presence of this calcified tissue in the ancestors of the extant Chondrichthyes. Scales Scales, once the hallmark of the Osteichthyes, would thus become a plesiomorphic character since a well developed shagreen of non-imbri- cating plates clothed the Acanthodii (Moy- Thomas and Miles 1971). Zangerl (1968) has not- ed in Pennsylvanian sharks composite ‘‘scales”’ of dermal membranous bone which supported elaborate clustered denticles. The latter pattern is not fundamentally dissimilar from the ‘‘scales”’ of Latimeria which represent thin lamellae of dermal bone with loosely attached odontodes (Smith et al. 1972; Miller, this volume). Lung Lungs, or more generally, air sacs, are widely distributed among Osteichthyes but also appear to have been present in placoderms, a group whose chondrichthian features have been de- bated (Miles and Young 1977). Thomson (1971) has theorized that the absence of lungs in extant Chondrichthyes may relate to their marine and free-swimming habitus. This same argument can account for the great reduction in calcified tissues and secondary loss of the airbladder as adapta- tions to a highly mobile and pelagic existence. Opercular Apparatus The absence of an opercular apparatus in the elasmobranchs may represent a specialization rather than retention of a more plesiomorphic condition. Holocephalans have an analagous opercular arrangement although it is clearly in- dependently derived and not homologous with those of the Osteichthyes (see W. Miller, this volume). It is not clear, however, whether the opercular apparatus in the coelacanth is homol- ogous with the osteichthyean structure. Intracranial Kinesis Apart from the general problem of the affini- ties of the coelacanth with the Osteichthyes, there is a more particular problem of their re- lationship to the Sarcopterygii (Romer 1966), a group including both the Dipnoi and the Rhipi- distia. The coelacanth and Rhipidistia were pre- viously considered to represent sister groups predominantly on the basis of a single character, the intracranial kinesis. The latter was consid- ered a synapomorphic feature which united the two otherwise superficially similar groups. Two recent developments contradict this interpreta- tion. Bjerring (1973) has shown that the intra- cranial kinesis in coelacanths and Rhipidistians is not homologous; it has different relationships to the segmentation of cranial somites in the two groups and thus is clearly not a synapomorphy. Additionally, Nelson (1969) has shown an anal- ogous system of subcephalic muscles which work the ventral portion of the kinesis in the extant polypterids, and Gardiner and Bartram (1977) have reviewed the subcephalic fissure in fossil palaeonisciforms, documenting the fact that such intracranial kineses were probably more widely distributed among palaeozoic fishes than previously suspected (Thomson 1969). During the course of reading Compagno’s re- buttal to this presentation, I became aware of Lgvtrup’s independent analysis of coelacanth relationships (Lg@vtrup 1977). This investigator came to the same conclusions regarding sister group status of the coelacanth and the Chon- drichthyes as the present author. Interestingly, he did so without reference to the structure of the pituitary gland or of the endocrine pancreas, which evidence would have strengthened his ar- gument. In sum then, recent palaentologic reappraisal has demonstrated that features once considered synapomorphic, thereby uniting the coelacanth with the extinct Rhipidistia, are in fact plesio- morphic characters. These more recent reap- praisals corroborate the conclusion obtained from the present investigations that the coel- acanth cannot be a sister group of the Rhipidis- tia, at least as judged by their more plesiomor- phic tetrapod descendents. The shared charac- 42 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, No. 134 ters between the coelacanth and the Chondrich- thyes, which I interpret as derived (apomor- phic), were completely unexpected, but repre- sent specializations in rather conservative organ systems—pituitary gland, the endocrine pan- creas, and in cation excretion and urea retention in the marine environment. Such signal special- izations shared only by the extant coelacanth and the Chondrichthyes strongly suggest that these are synapomorphic characters of great antiquity despite the traditional folklore of *“‘Old Four Legs,’ the pretetrapod cousin. LITERATURE CITED ACHER, R., J. CHAUVET, M. T. CHAUVET, AND D. CREPY. 1968. Molecular evolution of neurohypophysial hormones: comparison of active principles of the bony fishes. Gen. Comp. Endocrinol. 11:535—538. A.Lis, E. P., JR. 1912. The branchial, pseudobranchial and carotid arteries in Chimaera (Hydrolagus) colliei. Anat. Anz. 42:10-18. ALLUHON-GERARD, M. J. 1971. Types cellulaires et etapes de la différentiation de |’adénohypopophyse chez |’°embryon de rousette (Scyllium canicula, Chondrichthyens). Etude au microscope électronique. Z. Zellforsch. 120:525-545. BJERRING, H. C. 1973. Relationships of coelacanthiforms. In P. H. Greenwood, R. S. Miles, and C. Patterson, eds., Interrelationships of fishes. Zool. J. Linn. Soc. 53, Suppl. 1. Academic Press, London. BONAVENTURA, J., C. BONAVENTURA, AND B. SULLIVAN. 1974. Urea tolerance as a molecular adaptation of elasmo- branch hemoglobins. Science 186:57—59. Brown, G. W., JR., AND S. G. BRown. 1967. Urea and its formation in coelacanth liver. Science 155:570—573. Crorts, D. R. 1925. The comparative morphology of the cecal gland (rectal gland) of selachian fishes, with some ref- erences to the morphology and physiology of similar intes- tinal appendages throughout Ichthyopsida and Sauropsida. Proc. Zool. Soc. London, pp. 101-108. Derry, D. J., AND A. C. Jones. 1975. Effects of hypotha- lamic extracts, neurotransmitters and synthetic hypotha- lamic releasing hormones on adenyl cyclase activity in the lobes of the pituitary of the dogfish (Scyliorhinus canicula L.). J. Endocrinol. 64:49-57. DinGerkus, G., H. K. MoK, AND M. D. LaGios. 1978. The living coelacanth Latimeria chalumnae does not have a cloaca. Nature 276:261-262. Dosson, S., AND J. M. Dopp. 1977. Endocrine control of the testis in the dogfish Scyliorhinus canicula L. I. Effects of partial hypophysectomy on gravimetric, hormonal and biochemical aspects of testis function. Gen. Comp. Endo- crinol. 32:41-52. Dopp, J. M. 1975. The hormones of sex and reproduction and their effects in fish and lower vertebrates: twenty years on. /n E. J. W. Barrington and T. J. Griffith, eds., Trends in comparative endocrinology. Am. Zool. Suppl. Utica, New York. ,M.H.1. Dopp, AND N. JENKINS. 1979. Presence of a gonadotrophin in the Rachendach-hypophyse of the pi- tuitary gland of the rabbitfish Hydrolagus colliei, Abstract. Proc. Eur. Endocr. Soc., Sorento Meeting. Gen. Comp. Endocrinol. EppLe, A., AND J. E. BRINN, JR. 1975. Islet histophysiology: Evolutionary correlations. Gen. Comp. Endocrinol. 27:320- 349. FAHRENHOLZ, C. 1928. Eine Rachendachhypophyse bei Chi- maera monstrosa. Anat. Anz. 66:342-348. FANGE, R., AND K. FUGELLI. 1963. The rectal salt gland of elasmobranchs, and osmoregulation in chimaeroid fishes. Sarsia 10:27-34. FirtH, J.G., AND L. VOLLRATH. 1973. Determination of the distribution of luteinizing hormone-like activity within the dogfish pituitary gland by means of the Yenopus oocyte meiosis assay. J. Endocrinol. 58:347—348. FONTAINE, Y-A, AND E. BURZAWA-GERARD. 1977. Esquisse de l’evolution des hormones gonadotropes et thyretropes des verteébres. Gen. Comp. Endocrinol. 32:341-347. GARDINER, B. G., AND A. W. H. BARTRAM. 1977. The ho- mologues of ventral cranial fissures in osteichthyans. Jn S. M. Andrews et al., eds., Problems in vertebrate evolution. Linnean Soc. Symposium Series 4. Academic Press, New York. Goossens, N., K. DieRICKX, AND F. VANDESANDE. 1977. Immunohistochemical demonstration of the hypothalamo- hypophysial vasotocinergic system of Lampetra fluviatilis. Cell Tissue Res. 177:317-323. GoORBMAN, A. 1965. Vascualr relations between the neuro- hypophysis and adenohypophysis of cyclostomes and the problem of evolution of hypothalamic neuroendocrine con- trol. Arch. Anat. Microsc. Morphol. Exp. 54:163-194. GRIFFITH, R. W., AND C. J. BurpDICcK. 1976. Sodium-potas- sium activated adenosine triphosphatase in coelacanth tis- sues: high activity in rectal gland. Comp. Biochem. Phys- iol., B, 54:557-559. HANSEN, G. N. 1971. On the structure and vascularization of the pituitary gland in some primitive actinopterygians (Acipenser, Polyodon, Calamoichthys, Polypterus, Lepi- sosteus, and Amia). K. Dan. Vidensk. Selsk. Biol. Skr. 18:1-64 and 18 pls. HAYASHIDA, T. 1977. Immunochemical and biological stud- ies with growth hormone in a pituitary extract of the Coel- acanth, Latimeria chalumnae Smith. Gen. Comp. Endocri- nol. 32:221—229. , S. W. FARMER, AND H. PApPKOFF. 1975. Pituitary growth hormones: Further evidence for evolutionary con- servatism based on immunochemical studies. Proc. Nat. Acad. Sci. 72:4322—4326. , AND M. D. LaGios. 1969. Fish growth hormone: A biological, immunochemical, and ultrastructural study of sturgeon and paddlefish pituitaries. Gen. Comp. Endocri- nol. 13:403-411. HENDERSON, N. E. 1972. Ultrastructure of the neurohypo- physial lobe of the Hagfish, Epratretus stouti (Cyclosto- mata). Acta Zool. 53:243-266. Honma, Y. 1967. A consideration regarding the systematics of the holocephalian fish, from the endocrinological point of view. Proc. Jpn. Soc. Syst. Zool. 3:24-29. JACKSON, R. G., AND M. SaGe. 1973. Regional distribution of thyroid stimulating hormone activity in the pituitary gland of the Atlantic stingray, Dasyatis sabina. U.S. Nat. Mar. Fish. Serv. Fish. Bull. 71:93-97. JARVIK, E. 1977. The systematic position of acanthodian fish- es. In S. M. Andrews et al., eds., Problems in vertebrate LAGIOS: COELACANTHS AND CHONDRICHTHYES AS SISTER GROUPS 43 evolution. Linnean Soc. Symposium Series 4. Academic Press, New York. JASINSKI, A., AND A. GORBMAN. 1966. Hypothalamo-hypo- physial vascular and neurosecretory links in the Ratfish, Hydrolagus colliei (Lay and Bennett). Gen. Comp. Endo- crinol. 6:476—-490. KNOWLES, F., AND L. VOLLRATH. 1975. Cytology and neu- roendocrine relations of the pituitary of the dogfish, Scyl- iorhinus canicula. Proc. R. Soc. London, Ser. B, 191:507— 525. KosBaAyYASHI, H., AND H. UEMuRA. 1972. The neurohy- pophysis of the Hagfish, Eptatretus burgeri (Girard). Gen. Comp. Endocrinol. 3:114—124. Lacios, M. D. 1968a. Tetrapod-like organization of the pi- tuitary gland of the polypteriformid fishes, Calamoichthys calabaricus and Polypterus palmas. Gen. Comp. Endocrin- ol. 11:300-315. 1968b. The pituitary gland of the paddlefish, Poly- odon spathula. Copeia 1968(2):401—404. . 1970. The median eminence of the bowfin, Amia cal- va L. Gen. Comp. Endocrinol. 15:453—463. 1972. Evidence for a hypothalamo-hypophysial por- tal vascular system in the coelacanth Latimeria chalumnae Smith. Gen. Comp. Endocrinol. 18:73-82. . 1975. The pituitary gland of the coelacanth Latimeria chalumnae Smith. Gen. Comp. Endocrinol. 25:126-146. , AND S. STASKO-CONCANNON. 1979. Presumptive in- terrenal tissue (adrenocortical homolog) of the coelacanth Latimeria chalumnae. Gen. Comp. Endocrinol. 37:404— 406. Leypic, F. 1851. Ztir Anatomie und Histologie der Chimaera monstrosa. Arch. Anat. Physiol. Wiss. Med. 18:241-271. Legvtrup, S. 1977. The phylogeny of the vertebrata. John Wiley & Sons, London, p. 330. MELLINGER, J. C. A. 1960. Contribution a l’etude de la vas- cularisation et du développement de la region hypophysaire d'un selacien, Scyliorhinus caniculus L. Bull. Soc. Zool. Fr. 85:123-139. MELLINGER, J., AND M. P. DusBots. 1973. Confirmation, par limmunofluorescence, de la fonction corticotrope du lobe rostral et de la fonction gonadotrope du lobe ventral de lV hypophyse d'un poisson cartilagineux, la Torpille marbee (Torpedo marmorata). C. R. Acad. Sci., Ser. D, 276:1879- 1881. MEURLING P. 1967a. The vascularization of the pituitary in elasmobranchs. Sarsia 28:1—104. 1967b. The vascularization of the pituitary in Chi- maera monstrosa (Holocephali). Sarsia 30:83—106. Mices, R. S., AND G. C. YouNG. 1977. Placoderm inter- relationships reconsidered in the light of new ptyctodontids from Gogo, Western Australia. Jn S. M. Andrews et al., eds., Problems of vertebrate evolution. Linnean Soc. Sym- posium Series 4. Academic Press, New York. MILLoT, J., AND J. ANTHONY. 1955. Considerations phy- siomorphologiques sur la tete de Latimeria (Crossoptery- gien Coelacanthide). C. R. Acad. Sci. 241:114-116. , AND 1958. Crossopterygiens actuels. Pages 2553-2597 in P. P. Grassé, ed., Traité de Zoologie, Tome XIII, 3. Masson, Paris. , AND . 1960. Le cloaque chez les Coelacanthes. Bull. Mus. Nat. Hist. Nat. 32:287-289. , AND 1965. Anatomie de Latimeria chalum- nae. Tome II. Systeme nerveux et organes des sens. C. N. R. S., Paris. , AND 1972. La glande post-anale de Latime- ria. Ann. Sci. Nat. Zool. 14:305—-318. , AND 1973. Variations sexuelles de l’appareil excreteur du Coelacanthe, Latimeria chalumnae, Connex- ions avec l'appareil genital male. C. R. Acad. Sci. 276D. , AND . 1973. L’appareil excréteur de Latimeria chalumnae Smith (Poisson coelacanthide). Ann. Sci. Nat. Zool. 15:293-328. : , AND D. RoBINEAU. 1972. Etat commenté des captures de Latimeria chalumnae Smith (Poisson, crossop- terygien, coelacanthide) effectuees jusqu-au mois d’ octobre 1971. Bull. Mus. Nat. Hist. Nat. 53:533-548. Moss, M. L. 1970. Enamel and bone in shark teeth: with a note on fibrous enamel in fishes. Acta Anat. 77:161—-187. Moy-THomas, J. A., AND R. S. Mices. 1971. Paleozoic fish- es, 2nd ed. W. B. Saunders Co., Philadelphia. Netson, G. J. 1969. Gill arches and the phylogeny of fishes, with notes on the classification of vertebrates. Bull. Am. Mus. Nat. Hist. 141:475—552, 91 plates. Norris, F. W. 1941. The plagiostome hypophysis. Grinell College, Grinell, Iowa. PANG, P. K. T., R. W. GRIFFITH, AND J. W. Artz. 1977. Osmoregulation in elasmobranchs. Am. Zool. 17:365-377. PickForDb, G. E., AND F. B. GRANT. 1967. Serum osmolality in the coelacanth Latimeria chalumnae: urea retention and ion regulation. Science 155:568—570. PoLeNov, A. L. 1968. On the evolution of the median em- inence—the proximal neurosecretory contact region in some fishes (Elasmobranchii, Chondrosteoidei). Arch. Anat. Histol. Embryol. 53:553—S61. Romer, A. S. 1966. Vertebrate paleontology, 3rd ed. Uni- versity of Chicago Press, Chicago, Illinois. 1968. Notes and Comments on Vertebrate Paleon- tology. University of Chicago Press, Chicago, Illinois. SATHYANESAN, A. G. 1965. The hypophysis and hypothala- mo-hypophysial system in the chimaeroid fish Hydrolagus colliei (Lay and Bennett) with a note on their vasculariza- tion. J. Morphol. 116:413-499. SAwYeER, W. 1969. The active neurohypophysial principles of two primitive bony fishes, the bichir (Polypterus sene- galis) and the African lungfish (Protopterus aethiopicus). J. Endocrinol. 44:421-435. ScANeEs, C. B., S. DoBson, B. K. FoLLott, AND J. M. Dopp. 1972. Gonadotropic activity in the pituitary gland of the dogfish (Scyliorhinus canicula). J. Endocrinol. 54:343-344. SmitH, M. M., M. H. HoBDELL, AND W. A. MILLER. 1972. The structure of the scales of Latimeria chalumnae. J. Zool. London 167:501—S09. THomson, K. S. 1969. The biology of the lobe-finned fishes. Biol. Rev. Cambridge Philos. Soc. 44:91-154. 1971. The adaptation and evolution of early fishes. Q. Rev. Biol. 46: 139-166. THORSON, T. B., C. M. CowAN, AND D. E. Watson. 1973. Body fluid solutes of juveniles and adults of the euryhaline Bullshark Carcharhinus leucas from freshwater and saline environments. Physiol. Zool. 46:29—42. TsSUNEKI, K., AND A. GORBMAN. 1975. Ultrastructure of the anterior neurohypophysis and the pars distalis of the lam- prey, Lampetra tridentata. Gen. Comp. Endocrinol. 25:487—-S08. VAN KEMENADE, J. A. M., AND J. W. KREMERS. 1975. The pituitary gland of the coelacanth fish Latimeria chalumnae 44 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, No. 134 Smith: General structure and adenohypophysial cell types. Cell Tissue Res. 163:291-311. VOLLRATH, L. 1967. Uber die Neurosekretorische Innerva- tion der Adenohypophyse von Teleostiern, insbesondere von Hippocampus kuda und Tinca tinca. Z. Zellforsch. Mikrosk. Anat. 78:234—260. WALLis, M. 1975. The moiecular evolution of pituitary hor- mones. Biol. Rev. Cambridge Philos. Soc. 50:35—98. WINGSTRAND, K. G. 1966. Microscopic anatomy, nerve sup- ply and blood supply of the pars intermedia. Pages 1—27 in G. W. Harris and B. T. Donovan, eds., The Pituitary Gland III. University of California Press, Berkeley, California. ZAMBRANO, D. 1971. The nucleus lateralis tuberis system of the gobiid fish, Gillichthys mirabilis. 11. Innervation of the pituitary. Z. Zellforsch. 110:496—516. ZANGERL, R. 1968. The morphology and developmental his- tory of the scales of the Paleozoic sharks Holmesella and Ordus. In Current problems of lower vertebrate phylogeny. Proceedings of 4th Nobel Symposium. Interscience Pub- lishers, New York. OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES The Biology and Physiology of the Living Coelacanth No. 134, 8 pages December 22, 1979 COELACANTHS: SHARK RELATIVES OR BONY FISHES? By Leonard J. V. Compagno Department of Biological Sciences, Stanford University, Stanford, California 94305 Because of their extensive fossil record, stretching in time from the middle Devonian to the upper Cretaceous, coelacanths (Actinistia) were relatively well-known morphologically be- fore the capture of the sole living representative of the group, Latimeria chalumnae, in 1938. Before and since that time coelacanths were generally classified as lobe-finned fishes or Cros- sopterygia, along with the extinct rhipidistian fishes, and placed in the class (or subclass) Os- teichthyes or bony fishes. However, recent in- vestigations on the physiology and soft-part morphology of Latimeria have suggested a rad- ically divergent interpretation of their relation- ships to some observers, as a sister-group to the class Chondrichthyes (Fig. 1) or cartilaginous fishes (sharks, rays and chimaeras). L@vtrup (1977) listed a number of characters common to Chondrichthyes and Actinistia, which he im- plied were shared derived characters of the two groups. These include common presence of a fatty liver; a rectal gland; similarities in eye structure; similar patterns of gray matter distri- bution and absence of Mauthner’s fibers in the spinal cord; stalked olfactory bulbs in the brain; similar structure of the thymus and thyroid glands; very large eggs; and osmoregulation through retention of high concentrations of urea and trimethylamine oxide in the blood plasma 45 and tissue fluids. Lagios (this volume) points to detailed similarities in pituitary gland anatomy, histology and histochemistry between Latimeria and cartilaginous fishes, which he regards as shared derived characters of Chondrichthyes and Actinistia. Lgvtrup (1977) was sufficiently impressed by these similarities in soft-part morphology to sug- gest that no serious alternative could be found to ranking the Actinistia and Chondrichthyes as sister groups. However, this is true only if evi- dence from skeletal morphology is ignored or slighted, evidence which is contrary to the hy- pothesis of a chondrichthyian relationship for coelacanths but strongly supports their conven- tional placement (Romer 1966; Moy-Thomas and Miles 1971; Andrews 1973; Miles 1977) in the class Osteichthyes as ‘fleshy-finned’ fishes or Sarcopterygii (a group including the lungfish- es or Dipnoi, the coelacanths, and the rhipidis- tians). One practical difficulty in using soft-part mor- phology in interpreting the phylogeny of gnatho- stome fishes is that it restricts the number of groups that can be compared. No information on the soft-part characters cited above exists for the classes Placodermi or Acanthodii. In the class Osteichthyes there is no information on these characters for early lungfishes, primitive 46 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, No. 134 ACANTHODII ACTINISTIA ee ee THYES ACTINOPTERYG!I DIPNO! POROLEPIFORMES OSTEOLEPIFORMES Figure 1. Cladogram illustrating the hypothesis of im- mediate sister-group relationship of coelacanths (Actinistia) and chondrichthyians, after Lgvtrup (1977, figs. 5.7-5.8). Placodermi were not placed in this cladogram, but other major groups of gnathostome fishes are included. ray-finned fishes (palaeoniscoid actinopterygi- ans) or rhipidistians, although their condition in the first two groups can be inferred from living lungfishes, sturgeons (acipenserids), paddlefish- es (polyodontids) and bichirs (polypterids). In the case of rhipidistians, or at least Osteolepi- formes (which most writers agree gave rise to tetrapods or land vertebrates), the states of these soft-part characters can be inferred from living tetrapods (especially amphibians), but with the danger that many of them were changed in the transition from water to land and by sub- sequent evolution on land. In contrast, the skel- etal morphology of gnathostome fishes, includ- ing extinct groups, has been intensively investigated in the past 100 years and is rela- tively well known. Most of this account is a summary of some of the evidence from skeletal morphology and oth- er characters that indicates an osteichthyian and sarcopterygian placement of coelacanths. Some coelacanth characters are listed that occur in other osteichthyians and sarcopterygians, are lacking in chondrichthyians and often acantho- dians and placoderms, and which are probably shared derived characters that indicate a com- mon ancestry for coelacanths and other sarcop- terygians and osteichthyians. DERIVED OSTEICHTHYIAN CHARACTERS OF COELACANTHS Osteichthyian characters of coelacanths that are apparently shared derived characters of the class Osteichthyes were compiled from Schaef- fer (1968), Moy-Thomas and Miles (1971), Miles (1973), and Gardiner (1973). Some of these were used by Miles (1968, 1973) and Moy-Thomas and Miles (1971) to support the placement of the spi- ny-finned Acanthodii as the sister group of Os- teichthyes in a common group of Teleostomi. However, Jarvik (1977) has recently challenged the teleostome hypothesis and presented evi- dence that he interprets as supporting an elas- mobranch (shark and ray) or shark relationship for acanthodians. It is beyond the scope of this account to consider the relative merits of an elasmobranch or osteichthyian relationship for acanthodians, except to note that some of the characters cited by Jarvik as uniting sharks and acanthodians may be primitive gnathostome or chondrichthyian + teleostome characters, while others (especially similarities between the orbit- al palatoquadrate articulations in squalomorph sharks and acanthodians) may be convergent between these groups. To avoid taking up the question of acanthodian relationships here, tel- eostome derived characters are included with the following list of derived characters shared by coelacanths and other osteichthyians: 1. The neurocranium or braincase has a pair of lateral occipital fissures separating the occip- ital region from the otic region, which are pres- ent in acanthodians but apparently absent in chondrichthyians and placoderms. 2. The neurocranium has a ventral fissure that basally and transversely divides it into anterior prechordal and posterior parachordal parts, which is also present in acanthodians but absent in chondrichthyians and placoderms. 3. The neurocranium has a pair of spiracular grooves in the basisphenoid region, which are lacking in chondrichthyians and placoderms and possibly present in acanthodians. 4. The neurocranium is more or less tropi- basic, with a narrow basal plate, no suborbital shelves, and relatively compressed orbital walls in the orbital region. Acanthodians also have tropybasic neurocrania, but placoderms and primitive chondrichthyians have broad-based or platybasic crania, with a broad basal plate, sub- orbital shelves, and wide-spaced orbital walls. 5. There is a skull roof of large and small der- mal bones superficially covering the neurocra- nium, which has a similar pattern of homologous elements in different osteichthyian groups. Apart from the dermal bones of the secondary upper jaw, opercle, and palate, these include small rostral bones on the dorsal surface of the snout; a paired longitudinal series on the dorsal surface of the head, including nasals, frontals, parietals, and postparietals; circumorbital bones, COMPAGNO: COELACANTHS: SHARKS OR BONY FISHES? 47 including lachrimals, infraorbitals and postorbit- als: small elements bordering the posterior mar- gin of the skull roof, the extrascapulars; ele- ments lateral to the longitudinal, dorsal series, the supratemporals, intertemporals, and tabu- lars; and elements of the cheek region anterior to the opercle, the squamosals and quadratoju- gals. There are problems in homologizing some bones of the coelacanth skull roof with their equivalents in actinopterygians and rhipidistians (Schaeffer 1952; Moy-Thomas and Miles 1971), but those definitely present include rostrals, postrostrals, a frontonasal series, postparietals (intertemporals), tabulars, extrascapulars, lac- rimojugals, postorbitals, and squamosals. Acan- thodians primitively have the head covered with numerous small dermal bones resembling the small, non-imbricated scales of the body, chon- drichthyians primitively have the head covered with dermal denticles similar to those elsewhere on the body, while placoderms have a skull roof of large bony plates that are not homologous with those of osteichthyians. 6. Dermal bony plates are present on the pal- ate, including paired vomers and a medial para- sphenoid bone (absent in non-osteichthyians). 7. The palatoquadrates, or primary upper jaws, have basal or basitrabecular articulations with the neurocranium, posterior to their pala- tine (or orbital) articulations with the cranium. Early coelacanths have the basitrabecular artic- ulation but later ones (including Latimeria) have lost it. Sharks primitively have orbital articula- tions with the cranium, but these occupy a ba- sitrabecular position in many squalomorph sharks, and are lost in rays. Chimaeras (Holo- cephali) and lungfishes have the palatoquadrates fused to the cranium, and hence have any joints with the cranium obliterated. The nature of the acanthodian palatoquadrate articulation is dis- puted (see Jarvik 1977). 8. The palatoquadrates are expanded medi- ally to form most of the palatal roof. These are more laterally restricted in Chondrichthyes, but their condition is uncertain in Acanthodii. The palatoquadrates of placoderms are often highly modified (see Stensid 1963), but apparently are not expanded medially also. 9. The palatoquadrates are functionally re- placed by dermal bones lateral to them that form part of the superficial dorsal armor of the skull, the maxillae and premaxillae. These form sec- ondary upper jaws, with an arcade of teeth lat- eral to those of the palatoquadrates. However, maxillae were apparently lost in coelacanths (which have premaxillae), premaxillae were also lost in Dipnoi, and chondrichthyians, placo- derms and acanthodians lack these bones. Plac- oderms have upper tooth plates (anterior and posterior superognathals) that are not homolo- gous but are functionally similar to maxillae and premaxillae. 10. The primary lower jaws, or Meckel’s car- tilages, are functionally replaced by a number of dermal and cartilage-replacement bones forming a similar pattern in osteichthyian groups. These include the tooth-bearing dentaries, splenials, prearticulars, coronoids, articulars, surangulars, and angulars (the last two fused in coelacanths). All of these are absent in acanthodians, chon- drichthyians and placoderms, although many of the latter have lower tooth plates (inferogna- thals) that partially replace the Meckel’s carti- lages. 11. Dorsal elements of the hyoid visceral arch are modified to form a hyomandibula, with a proximal or dorsal articulation on the lateral commissures of the neurocranium, lateral to and partly or entirely above the lateral head vein. A hyomandibula is absent in holocephalians but present in elasmobranchs, at least some placo- derms, and acanthodians. The neurocranial ar- ticulation of the hyomandibula is below the lat- eral head vein in placoderms and elasmobranchs, but its position is disputed in acanthodians. The different articulations of the hyomandibula in elasmobranchs and osteichthyians and its ab- sence in holocephalians (in which the dorsal ele- ments of the hyoid arch are unmodified and sim- ilar to dorsal branchial arch elements) suggest that the hyomandibula was separately evolved at least in chondrichthyians and osteichthyians. 12. The hyoid arch has an intermediate ele- ment (interhyal or stylohyal) between the hy- omandibula and ceratohyal on each side, a sym- plectic below the hyomandibula and lateral to the interhyal, a hypohyal between basihyal and ceratohyal, and a glossohyal behind the basi- hyal. An interhyal may be present in acantho- dians, but interhyals, symplectics, hypohyals and glossohyals are otherwise absent in non-os- teichthyians. 13. The gill cover or hyoid septum of the hyoid arch is greatly expanded posteriorly as an operculum that covers all the gills. It has at least two large dermal bones, the opercular and sub- 48 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, No. 134 opercular, and branchiostegal plates ventrally. It functionally replaces the interbranchial septa of the gills (which are greatly reduced) as a pumping element of the respiratory apparatus. Elasmobranchs lack an expanded hyoid oper- culum but have separate external gill openings and unreduced interbranchial septa (probably a primitive condition among gnathostome fishes). Acanthodians have a partial hyoid operculum with small bony plates similar to branchioste- gals, holocephalians have a soft hyoid opercu- lum covering the partially reduced interbranchi- al septa, and placoderms have an internal gill chamber with a single external aperture on each side, but none of these fishes have an operculum of osteichthyian type. 14. The gill arches have suprapharyngobran- chials dorsally in addition to the infrapharyn- gobranchials present in other fishes. 15. The shoulder girdle has a series of dermal bones, attached to the scapulocoracoid or pri- mary shoulder girdle and functionally replacing it in part, with a dorsal attachment on each side to the dermal cranium. The dermal shoulder gir- dle of osteichthyians includes posttemporals, supracleithra, cleithra, clavicles, and in some groups anocleithra and postcleithra. The scapu- locoracoid only is present in acanthodians and chondrichthyians, but in placoderms an analo- gous but very different dermal shoulder girdle is present, which probably evolved separately from the osteichthyian one. 16. The incurrent and excurrent apertures of the nostrils are separate openings. In chondrich- thyians and probably placoderms these are parts of a single opening, but their condition is uncer- tain in acanthodians. 17. A swim bladder or lungs are present, but lacking in chondrichthyians and uncertain in placoderms or acanthodians. 18. The vertebrae have dorsal ribs, which are absent in non-osteichthyians as far as is known. 19. The fins are primarily supported by bony fin rays or lepidotrichia, which are absent in non-osteichthyians. 20. The scales are relatively large, bony plates that are arranged in diagonal rows on the body, with the anterior ends of the scales of each row deeply inserted in the skin beneath the pos- terior ends of the scales of the next most anterior row (imbric. ted). Acanthodians have small bony scales and chondrichthyians dermal denticles, in both groups sot imbricated or in diagonal rows. Placoderms are naked or have small to large scales that are not imbricated and either not in rows or in transverse rows. 21. At least some cranial elements are endo- chondrally ossified, in addition to the perichon- dral ossification found in non-osteichthyians with bone in the endoskeleton. Chondrichthyi- ans lack perichondral bone but have it replaced by perichondral calcified nodules. 22. The inner ear has large, regular otoliths, possibly also present in acanthodians. DERIVED SARCOPTERYGIAN CHARACTERS OF COELACANTHS - Within the Osteichthyes coelacanths are mor- phologically closest to the rhipidistian fishes, in- cluding the Osteolepiformes, Rhizodontiformes, Porolepidormes, and Onychodontiformes, and with them are generally placed in a major group (often ranked as a subclass or superorder of Os- teichthyes), the Crossopterygii. The lungfishes or Dipnoi, although specialized and aberrant in many characters, are morphologically closer to rhipidistians and coelacanths than actinopteryg- ians and are sometimes included with coel- acanths and rhipidistians in a common group, the Sarcopterygii (Romer 1966), which is ranked as a subclass of Osteichthyes coordinate with Actinopterygii. Derived sarcopterygian charac- ters linking these fishes are compiled from Ro- mer (1966) and Miles (1977): 1. The neurocranium has a lateral crest on each side that supports the edges of the dermal skull roof, roofs the gill chambers, and laterally defines the dorsal or supraotic muscle fossae. 2. The basal plate of the neurocranium lacks aortic canals, which are present in early acti- nopterygians as well as acanthodians, placo- derms, and early chondrichthyians. 3. The neurocranium has an adotic process on the lower edge of the jugular groove near the first gill arch. 4. The head of the hyomandibula has a broad articular condyle or pair of condyles, with a lat- eral articulation on the lateral commissure that straddles the lateral head vein. 5. Epaxial lepidotrichia are present in the caudal fin, forming a variably developed epi- chordal lobe (absent in Actinopterygii). 6. The pectoral fins, and usually the pelvic fins also, have fleshy, lobate bases. The pectoral basals and radials are very narrow and project into the fin. Paired fins are primitively broad- COMPAGNO: COELACANTHS: SHARKS OR BONY FISHES? 49 based and not lobate in actinopterygians, acan- thodians, early chondrichthyians, and apparent- ly placoderms, but lobate fins have indepen- dently and repeatedly evolved in several actinopterygian groups, in extinct xenacanth sharks and some neoselachian groups, in the ex- tinct holocephalian Chondrenchelys and living chimaeras, and in antiarch placoderms. There are a number of derived characters in coelacanths that suggest a close relationship to rhipidistians and which are crossopterygian characters of these groups. These are compiled from Millot and Anthony (1958), Moy-Thomas and Miles (1971), Gardiner (1973), and Andrews (1973): 1. The lateral occipital fissures of the neuro- cranium are partially filled. 2. The neurocranial ventral fissure is expand- ed and modified as an intracranial joint, with the notochord enlarged anteriorly and expanded for- ward to the basisphenoid ossification, the basi- occipital bone laterally excavated, and a ventral basicranial muscle present and connecting the cranial halves. 3. The neurocranium lacks an ossified prootic bridge. 4. The hyomandibula has two articular con- dyles, a dorsal one and a ventral one, the latter opposite the lateral head vein. Also, Andrews (1977) notes that the vertebral structure of coelacanths, as exemplified by Lar- imeria, is basically similar to that of rhipidis- tians, and differs from that of other fishes, but she was uncertain whether this common verte- bral type is a shared derived character of cros- sopterygians. The exact relationship of coelacanths to other sarcopterygians remains uncertain. Three hy- potheses of coelacanth relationships seem most likely and are expressed here as cladograms (Figs. 2-4). The first, suggested by the classifi- cation of Romer (1966), is that rhipidistians and actinistians are sister groups (Fig. 2), both groups are sister to the Dipnoi, and that all three groups (the Sarcopterygii) are sister to the Ac- tinopterygii. The grouping of rhipidistians and coelacanths in a common crossopterygian group primarily depends on the interpretation of the intracranial joint as a unique derived character of the Crossopterygii, with the condition of the ventral fissure of early Dipnoi considered prim- itive. The ranking of rhipidistians and coela- canths as sister groups rests on the presence of ACTINOPTERYGII DIPNOI SARCOPTERYGI! CROSSOPTERYGI ACTINISTIA RHIPIDISTIA FiGuRE 2. Cladogram of osteichthyian relationships, illus- trating the hypothesis of sister-group relationship of coel- acanths (Actinistia) and rhipidistians; from the classification of Romer (1966). many divergent characters in coelacanths, in- cluding their lack of choanae, reduced dermal cheek bones, lack of maxillae, quadratojugals, and mandibular bones, reduction in number and increase in size of sensory canal pores, devel- opment of dorsal antotic articulations of the pal- atoquadrates to the neurocranium, presence of extracleithra in the dermal shoulder girdle, and presence of relatively few, unbranched lepido- trichia in the fins (listed from Andrews 1973). However, many of these characters, except per- haps absence of choanae, may be unique derived characters of coelacanths and may not preclude the descent of coelacanths from some rhipidis- tian group. Andrews (1973) accepted the grouping of coel- acanths and rhipidistians as crossopterygians, but did not discuss the relationships of crossop- terygians to other osteichthyians. She suggested a new hypothesis (Fig. 3), that coelacanths are the sister group of porolepiform rhipidistians and that both form the sister group of osteo- lepiform, rhizodontiform, and onychodontiform rhipidistians. This was based on the common presence of a specific ‘binostian’ type of ‘‘skull table’’ (the postparietal, tabular and extratem- poral bones of the dermal skull roof) in Porolep- iformes and Actinista, as contrasted by the ‘quadrostian’ type in Rhizodontiformes, Osteo- lepiformes and Onychodontiformes. Andrews also mentioned primitive, undescribed Devonian coelacanths with rhipidistian or even porolepi- form characters, but did not consider coel- acanths as straight porolepiform derivatives be- cause known coelacanths showed no sign of loss of dentine folding in their teeth (folding present in porolepiform teeth), symphysial tooth whorls of Porolepiformes, and the elongated, dipnoan- like pectoral fin axes of Porolepiformes (all of which characters she considered as unique de- rived features of Porolepiformes). Miles (1977), in a recent study of primitive Devonian lungfishes, suggested a third alterna- 50 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, No. 134 TINISTIA BINOSTIA porns POROLE PIFORMES ONYCHODONTIFORMES QUADROSTIA RHIZODONTIFORMES OSTEOLEPIFORMES FiGure 3. Cladogram of crossopterygian relationships, il- lustrating the hypothesis of sister-group relationship of coel- acanths (Actinistia) and Porolepiformes, after Andrews (1973, fig. 5). tive, that Dipnoi and rhipidistians (or “‘choa- nates’’) are sister groups, both form a sister group to coelacanths, and that all three groups (the Sarcopterygii) are the sister group of Actin- opterygii (Fig. 4). He bases his hypothesis of interrelationships within the Sarcopterygii on the shared possession in early lungfishes and rhipidistians of a supraotic cavity in the neuro- cranium, cosmine on the scales and dermal bones, and submandibular bones, which he re- gards as shared derived characters of these groups that are lacking in coelacanths. This hy- pothesis requires the intracranial joint of coela- canths and rhipidistians (as well as other *‘cros- sopterygian’’ characters of these groups listed above) to be primitive for sarcopterygians (and a unique derived character of the group), and that lungfishes have secondarily lost the joint. However, the condition of the ventral fissure in the primitive lungfishes studied by Miles does not suggest secondary loss of an intracranial joint, but rather the primitive osteichthyian or even teleostome condition. This would support the hypothesis that the intracranial joint of coel- acanths and rhipidistians is a shared derived character of these fishes, but requires that coel- acanths would have secondarily lost the su- praotic cavity, submandibular bones, and cos- mine of lungfishes and rhipidistians, and that these characters be considered as primitive ones for Sarcopterygii. Miles (1977) initially support- ed the hypothesis (Fig. 2) ranking coelacanths and rhipidistians as sister groups, but eventually supported the hypothesis of lungfishes and rhip- idistians as sister groups (Fig. 4). He rejected Andrews’ (1973) arrangement of Porolepiformes and Actinista as sister groups because of the ab- sence of the characters uniting lungfishes, Po- rolepiformes, and other rhipidistians in coel- acanths. Miles accepted Andrews’ phylogeny of rhipidistians (with coelacanths deleted) and sug- gested that the ‘binostian’ condition in Porole- ACTINOPTERYGII | ACTINISTIA a RHIPIDISTIA FIGURE 4. trating the hypothesis of sister-group relationship of coel- acanths (Actinistia) and rhipidistians + Dipnoi, after Miles (1977, fig. 158). Cladogram of osteichthyian relationships, illus- piformes and coelacanths may be a primitive condition of sarcopterygians and hence of no weight in supporting the placement of these fish- es In a Common group. A radically different interpretation of coel- acanth relationships was proposed by Bjerring (1973), who thought that coelacanths are not re- lated to rhipidistians because their intracranial joints are not homologous. His arguments are complex and detailed but primarily hinge on his interpretations of the embryonic components of the posterior part of the neurocranium in these fishes, based in turn on extrapolations from cra- nial development of other fishes and other ver- tebrates and hypotheses on the vertebral com- ponents of the gnathostome neurocranium (but not on direct morphogenetic evidence from the cranium of coelacanths or rhipidistians). It also depends on his view that an intracranial! joint of crossopterygian type (or types), the associated basicranial muscle, and the greatly expanded anterior end of the notochord in coelacanths and rhipidistians are primitive gnathostome charac- ters, and that their absence in other gnathostome fishes is derived. However, the presence of a ventral fissure but absence of an intracranial joint in early actinopterygians (Gardiner 1973; Gardiner and Bartram 1977) and Dipnoi (Miles 1977, but see above), the absence of the joint in acanthodians (Miles 1968, 1973; Jarvik 1977), placoderms (Stensid 1963; Moy-Thomas and Miles 1971) and all known chondrichthyians, and the suite of derived osteichthyian characters uniting the Dipnoi, Actinista, Rhipidistia, and Actinopterygii all strongly suggest that an intra- cranial joint of crossopterygian (or sarcopteryg- ian) type is derived (Gardiner 1973; Miles 1977). The observed morphological differences be- tween rhipidistian and coelacanth intracranial joints may support the hypothesis that coel- acanths cannot be derived from any known rhip- idistian type, without negating their placement COMPAGNO: COELACANTHS: SHARKS OR BONY FISHES? 5] as sister groups of either the Rhipidistia or Rhip- idistia + Dipnoi. OTHER CHARACTERS SEPARATING COELACANTHS AND CHONDRICHTHYIANS In addition to shared derived characters plac- ing them with Sarcopterygii and Osteichthyes, coelacanths have many other differences from Chondrichthyes. These include unique derived characters of Chondrichthyes as well as char- acters of uncertain polarity. The following list is hardly exhaustive as it includes mostly skel- etal characters and can be expanded by com- parisons of other systems in coelacanths and chondnrichthyians. 1. Chondrichthyian neurocrania have lateral- ly expanded ethmoid regions with globular nasal capsules but no medial rostral cavity and organ, the cranial cavity communicating with the ex- terior through an anterior fontanelle, and an en- dolymphatic or parietal fossa on the cranial roof, with paired perilymphatic and endolymphatic foramina. Coelacanth crania have a medial ros- tral cavity and organ in the ethmoid region, which is not expanded laterally with globular nasal capsules, no anterior fontanelle, and no parietal fossa on the cranial roof. 2. In chondrichthyians the inner ears com- municate with the exterior through paired en- dolymphatic pores on the dorsal surface of the head, but not in coelacanths. 3. The palatoquadrates of chondrichthyians usually have anteriomedial palatine processes that meet in an upper symphysis at the midline of the mouth, but those of coelacanths (and oth- er osteichthyians) lack these processes and do not meet in an upper symphysis. The palato- quadrates of coelacanths (and rhipidistians) have several parts separated by sutures, those of chondrichthyians form a single unit. The man- dibular joint (posterior joints of palatoquadrates and Meckel’s cartilages) is double in chondrich- thyians (and acanthodians) but single in coela- canths (and other osteichthyians). 4. The gill arches of coelacanths (and oste- ichthyians generally) have small bony plates with small teeth, while those of chondrichthyi- ans have dermal denticles, denticle gill rakers, papillar gill rakers, or chondropapillar gill rak- ers. The basibranchial skeleton of coelacanths has a single basibranchial element present, but multiple basibranchials are primitively present in chondrichthyians. 5. The nostrils are ventral on the snout of chondrichthyians (as some placoderms and pos- sibly acanthodians), dorsolateral in coelacanths (as in other osteichthyians generally). 6. Chondrichthyians generally have subter- minal mouths, coelacanths terminal mouths (as in other osteichthyians generally). 7. Chondrichthyians have highly differentiat- ed teeth with distinct roots and crowns, bound to the primary jaws by the dental membrane only and arranged in transverse rows or tooth families. Coelacanths have simple conical teeth that are fused to the jaw bones and are not ar- ranged in rows. 8. Chondrichthyians have paired intromittant organs or claspers developed as posterior exten- sions of the pelvic fins, which are lacking in coelacanths (and other osteichthyians). 9. Chondrichthyians have bone, when pres- ent, confined to the roots of denticles and teeth but otherwise lack it; it is replaced by calcified nodules in the endoskeleton. Coelacanths, as with other osteichthyians, placoderms, and acanthodians, have bone in the endoskeleton and the external dermal covering. FALSIFICATION OF THE HYPOTHESIS OF COELACANTH-CHONDRICHTHYIAN RELATIONSHIPS The numerous differences between coel- acanths and chondrichthyians and similarities between coelacanths, rhipidistians, dipnoans, and actinopteryians make it unlikely that simi- larities in soft-part morphology between Actin- istia and Chondrichthyes reflect an immediate common ancestor for them shared by no other group of gnathostome fishes. Acceptance of the cladogram implied by these soft-part similarities (Fig. 1) requires the acceptance of either of two grossly unparsimonious hypotheses: 1. The osteichthyian, sarcopterygian and crossoptery- gian characters of coelacanths, if regarded as derived, were separately evolved in these fishes from a primitive, chondrichthyian-like ancestor and are closely convergent in these characters with dipnoans, rhipidistians, and actinopterygi- ans. Coelacanths, at their onset in the Devonian, are no more chondnichthyian-like in these char- acters than their Mesozoic and living descen- dents, and if anything are less divergent from other sarcopterygians. 2. The osteichthyian, sar- copterygian and crossopterygian characters of coelacanths are a result of common ancestry with rhipidistians, dipnoans and actinopterygi- ans, but chondrichthyians are still the immediate sister group of coelacanths. The absence, in this case, of osteichthyian, sarcopterygian and cros- sopterygian characters in chondrichthyians is the result of secondary loss without a trace. However, there is no evidence from living or fossil chondrichthyians that they ever had a ven- tral fissure, intracranial joint, spiracular groove, primitively tropibasic neurocranium, dermal head, palatine, jaw, and shoulder bones of os- teichthyian type, an osteichthyian type of pala- toquadrate or hyomandibular, symplectics, in- terhyals, hypohyals, urohyals, an osteichthyian hyoid operculum, suprapharyngobranchials, dorsal ribs, a swim bladder or lungs, separate narial incurrent and excurrent apertures, lepi- dotrichia, or scales or otoliths of osteichthyian type. Chondrichthyes, from Devonian to Re- cent, show no sign of a shift from an osteich- thyian morphotype, and at most are related to the Osteichthyes at a basal level, as a sister group of acanthodians and osteichthyians (Schaeffer 1975; Schaeffer and Williams 1977). In the light of the great divergence between coel- acanths and chondrichthyians, it seems likely that some of their similarities in soft-part mor- phology may reflect primitive gnathostome (or chondrichthyian + teleostome) traits, and other convergences between these fishes. ACKNOWLEDGMENTS I would especially like to thank Bobb Schaef- fer (Department of Paleontology, American Mu- seum of Natural History) and Warren Freihofer (then of the Department of Ichthyology, Cali- fornia Academy of Sciences) for reviewing an early draft of this paper, and for offering helpful suggestions. I would also like to thank Michael Lagios (Children’s Hospital of San Francisco) for allowing me to read a manuscript copy of his paper on coelacanth relationships, and John 52 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, No. 134 McCosker (Steinhart Aquarium, California Academy of Sciences) for making it possible for me to contribute this account. LITERATURE CITED ANDREWS, S. M. 1973. Interrelationships of crossopterygi- ans. Zool. J. Linn. Soc., London 53(supp. 1):137—177. . 1977. The axial skeleton of the coelacanth, Latime- ria. Linn. Soc. London Symp. (4):271-288. BJERRING, H. C. 1973. Relationships of coelacanths. Zool. J. Linn. Soc., London 53(supp. 1):179-204. GARDINER, B. G. 1973. Interrelationships of teleostomes. Zool. J. Linn. Soc., London 53(supp. 1):105—135. , AND A. W. H. BARTRAM. 1977. The homologies of ventral cranial fissures in osteichthyians. Linn. Soc. Lon- don Symp. (4):227-245. JARVIK, E. 1977. The systematic position of acanthodian fish- es. Linn. Soc. London Symp. (4):199-225. Lgvrrup, S. 1977. The phylogeny of vertebrata. John Wiley & Sons, London, New York. xii + 330 pp. MILEs, R. S. 1968. Jaw articulation and suspension in Acan- thodes and their significance. Pages 109-127 in T. Orvig, ed., Current problems of lower vertebrate phylogeny. Pro- ceedings of 4th Nobel Symposium. Interscience, New York. 1973. Relationships of acanthodians. Zool. J. Linn. Soc., London 53(supp. 1):63-103. 1977. Dipnoan (lungfish) skulls and the relationship of the group: a study based on new specimens from the Devonian of Australia. Zool. J. Linn. Soc., London 61(1— 3):1-328. MILLoT, J.,. AND J. ANTHONY. 1958. Anatomie de Latimeria chalumnae. Tome 1. Squelette, muscles et formations de soutien. C. N. R. S., Paris. 122 pp. Moy-TuHomas, J. A., AND R. S. Mires. 1971. Palaeozoic fishes, 2nd ed. W. B. Saunders Co., Philadelphia, Toronto. x1 + 259 pp. Romer, A. S. 1966. Vertebrate paleontology, 3rd ed. Uni- versity of Chicago Press, Chicago, London. viii + 468 pp. SCHAEFFER, B. 1952. The Triassic coelacanth fish Diplurus, with observations on the evolution of the Coelacanthini. Bull. Am. Mus. Nat. Hist. 99(2):25-78. . 1968. The origin and basic radiation of the Osteich- thyes. Pages 207-222 in T. Orvig, ed., Current problems of lower vertebrate phylogeny. Proceedings of 4th Nobel Sym- posium. Interscience, New York. . 1975. Comments on the origin and basic radiation of the gnathostome fishes with particular reference to the feed- ing mechanism. Collog. Int. C. N. R. S. (218):101-109. , AND M. WILLIAMS. 1977. Relationships of fossil and living elasmobranchs. Am. Zool. 17:293-302. STENSIO, E. A. 1963. Anatomical studies on the arthrodiran head. Part I. K. Sven. Vetensk. Akad. Handl., Ser. 4, 9(2): 1-419. OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES The Biology and Physiology of the Living Coelacanth No. 134, 3 pages December 22, 1979 REPLY TO THE REBUTTAL OF LEONARD COMPAGNO. *“COELACANTHS: SHARK RELATIVES OR BONY FISHES’”’ A significant advantage to a comparative and in particular a systematic analysis based on cal- cified tissues, to wit the skeletal and calcified dermal structures of fishes, is the ability to uti- lize such material from the fossil record. It must be emphasized, however, that this is perhaps the only advantage to such an approach. The other various organ systems, peptide sequences of hormones and other proteins, and biochemical specializations which characterize a living or- ganism, are necessarily excluded from consid- eration by this traditional method of systematic analysis. Moreover, many extant taxa poorly represented in the fossil record can hardly be evaluated by this method. This is not to say that such a traditional method has not been success- fully used, but only that it presents a very lim- ited viewpoint. The systematics developed on the basis of calcified tissues are more useful, moreover, when the members of the respective taxa show substantial divergence so that a pat- tern of baseline or plesiomorphic character traits can be distinguished from subsequent special- izations. When, as in the case of the coelacanth, one is limited to a single surviving species and a rather homogeneous fossil record, it becomes necessary to evaluate evidence other than pa- leontologic remains. By Michael D. Lagios Children’s Hospital, San Francisco, California 94118 53 Compagno notes that the extinct classes Plac- odermi and Acanthodii cannot be evaluated by a comparative approach using organ systems other than calcified tissues. This is true, but it must be recalled that traditional systematics has done little to clarify the relationships of these two groups. Indeed, numerous papers have attempted to place the Acanthodii in the Osteichthyes or Chondrichthyes, depending on the interpretation of the data. Ten of the 17 de- rived skeletal characters which Compagno cites for the coelacanth and the Osteichthyes were present in the Acanthodul, which Jarvik (1977) regards as a Chondrichthian group. Many of the characters which have been uti- lized to distinguish bony fishes from the Chon- drichthyes reflect their divergent specializations over the last 300 million years rather than base- line differences of the groups. The Chondrich- thyes are generally characterized as benthic ‘nose’ ’-fishes with subterminal mouths, a sha- green of denticles, claspers, unrestricted noto- chord with neither centra nor ribs, and serially replaced teeth which are loosely held to the jaws. Yet the archetypal palaeozoic shark Clad- oselache is an obvious non-benthic ‘eye’ ’-pred- ator with reduced rostrum, terminal mouth, composite multicuspid scales, an absence of 54 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, No. 134 claspers, and, as speculated by Keith Thomson (1972), a probable hydrostatic organ. In con- trast, the *‘Teleostome’’ (= Osteichthyes + Acanthodii) Acanthodes shared with Cladosel- ache an unrestrictive notochord with calcified neural arches but no ribs, a similar jaw suspen- sion, and in some forms, composite scales or denticles on a base of acellular bone, and serial whorls of multicusped teeth attached to the jaw cartilages by connective tissue. Among the derived sarcopterygian characters of coelacanths which Compagno cites are the intracranial kinesis which has traditionally been used as the most important synapomorph char- acter linking the coelacanth with the rhipidis- tian fishes. As discussed at length in my paper (this volume), the significance of this feature has been interpreted very differently and recent evi- dence suggests that such kineses including the ventral fissure are widely distributed in archaic gnathostome groups. The superficial similarities of the coelacanth lobate fin with those of the Rhipidistia, and with some imagination early amphibian groups, has also traditionally been used as evidence for an association. As Compagno himself notes, such limb structure independently evolved in several other groups including certain mesozoic sharks. The question of bone in the Chondrichthyes. As noted in my discussion, bone is an archaic plesiomorphic character of vertebrates and its virtual absence in modern sharks no doubt re- lates to secondary adaptation. Similar great re- ductions in mineralized tissues have occurred in true bony fishes, particularly those that are me- sopelagic in habitat, and has frequently been noted in the evolution of several different fish taxa. The skeletal bone as well as the dermal ornamentation of the living coelacanth is itself greatly reduced and much of the bone including that of the dentary (see Miller, this volume) is largely acellular and membranous in type. Should Jarvik’s interpretation of the Acanthodii (1977) be sustained, it would make the entire question of bone in modern sharks moot, as ear- ly acanthodians have both well developed min- eralized skeletal tissue, including endochondral bone and dermal ornamentation. Compagno chooses to consider the ‘‘scales”’ or the branchial arches of bony fishes as distinct from the similarly located ‘‘denticles’’ of the chondrichthian gill arches, although both are epidermal derivatives composed of dentine. The ‘“‘scales’’ of the coelacanth gill arch are entirely comparable to the odontodes which ornament the body scales. Their gross appearance and structure are most ‘“‘chondrichthian-like’’ and are reminiscent of the elaborate composite scales described by Zanger! (1968) in mesozoic sharks. Characteristically, in a cladistic analysis no weight is given to particular traits in analyzing differences between taxa. Although this makes a great deal of sense in terms of the internal logic of cladism, it ignores the empirically-observed degrees of conservation and polymorphism for particular character traits. Forexample, somatic color pattern can be subjected to cladistic anal- ysis but the diversity of this trait is so enormous that no reasonable weight can be given it except on the interspecific level. Pigmentation obvious- ly reflects specific adaptation and as such is a very labile character trait. Although I don’t wish to suggest that calcified tissues represent an or- gan system or similar lability, they are nonethe- less decidedly less stabile as compared to the limited patterns of endocrine organization. Nu- merous examples of convergence in cranial mor- phology and locomotor skeletal structures are known, e.g., the caudal peduncle of thunnid fishes and isurid sharks, the limb structure of elephants and brontosauri, etc. Such conver- gence obviously reflects adaptation to specific environmental needs. In contrast endocrine or- ganization represents a ‘“‘nonsense’’ feature in terms of function. Insulins produced by the dis- crete and separate islet tissue, the Brockmann’s bodies, of teleost fishes and those of the diffuse small intrapancreatic islets of mammals are not only functionally equivalent in their respective taxa, but, due to marked similarities in amino acid sequence, show considerable biologic ac- tivity in heterologous taxa as well. The gonad- otropins of sharks, synthesized by cells in a dis- crete, isolated ventral lobe are entirely equivalent functionally to those of other vertebrate groups. Although comparative work on the gonadotro- pins of the coelacanth and other vertebrates has not been performed it is clear that the growth hormone of the coelacanth shows numerous an- tigenic similarities with those of tetrapods and preserves significant biologic activity in the lat- ter (Hayashida, this volume). In summary, the coelacanth has historically been interpreted as a crossopterygian predom- inantly on the basis of a number of superficial LAGIOS: REPLY TO REBUTTAL OF COMPAGNO skeletal similarities on which Compagno has based his rebuttal. These similarities include many, such as lobed fins, scales, and cranial ki- nesis, which have been independently evolved by various taxa and could represent examples of convergence in the coelacanth. Analysis of organ systems other than skeletal tissues, such as endocrine and visceral anatomy, which em- pirically show a greater degree of morphologic conservatism among extant taxa, suggest that the sister group of the coelacanth is to be found among the Chondrichthyes. Modern Chondrich- thyes, with their great reduction in calcified tis- sues, simplified scales and serially replaced teeth, hardly seem a likely sister taxon to “‘Old Four Legs,’ yet most of the *“‘diagnostic”’ char- ss acter traits of the Chondnchthyes represent spe- cializations for an active, predaceous habitus. The conclusions which have been reached on the basis of endocrine organization, soft part morphology and biochemistry in the coelacanth are admittedly at odds with the prevailing con- ception of its systematic position. Though soft part morphology cannot be compared among extinct taxa, it nonetheless carries significant weight in any systematic analysis. In the opinion of this observer more so than the more labile calcified tissues which have conventionally been used to classify the coelacanth as a sister group to the Rhipidistia and as a member of the Os- teichthyes. OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES The Biology and Physiology of the Living Coelacanth No. 134, 12 pages December 22, 1979 VENTRAL GILL ARCH MUSCLES AND THE PHYLOGENETIC RELATIONSHIPS OF LATIMERIA By E. O. Wiley Museum of Natural History, The University of Kansas, Lawrence, Kansas, 66045 ABSTRACT The phylogenetic relationships of Latimeria chalumnae as evidence by ventral gill arch musculature is reviewed. I conclude that Latimeria and, by inference, other ac- tinistians (coelacanths) are the sister group of all other osteichthyans. Osteological evidence which supports this hypothesis is reviewed. Two alternate views are examined: (1) that actinistians are sarcopterygians and (2) that actinistians are relatives of Chon- drichthyans. Alternate (1) is reasonable if the intracranial joint is considered synapo- morphic (derived) for actinistians and rhipidistians. Alternate (2) is supported only by characters which can be shown to be plesiomorphic (primitive) or of uncertain quality. I conclude that (a) Latimeria is an osteichthyan and (b) that Latimeria and other actinistians are the sister group of all other osteichthyans (the Euosteichthyes). INTRODUCTION The phylogenetic relationship of Latimeria chalumnae has been debated as much by past investigators as it is debated in this book. Wes- toll (1949) placed coelacanths with the sarcop- terygian groups Dipnoi (lungfishes) and choa- nata (Rhipidistia and Tetrapoda). He considered the intracranial joint a primitive character of Sarcopterygii. Romer (1966) and Thomson 56 (1969) considered Latimeria the closest living relative to the tetrapods and closely related to the extinct rhipidistians. The alignment of coel- acanths with rhipidistians and tetrapods has most recently been advocated by Andrews (1973) and Miles (1975). Other investigators have expressed doubts. Although Nelson (1969) grouped Latimeria in Sarcopterygil, he suggest- ed that Latimeria might be derived from a non- WILEY: GILL ARCH MUSCLES sarcopterygian teleostome.! Von Wahlert (1968) suggested that coelacanths were the sister-group of choanates + actinopterygians and that Dip- noi was the sister-group of the three (Fig. 1a). Bjerring (1973) questioned the homology of the intracranial joints of rhipidistians and coel- acanths and suggested that the relationships of coelacanths were unknown. Andrews (1977) came to a similar conclusion, pointing out that the vertebrae of Latimeria invite comparison with all the major osteichthyan groups. Miles (1977) concluded that coelacanths were sarcop- terygians, but presented evidence that Dipnoi was more closely related to choanates than are coelacanths (Fig. 1b). Lg@vtrup (1977) came to an entirely different conclusion—that coel- acanths are the sister-group of chondrichthyans (Fig. lc). A similar viewpoint is expressed in this volume by Lagios. My own work with the ventral gill arch mus- cles of gnathostomes (Wiley 1979; specimens examined are listed in that paper) led me to con- clusions that differ in one respect or another from all the above authors. I concluded that coelacanths are the sister-group of all other Re- cent osteichthyans (Fig. 1d). Below I will review the evidence supporting this hypothesis and at- tempt to resolve the apparent character conflicts between this hypothesis and competing hypoth- eses. Phylogenetic Analysis Phylogenetic (or ‘‘cladistic’’) analysis is a method to assess the relative genealogical rela- tionships between organisms. It is designed to ask if two groups of organisms shared a common ancestral species not shared by another organ- ism. If these two groups show similarities in morphology or other attributes not shown by other groups, then the method assumes it 1s more parsimonious to consider these similarities ' A word about nomenclature is in order. Teleostomi in- cludes, in this paper, acanthodians, coelacanths, dipnoans, rhipidistians, tetrapods and actinopterygians: in short, all non-chondrichthyan and non-placoderm gnathostomes. Osteichthyans are all teleostome groups except acanthodians. Sarcopterygii, in the common usage, includes choanates (rhip- idistians and tetrapods), Dipnoi and coelacanths. But, coela- canths are excluded from the Sarcopterygii in this paper for reasons explained below. Thus, ‘‘Sarcopterygii’’ used without qualification refers only to Choanata and Dipnoi. The term ‘‘Euosteichthyes” refers to Sarcopterygii (as defined herein) plus Actinopterygii. These terms follow Wiley (1979). LY iS SS 9 ~ => S is ~ ~ OS OS SS OS an ee Sy Q) 1S) 1S) NE (Se {S) 9 1S) Na fe) b. SS S i~ x We Ss S < ~ OQ) (@) Oo To oS oT VD LY & > GG oO oS = GS soos ND C. de FiGure 1. Four hypotheses of the interrelationships of chondrichthyans (Chon), dipnoans (Dipn), actinistians (coel- acanths, Coel), choanates (Choa) and actinopterygians (Acti): (a) from von Wahlert (1968); (b) from Miles (1978); (c) from Levtrup (1977); (d) from Wiley (1979). as having developed in the common ancestor of the two groups than it is to assume that these similarities are the result of convergence or par- allelism. Characters hypothesized to demon- strate a unique common ancestry relationship are synapomorphous or derived characters. These may be defined as homologous characters developed in the ancestral species of the taxa showing them and in no other ancestral species. A number of alternate hypotheses based on dif- ferent characters may result from such an anal- ysis. In that case the phylogeneticist attempts to resolve the conflict by demonstrating (1) that supposed synapomorphies are actually primitive or plesiomorphous characters or (2) that sup- posed synapomoprhies are actually non-homol- ogies. Failing this, the hypothesis with the great- est number of hypothesized synapomorphies is considered the best estimate of phylogenetic de- scent. For those interested in the finer aspects of this method and justification for using the method rather than alternates I can recommend Hennig (1966) and various articles in the journal Systematic Zoology. My thoughts on the meth- 58 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, No. 134 FIGURE 2. Diagrammatic ventral views of the gill arch hypobranchial muscles of (a) a generalized shark; (b) a generalized dipnoan; (c) a generalized actinopterygian. Bone and cartilage stippled, basibranchial copulae not shown. Abbreviations: Cl, ceratobranchial 1; CA, coracoarcualis; CB1, 2, coracobranchiales | and 2; CM, coracomanbibularis; HA, hyoid arch; LJ, lower jaw; PG, pectoral girdle; STH, sternohyoideus. od are outlined in Wiley (1975, 1976) and En- gelmann and Wiley (1977). Ventral Gill Arch Muscles Branchial and hypobranchial ventral gill arch muscles serve to abduct and adduct the visceral arches, to open and aid in opening the lower jaw and to aid in moving the pectoral girdle. Hypo- branchial muscles are derived from the ventral portions of the more anterior post-otic body myotomes and are innervated by one or more spino-occipital (post-vagus) nerves. These mus- cles are shown in Table 1. Branchial muscles are derived from lateral plate mesoderm. Those of the first arch are innervated by a branch of the glossopharyngeal (IX) and/or a branch of the vagus (a branch of X,). Branchial muscles of more posterior arches are innervated by the TABLE |. Wiley (1979). post-trematic branch of the vagus for their re- spective arch (i.e., the X,—X, for those species with five arches). These muscles are shown in Table 2. Note in the discussion that some of the muscles shown in these tables are homologues of others and that none are found in all taxa. Hypobranchial Muscle Patterns Sternohyoideus Muscles.—Latimeria cha- lumnae has a pair of sternohyoidei. The sterno- hyoideus of each side originates on the clavicle medial and dorsal to the coracomandibularis. It inserts on the ‘“‘urohyal’’ (not homologous with the teleost urohyal) just anterior to the insertion of the transversus ventralis 2. All gnathostomes have sternohyoidei. These muscles in Latimeria are unique in their inser- tion. The usual insertion for primitive members HyPOBRANCHIAL MUSCLES ASSOCIATED WITH THE VENTRAL GILL ARCHES OF GNATHOSTOMES. Names follow Muscle Origin Insertion Coracoarcualis Pectoral girdle Coracomanbibularis Pectoral, girdle, and/or Sternohyoides Lower jaw sternohyoides, or 3rd hypobranchial Sternohyoideus Pectoral girdle Coracobranchiales Pectoral girdle (last CB only) Hyoid arch or urohyal Ceratobranchial of respective arch and medial hypobranchial muscles WILEY: GILL ARCH MUSCLES 59 TABLE 2. BRANCHIAL MUSCLES ASSOCIATED WITH THE VENTRAL GILL ARCHES OF GNATHOSTOMES. Names follow Wiley (1979). Muscle Origin Interarcualis | Interarcualis 2 Interarcualis 3 Interarcualis 4 Interarcualis 5 Pharyngoclavicularis Ph. externus Ph. internus Obliquus ventralis | Obliquus ventralis 2 Obliquus ventralis 3 Obliquus ventralis 4 Rectus communis Rectus ventralis Subarcualis rectus Hyoid arch Ceratobranchial 1 Ceratobranchial 2 Ceratobranchial 3 Ceratobranchial 4 Pectoral girdle Hyoid arch Hypobranchial 2 Hypobranchial 3 Hypobranchial 4 Variable Ceratobranchial 5 Ceratobranchial 5 Transversus ventralis 2 Midline Transversus ventralis 3 Midline Transversus ventralis 4 Midline Transversus ventralis Midline posterior Ventral superficial constrictors fascia of arch Insertion Innervation Ceratobranchial 1 DXGEIENG Ceratobranchial 2 x Ceratobranchial 3 X, Ceratobranchial 4 > Ceratobranchial 5 xX 4 Ceratobranchial 5 X, and/or recurrens Ceratobranchial 5 xe Ceratobranchial 1* IX + X, Ceratobranchial 2 Ceratobranchial 3 Ceratobranchial 4 Variable Ceratobranchial 3 Ceratobranchial 2,3 Ceratobranchial 2 Ceratobranchial 3 Ceratobranchial 4 Ceratobranchial 5 4 XX MK KKM RH = respective arch DXGtopxge= and/or midline * Hypobranchial | in actinopterygians. ** For species with five arches. of other gnathostome groups is on the lower part of the hyoid arch (Figs. 2a, b and c). The ster- nohyoidei provide no insight into the phyloge- netic relationships of Latimeria because, where they differ from other gnathostomes (in inser- tion), they are unique. Coracomandibularis Muscles.—The cora- comandibulares of Latimeria originate on the lower lateral side of the clavicles and insert on the lower jaws dorsal to the intermandibularis musculature and ventral to the sternohyoidei. Primitive members of all gnathostome groups have corcomandibulares. The condition in Lati- meria is similar to that of many chondrichthyans (Fig. 2a) and all dipnoans (Fig. 2b). Actinop- terygians (including polypterids) differ in having the origins of these muscles on the third hypo- branchial (Fig. 2c, thus the term branchioman- dibularis). In teleosts and gars the muscle is modified or absent. This is also true for am- niotes. Since both the primitive sarcopterygian and chondrichthyan conditions are similar to each other and to Latimeria, this muscle also provides no insight into the phylogenetic rela- tionship of Latimeria. Coracobranchialis Muscles.—Coracobran- chiales are found only in chondrichthyans (Fig. 2a). Their absence in osteichthyans and in ag- nathans supports the idea that the presence of these muscles in chondrichthyans is a synapo- morphy for that group. The lack of coracobran- chiales in Latimeria, therefore, neither supports nor refutes any hypothesis concerning its rela- tionships. Coracoarcualis Muscles.—Coracoarcuales are well developed in chondrichthyans (Fig. 2a) and are probably found in osteichthyans as part of the sternohyoidei or coracomandibulares (Wiley 1979). There are indications of coracoarcuales in Neoceratodus but none in Latimeria (at least according to Millot and Anthony 1958). At best, the presence or absence of coracoarcuales are ambiguous characters because the “‘loss’’ could support a monophyletic Osteichthyes or the ‘gain’? could support a monophyletic Chon- drichthyes. Summary of Hypobranchial Muscle Pat- terns.—Hypobranchial muscle patterns are use- ful in supporting hypotheses of the monophyly of actinopterygians (coracomandibularis origin on the third hypobranchial) and chondrichthy- ans (presence of coracobranchiales). The ab- 60 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, No. 134 FIGURE 3. Diagrammatic ventral views of the gill arch branchial muscles of (a) a generalized shark; (b) a hypothetical primitive dipnoan; (c) Latimeria chalumnae; and (d) a primitive actinopterygian. Bone and cartilage stippled, basibranchial copulae not shown. Abbreviations: HA, hyoid arch; IV1, 2, interarcuales ventrales | and 2; OVI, 4: obliqui ventrales 1 and 4; PC, pharyngoclavicularis; SC1, 3, 5, superficial constrictors 1, 3, and 5; SR, subarcualis rectus; TV2, 3, 4, transversi ventrales 2, 3, and 4; TVP, transversus ventralis posterior. sence of coracoarcuales in Latimeria may sup- port a monophyletic Osteichthyes, but since agnathans do not have comparable hypobran- chial muscles, it is not possible to evaluate whether the absence or reduction of coracoar- cuales in osteichthyans is primitive or derived. I conclude that hypobranchial muscles are not useful in evaluating the phylogenetic relation- ships of coelacanths. Branchial Muscle Patterns General Observations.—There are two basic sources for branchial muscles. Both are onto- genetic derivatives of lateral plate mesoderm. The first of these sources is the muscle sheet associated with each gill arch. The second is the oesophageal musculature. The muscle fibers as- sociated with each gill arch give rise to various muscles in various gnathostome groups. They include the superficial constrictors, transversi ventrales and their ontogenetic derivatives, and the interarcuales ventrales. The oesophageal musculature gives rise to the posterior transver- sus and the pharyngoclavicularus and its deriv- atives. The following discussion is based on Wiley (1979) and references therein. Superficial Constrictor Muscles.—Latimeria chalumnae has four pairs of superficial constric- tors, one for each of the first four gill arches (Fig. 3b). The presence of superficial constric- tors on the first four gill arches is characteristic of dipnoans (Fig. 3c), amphibians (larvae), and WILEY: GILL ARCH MUSCLES actinopterygians (where the superficial constric- tors are present as the interbranchial muscula- ture). Chondrichthyans differ in having superfi- cial constrictors on all arches (i.e., on the fifth, sixth, or seventh depending on the number of arches, Fig. 3a). On ontogenetic and phyloge- netic grounds it is reasonable to postulate that the lack of superficial constrictors on the last gill arch is a secondary loss character. Therefore, this character is derived for Osteichthyes and evidence that Latimeria is more closely related to other osteichthyans than to chondrichthyans. Transversus Ventralis and Obliquus Ventra- lis Muscles.—Latimeria chalumnae has two anterior transversus muscles, one associated with the second gill arch (transversus ventralis 2) and one associated with the third gill arch (transversus ventralis 3). Latimeria also has a pair of posterior transversus muscles associated with the fifth gill arch (transversus ventralis pos- terior). These muscles are illustrated in Fig. 3b. Dipnoans, tetrapods (amphibian larvae) and actinopterygians have the same two anterior transversus muscles as Latimeria (Figs. 3c, d). (Actinopterygians have obliqui ventrales, onto- genetic derivatives of the transversi ventrales.) In contrast, chondrichthyans lack anterior trans- versus muscles. In chondrichthyans the super- ficial constrictors either meet in an undifferen- tiated condition at the midline (Fig. 3a) or (in some sharks and holocephalians) end with the holobranchs. This represents the more primitive phylogenetic condition. Thus, I interpreted the presence of transversi ventrales 2 and 3 as shared derived characters linking Latimeria with Osteichthyes (Wiley 1979). I note that dip- noans, amphibian larvae and actinopterygians have an obliquus ventralis 1 and a transversus ventralis 4 (Figs. 3c, d). Neither muscle is found in chondrichthyans or Latimeria (Figs. 3a, b). I conclude that this evidence supports a mono- phyletic Euosteichthyes and excludes Latimeria from being a sarcopterygian (Wiley 1979). Fi- nally, a transversus ventralis posterior is char- acteristic of all gnathostomes and therefore use- less in assessing phylogenetic relationships. Pharyngoclavicularis Muscles.—Latimeria chalumnae lacks paired pharnygoclaviculares. This similarity is shared with agnathans and chondrichthyans and is primitive (Wiley 1979; Figs. 3a, b). In contrast, euosteichthyans have pharyngoclaviculares. In sarcopterygians, there are multiple pharyngoclaviculares on each side 61 (Fig. 3c) while in actinopterygians there is a sin- gle pharyngoclavicularis on each side (Fig. 3d). The presence of a pharyngoclavicularis muscle system can be interpreted as a derived character and evidence of a monophyletic Euosteichthyes (Wiley 1979). Interarcualis Ventralis Muscles.—Latimeria chalumnae has interarcuales ventrales connect- ing the hyoid arch with the first gill arch and each succeeding gill arch (Fig. 3b). This similar- ity is shared with dipnoans (Fig. 3c) and, in a reduced condition (i.e., some are missing), with tetrapods (amphibian larvae). These muscles are missing in chondrichthyans and in actinopteryg- lans (Figs. 3a, d). Without considering other evi- dence, the presence of interarcuales ventrales is a unique similarity shared among dipnoans, tet- rapods and Latimeria. Thus, it could be inter- preted as a synapomorphy linking Latimeria with Sarcopterygii. However, when other mus- cles are considered it is more parsimonious to consider the presence of interarcuales ventrales a derived character of Osteichthyes and thus primitive within Osteichthyes. This calls for the ad hoc hypothesis that the lack of interarcuales ventrales in actinopterygians is a loss character derived for that group rather than homologous with the lack of interarcuales ventrales in chon- drichthyans. The lack of these muscles in chon- drichthyans is interpreted as a primitive gna- thostome character. A SUMMARY OF THE EVIDENCE The phylogenetic hypothesis presented in Fig. 4 summarizes the ventral gill arch muscle evi- dence for the phylogenetic relationships of ma- jor gnathostome groups. It also summarizes evi- dence presented below and by Wiley (1979) concerning the interrelationships of certain groups within Actinopterygii and Sarcopterygil. Gnathostomes differ from agnathans in having ventral branchial gill arch musculature. Chon- drichthyans share the presence of coracobran- chiales. Osteichthyans share three derived char- acters: 1) the presence of discrete transversi or obliqui ventrales on the second and third gill arches, 2) the presence of interarcuales ven- trales (secondarily lost in Actinopterygii), and 3) the loss of superficial constrictors on the fifth gill arch. Additionally, the loss or reduction of the coracoarcuales might be derived condition. Euosteichthyans share three derived characters 62 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, No. 134 Ficure 4. An hypothesis of vertebrate interrelationships. Synapomorphies (black bars) uniting taxa are: (1), general synapomorphies of gnathostomes; (2), ventral branchial mus- cles present in the gill arches; (3), dermal bones associated with the endochondral palate; (4), parasphenoid present; (5), premaxilla present; (6), angular, prearticular and coronoid bones present in the lower jaw: (7), cleithrum and clavicle present in the pectoral girdle; (8), dermal operculum and car- tilaginous suboperculum present; (9), accessory hyoid element present; (10), interarcuales ventrales 1—5 present; (11), trans- versi ventrales 2 and 3 present; (12), fifth superficial constric- tor absent; (13), the lateral commissure in anterior position: (14), ventral cranial fissure present; (15), dermal subopercu- lum present; (16), maxilla present; (17), obliquus ventralis 1 present; (18), pharyngoclavicularis present; (19), transversus ventralis 4 present; (20), origin of the coracomandibularis on the third hypobranchial; (21), obliqui ventrales (rather than transversi ventrali) present; (22), interarcuales ventrales ab- sent; (23), interbranchiales (rather than superficial constric- tors) present; (24), cosmine present; (25), presence of a su- praotic cavity associated with the endolymphatic system; (26), presence of submandibular bones. Synapomorphies are drawn from several sources, see text. not found in Latimeria: 1) the presence of trans- versi or obliqui ventrales on the fourth gill arch, 2) the presence of obliqui ventrali connecting the hyoid arch and the first gill arch, and 3) the pres- ence of pharyngoclaviculari. Therefore, Lati- meria and other actinistians are osteichthyans and the sister-group of all other osteichthyans. Of these characters, the acceptance of the pres- ence of interarcuales ventrali as a synapomor- phy of osteichthyans is ad hoc, a by-product of considering the phylogenetic relationships of Latimeria as evidenced by other characters. Other Corroborating Evidence Additional evidence relating to the problem of actinistian relationships falls, naturally, into those lines which tend to corroborate and those lines which tend to refute the scheme of rela- tionships presented above. Below, I will briefly summarize some of the morphological evidence which corroborates actinistians as osteichthyans and as the sister-group of Euosteichthyans. 1. The Neurocranium.—Latimeria chalum- nae and other actinistians are similar in general ossification pattern to other teleostomes but dif- fer from chondrichthyans and placoderms in a number of neurocranial characters. These in- clude: a. The presence of a ventral cranial fissure (the ventral part of the intracranial joint; Gar- diner and Bartram 1977). b. The lateral commissure is located in the anterior part of the otic region, carries the hyo- mandibular articulation and forms part of the tri- gemino-facial chamber (Jollie 1971; Miles 1973). This is distinctly similar to other osteichthyans and distinctly different than chondrichthyans (Jollie 1971). 2. Dermal Bones in the Braincase.—Os- teichthyans among all vertebrates have a para- sphenoid (Miles 1973) and dermal bones asso- ciated with the endochondral palate (Schaeffer 1968; Gardiner 1973). 3. Hyoid and Visceral Arches.—Schaeffer (1968) and Jollie (1971) have called attention to the fact that only teleostomes among all verte- brates have accessory elements between the hyomandibular and ceratohyal (i.e., the inter- hyal and/or sympletic, also see Miles 1973:72 for discussion). Actinistians differ from acanthodi- ans but resemble euosteichthyans in having both supra- and infrapharyngobranchials (cf. Gardi- WILEY: GILL ARCH MUSCLES ner 1973) and in having anteriorly directed in- frapharyngobranchials. All three conditions are derived relative to non-osteichthyan gnathos- tomes. 4. Lower Jaw.—The lower jaw of Latimeria is composed of several elements. Ossifications of Meckel’s cartilage include the mentomecke- lian and the articular. Both are primitive based on their presence in placoderms (Moy-Thomas and Miles 1971). Dermal elements include a den- tary, infradentary (or splenial), angular, prear- ticular and a coronoid series (Nelson 1973). Of these the dentary and angular may be primitive based on the presence of similar bones in plac- oderms (i.e., the intergnathal may represent both, see illustration and discussion of Moy- Thomas and Miles 1971). The remaining bones are, to my knowledge, found only in Osteich- thyes and represent unique similarities. The or- ganization of the tooth-bearing bones of actinis- tians is comparable only to primitive dipnoans (Schultze 1969) and Nelson (1973) has suggested that the organization of the actinistian lower jaw is uniquely primitive. This is compatible with the interpretation that actinistians are the sister- group of all other osteichthyans. 5. Shoulder Girdle.—Latimeria, like other osteichthyans, has a well-developed dermal shoulder girdle which includes a clavicle, clei- thrum and one dorsal element. Neither acantho- dians nor placoderms have comparable ossifi- cations. (Placoderms do have trunk shield armour and acanthodians have dermal plates but apparently neither is homologous to the dermal elements of osteichthyans: Moy-Thomas and Miles 1971; Miles and Young 1977.) 6. Upper Jaw.—Latimeria has a premaxilla (Millot and Anthony 1965; von Wahlert 1968; Miles 1977) but lacks any element which can be called a maxilla. Attempts to identify a vestigial maxilla (i.e. Nielsen 1936; as discussed by Miles 1977:188) are simply not convincing and I con- clude that there is no good evidence that acti- nistians ever had a maxilla. In contrast, dipnoans have a series of tooth-plates which might be homologized with the maxilla of other euos- teichthyans (cf. discussion by Miles 1977:186— 188). I interpret the presence of a premaxilla as a shared derived character of Osteichthyes based on its absence in acanthodians, placo- derms and chondrichthyans. I interpret the pres- ence of a maxilla as a shared character of Euos- teichthyes based on its absence in actinistians 63 and all other non-euosteichthyan vertebrate groups. 7. The Opercular Series.—Patterson (1977) has discussed the homology and development of various skull bones which forms the basis for this section. There are, among gnathostomes, two series of opercular elements—one series is cartilaginous and one is dermal. The dermal os- sifications are neither the ontogenetic nor phy- logenetic derivatives of the cartilaginous ele- ments. Holocephalians have a large cartilaginous operculum, sharks have numerous small hyoid ray cartilages. Latimeria chalumnae has both a cartilaginous and a dermal operculum, and it has (at least according to Millot and Anthony 1965) a separate cartilaginous suboperculum. The presence of a dermal operculum and a cartilagi- nous suboperculum are unique osteichthyan fea- tures while the presence of dermal bone in the opercular flap seems a unique teleostome feature (acanthodians have many smaller dermal ele- ments, Schaeffer 1968; Miles 1973: similar struc- tures are missing in placoderms). Further, the presence of a dermal operculum and cartilag- inous suboperculum can be interpreted as shared derived characters of osteichthyans. Within Osteichthyes, Latimeria and other acti- nistians are primitive in lacking a dermal sub- operculum. Thus, a dermal suboperculum can be interpreted as a shared derived character of Euosteichthyes. 8. Dermal Skull Roofing Bones .—Graham- Smith (1978) recently analyzed various skull roofing bone patterns in placoderms and os- teichthyans, excluding actinistians. His analyses indicate that many patterns of fusion and frag- mentation exist. While it is apparent that acti- nistians differ in several respects from other os- teichthyans, it is not clear to me what the significance of this variation means in terms of phylogenetic relationships. Refuting Evidence Refutation of the hypothesis that actinistians are osteichthyans and the sister group of euos- teichthyans falls into two basic classes. First, it has been argued that actinistians are either the sister group of choanates (Romer 1966; Andrews 1973; Miles 1975) or at least the sister group of dipnoans + choanates (Miles 1977). Second, it has been argued that actinistians are (among Recent groups) the sister group of chondrich- thyans (Lgvtrup 1977; Lagios, this volume). 64 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, No. 134 Actinistians as the Sister Group of Choanates This may be termed the ‘‘traditional’’ hypoth- esis and is based on such features as the orga- nization of the archiopterygial fin, the presence of an intracranial joint in actinistians, rhipidis- tians and ichthyostegid amphibians, and the gen- eral similarity of actinistians and rhipidistians. However, Miles (1977) has presented evidence that dipnoans are more closely related to choa- nates than are actinistians. His hypothesis is based on three characters shared by dipnoans and choanates that are not shared by actinis- tians: (1) the presence of cosmine, (2) the pres- ence of a supraotic cavity associated with the endolymphatic system in the cranium, and (3) the presence of a series of submandibular bones. Of these three characters, (1) and (2) are unique to dipnoans and choanates while (3) is shared with actinopterygians (Jessen 1968; Miles 1977 is of the opinion that the series in actinoptery- gians is not homologous with that of choanates and dipnoans).” Miles’ (1977) evidence supports the conclusion that dipnoans are the closest Re- cent relatives of tetrapods and the closest rela- tives of choanates among all vertebrates. Latimeria as a Sarcopterygian The major line of evidence supporting this hy- pothesis is the presence of an intracranial joint in actinistians and rhipidistians (and perhaps in some primitive tetrapods such as the ichthyo- stegids). This joint separates the brain case into two regions and is impressively similar in the two groups. However, there are differences. In the cranium, these differences involve the to- pographic relationship of the joint and the tri- germinal nerve (Schaeffer 1968; Bjerring 1973). In rhipidistians, the maxillary and mandibular branches of the trigeminal exit the posterior di- vision of the braincase via a foramen which is located posterior to the intracranial joint. In ac- tinistians, these branches of the trigeminal exit through the intracranial joint itself. There are also differences between the topographic posi- tion of the joint and the dermal skull roof (Bjer- ring 1973). In rhipidistians, the joint’s external position is between the parietals and frontals. In * If, however, Actinistia is the sister group ofall other os- teichthyans, there is no phylogenetic reason to consider the submandibular series as anything but a synapomorphy of Euosteichthyes and its absence in Latimeria as a primitive character actinistians, the external position of the joint is between two bones of uncertain homology (var- iously termed frontals and parietals or an ante- rior portion of the frontal and a posterior portion of the frontal fused with the parietal; cf. An- drews 1973, and Bjerring 1973, respectively). In actinistians the joint is anterior to the junction of the infraorbital and supraorbital sensory ca- nals while in rhipidistians the joint is posterior to this junction. Bjerring (1973) argued, on the basis of an ar- cual theory of cranial ossification, that the intra- cranial joint of actinistians is non-homologous with that of rhipidistians. Arcual theories of cra- nial ossification stem from Goethe (see DeBeer 1937) and state that various cranial ossifications correspond to metameres in the cranium that are serial equivalents of the vertebral metameres. Bjerring (1973) identified the topographic posi- tion of the actinistian joint as lying between the presumptive mandibular (2nd) and hyoid (3rd) metameres and the topographic position of the rhipidistian joint as lying within the mandibular metamere between the antotic pilae and the acorochordals. However, there is a problem with this hypothesis. Gardiner and Bartram (1977) demonstrated that the ventral part of the intracranial joint is the homologue of the ventral otic fissure. The ventral fissure is a plesiomor- phic osteichthyan feature present in acanthodians actinopterygians and dipnoans as well as being represented as part of the intracranial joint. This means that the arcual hypothesis must be refut- ed since it is tied to the hypothesis that the entire joint (dorsal and ventral portions) is non-ho- mologous in actinistians and rhipidistians. But, this does not mean that the entire joint must be thought of as homologous in the two groups. Wiley (1979) suggested that the differences ex- hibited in the dorsal part of the joints of actinis- tians and rhipidistians might mean that the dor- sal part of the joints were non-homologous. Comments.—We are faced with three mu- tually exclusive hypotheses. (1) Actinistians are the sister group of Euosteichthyes. (2) Actinis- tians are the sister group of choanates and dip- noans are the sister group of the two. (3) Actinis- tians are the sister group of dipnoans and choanates. If we opt for hypothesis (1), then we must reject the intracranial joint of actinistians as being homologous with that of rhipidistians and accept that it has been gained twice. If we opt for hypothesis (2), then the lack of the var- WILEY: GILL ARCH MUSCLES ious Ossifications and muscles in Latimeria out- lined above must be thought of as being lost characters which are synapomorphous for the Actinistia. Such an hypothesis might be reason- able if it could be shown that the variation in trigeminal nerve exits in rhipidistians (Thom- son, 1967) made the differences between actinis- tian and rhipidistian joints trivial and if enough weight was placed on the character to preclude the use of the characters embodied in (1) as syn- apomorphies. Hypothesis (3) seems the most uneconomical of the three. It assumes that the intracranial joint of sarcopterygians was lost twice and it assumes that all of the characters uniquely exhibited by euosteichthyans have been independently derived in actinopterygians on the one hand and dipnoans + choanates on the other hand. All of this discussion, of course, rests on the assumption that the rhipidistian fishes are the sister group of tetrapods. If it could be shown that rhipidistians are more primitive than dip- noans, then other interpretations are possible, not all of which are inconsistent with the inter- pretation that Latimeria is a sarcopterygian. Latimeria as the Sister Group of Chondrich- thyes The Pituitary Complex.—Lagios (1975, and this volume) has compared the organization of the pituitary gland of Latimeria with other gnathostomes. Like chondrichthyans, the pars distalis of Latimeria is organized into two dis- tinct components, a ventral component (the ros- tral lobe) and a dorsal component. The dorsal component incorporates a dorsal lobe composed primarily of acidiophilic cells and a proximal lobe composed of mixed cell types. Lagios (1975) suggests a homology between the rostral lobe of Latimeria and what is usually termed the ventral lobe of chondrichthyans. He rejects a homology between the ventral lobe of Polyp- terus and certain teleosts and the ventral lobe of chondrichthyans and Latimeria because the actinopterygian lobe contains cells which se- crete ‘“‘prolactin’’ whereas the ventral lobes of Latimeria and chondrichthyans lack these cells. Comments.—I see no reason to reject the hy- pothesis that the ventral lobe of the pars distalis of chondrichthyans is homologous with the sim- ilar structure found in Latimeria. However, I think there is good reason to reject the notion that this homology is a synapomorphy indicating 65 a sister group relationship between these groups. Olsson (1968) has pointed out that the ventral lobe of the pars distalis is an embryonic remnant of Rathke’s pouch. Those gnatho- stomes with ‘“‘compact”’ or ‘‘follicular’’ pituitar- ies (cf. Olsson 1968) go through an embryonic stage in which the adenohypophysis is formed by an invagination of the stomodeum (the ectoder- mal lining of the anterior pharyngeal cavity). The adenohypophysis primordium pushes dorsal- ly until it contacts the neurohypophysis. It then loses its connection with the stomodeum and may become ‘‘duct type,” ‘‘follicle type’ or ‘““compact type’? in its organization depending on subsequent ontogenetic change. Thus, Lati- meria, chondrichthyans, and polypterids (all of which have the ‘‘duct type’’) can be interpreted as having and ‘“‘incomplete’’ ontogeny of the pars distalis that is the primitive expression of a phylogenetically derived condition found in all euosteichthyans except (to my knowledge) po- lypterids. Following Olsson (1968), I conclude that the bipartite organization of the pars dis- talis of Latimeria, chondrichthyes and polyp- terids is a primitive gnathostome feature and thus not evidence for a sister group relationship between Latimeria and chondrichthyans. In- deed, if we follow Wingstrand (1966), this bi- partite organization is a primitive vertebrate feature and the compact pituitaries of lampreys and hagfishes cannot be considered homologous with the compact pituitaries of various euos- teichthyan groups. Finally, the presence of ‘prolactin’? cells in the ventral lobes of acti- nopterygians does not denote the non-homolo- gous nature of this lobe to that of actinistians so much as it suggests that these cells are derived relative to the lack of ‘“‘prolactin’’ cells. The Rectal Gland.—The presence of a rectal gland in Latimeria and elasmobranchs has been cited as evidence for a sister group relationship between chondrichthyans and actinistians by Lgvtrup (1977) and by Lagios (this volume). The similarities are striking, but there are problems with interpreting these similarities as synapo- morphies. Holocephalians do not have rectal glands. As pointed out by Lagios (this volume), holocephalians do have cation excreting tissue in the hindgut which may be reasonably consid- ered as a phylogenetically more primitive expression of the rectal gland of elasmobranchs. The problem with interpreting the rectal glands of Latimeria and elasmobranchs as synapomor- 66 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, No. 134 phies is that we must accept the hypothesis that Latimeria is more closely related to elasmo- branchs than are holocephalians. In other words, we must assume that Latimeria fits be- tween holocephalians and sharks and not below holocephalians and sharks in the phylogeny. Since there is reasonable evidence at hand to reject this hypothesis in favor of an alternate hypothesis that among these three taxa, holo- cephalians are more closely related to elasmo- branchs than is Latimeria we are left with three hypotheses. (1) The rectal gland is a primitive gnathostome feature and thus no evidence for relationships within the Gnathostomata. (2) The rectal glands are non-homologous, and thus no evidence of relationship at any level. (3) The rectal glands are synapomorphic for Actinistia + Chondrichthyes and have been independently reduced in holocephalians. The third hypothesis is uneconomical for the same logical reasons that I have given for Miles’ (1977) hypothesis concerning the intracranial joint. Other Evidence.—Without going into great detail or out-group comparisons, L@vtrup (1977) listed the following characters as similarities shared by Latimeria and chondrichthyans: a fat- ty liver, eye structure, shape of the gray matter in the spinal chord, the absence of Manthner’s fibers, the position of the first pair of cranial nerves in a hollow stalk which is a continuation of the telecephalic cavity, the structure of both the thymus and thyroid, large eggs, and osmo- regulation through urea retention. Comments.—I would not pretend to be com- petent to evaluate all of these characters. I would point out that Lgvtrup’s (1977) list of characters does not have the benefit of out- group comparison. That is, critical comparisons have not been made with agnathans. I note, for example, that a cursory glance at lampreys (Jol- lie 1973, Figs. 13-15) indicates they have a sim- ilar pattern for the first two cranial nerves as chondrichthyans and Latimeria. Thus a case could be made for this character being primitive. The distribution of the gray matter in the spinal cord is probably a derived character for gna- thostomes. I base this conclusion on the fact that the condition in teleosts and amphibians is not that much different from Latimeria and sharks and that gnathostomes (in having distinct ‘‘X- shaped” distribution of gray matter) differ from cyclostomes and amphioxus. Indeed, the distri- bution of gray matter in sharks seems to me to be as similar to teleosts as to Latimeria and the specific similarities between Latimeria and chondrichthyans could be explained as the re- tention of a primitive character. The simple structure of the thymus and thyroid glands might likewise be primitive, but comparisons will have to be made with agnathans and euosteichthyans before the status of these characters can be as- sessed. Jollie (1973) mentions Mauthner’s cells in teleosts and salamanders. What its distribution is within Euosteichthyes and Agnatha is un- known to me. Again, more comparisons will have to be made. Other characters, such as a fatty liver and large eggs can not be evaluated without additional information (for example—is the oil retained in the liver if similar biochem- istry connoting similar and homologous bio- chemical pathways’). Finally, there is the question of osmoregula- tion by urea. I have previously dismissed this evidence because lungfishes also retain urea (Wiley 1979). Thus, urea retention, if considered homologous at al, can be interpreted as a prim- itive character of gnathostomes (see Griffith and Pang, this volume). SUMMARY In view of the evidence summarized in this paper, Latimeria chalumnae may be considered (1) an osteichthyan and (2) the sister group of all other Recent osteichthyans. Twelve shared de- rived characters (synapomorphies) corroborate the first hypothesis and five shared derived char- acters corroborate the second hypothesis (Fig. 4). The phylogenetic relationships of Latimeria chalumnae advocated here are summarized in the form of a classification, shown below. INFRAPHYLUM Gnathostomata SUPERCLASS Chondrichthyes SUPERCLASS Teleostomi CLASS Actinistia ORDER Coelacanthiformes Latimeria chalumnae CLASS Euosteichthyes SUBCLASS Sarcopterygii SUBCLASS Actinopterygii ACKNOWLEDGMENTS I thank Bobb Schaeffer (American Museum of Natural History) and Hans-Peter Schultze (University of Kansas) for reading and providing valuable comments on earlier drafts. I have ben- WILEY: GILL ARCH MUSCLES efited from numerous discussions with Donn Rosen (AMNH) concerning sarcopterygian re- lationships and with James Atz (AMNH) con- cerning pituitary glands.-Review and discussion does not imply acceptance of my hypothesis. LITERATURE CITED ANDREws, S. M. 1973. Interrelationships of crossopterygi- ans. Pages 137-177 in P. H. Greenwood, R. S. Miles, and C. Patterson, eds., Interrelationships of fishes. Academic Press, London. 1977. The axial skeleton of the coelacanth, Latime- ria. Pages 271-288 + 3 pls. in S. M. Andrews, R. S. Miles, and A. D. Walker, eds., Problems in vertebrate evolution. Academic Press, London. DeBeer, G. R. 1937. The development of the vertebrate skull. Clarendon Press, Oxford. BJERRING, H. C. 1973. Relationships of coelacanthiforms. Pages 179-205 in P. H. Greenwood, R. S. Miles, and C. Patterson, eds., Interrelationships of fishes. Academic Press, London. ENGELMANN, G. F., AND E. O. WILEY. 1977. The place of ancestor-descendant relationships in phylogeny reconstruc- tion. Syst. Zool. 26:1-11. GARDINER, B. G. 1973. Interrelationships of teleostomes. Pages 105-135 in H. P. Greenwood, R. S. Miles, and C. Patterson, eds., Interrelationships of fishes. Academic Press, London. , AND A. W. H. BARTRAM. 1977. The homologies of ventral cranial fissures in osteichthyans. Pages 227-245 in S. M. Andrews, R. S. Miles, and A. D. Walker, eds., Prob- lems in vertebrate evolution. Academic Press, London. GRAHAM-SMITH, W. 1978. On the lateral lines and dermal bones in the parietal region of some crossopterygian and dipnoan fishes. Philos. Trans. R. Soc. London 282:41-105. HENNIG, W. 1966. Phylogenetic systematics. University of Illinois Press, Urbana. JESSEN, H. 1968. The gular plates and branchiostegal rays in Aim, Elops and Polypterus. Pages 427-438 in T. Orvig, ed., Nobel Symposium 4. Current problems in lower vertebrate phylogeny. Interscience Publications, New York. Jotitie, M. 1971. Some developmental aspects of the head skeleton of the 35-37 mm Squalus acanthias foetus. J. Mor- phol. 133:17—40. . 1973. Chordate morphology. Robert E. Krieger Pub- lishing Company, Huntington, New York. Lacios, M. D. 1975. The pituitary gland of the coelacanth Latimeria chalumnae Smith. Gen. Comp. Endrocrinol. 25:126-146. 1979. The coelacanth and the chondrichthyes as sis- ter groups: a review of shared apomorph characters and a cladistic analysis and re-interpretation. This volume. Lgvrrup, S. 1977. The phylogeny of Vertebrata. John Wiley and Sons, New York. Mices, R. S. 1973. Relationships of acanthodians. Pages 63- 103 in P. H. Greenwood, R. S. Miles, and C. Patterson, eds., Interrelationships of fishes. Academic Press, London. . 1975. The relationships of the Dipnoi. Colloq. Int. C. N.R. S. 218:133-148. . 1977. Dipnoan (lungfish) skulls and the relationships 67 of the group: study based on new species from the Devonian of Australia. Zool. J. Linn. Soc., London 61: 1-328. . AND G. C. YOUNG. 1977. Placoderm interrelation- ships reconsidered in the light of new ptyctodontids from Gogo, Western Australia. Pages 123-198 in S. M. Andrews, R. S. Miles, and A. D. Walker, eds., Problems in vertebrate evolution. Academic Press, London. MILLor, J., AND J. ANTHONY. 1958. Anatomie de Latimeria chalumnae 1. Squelette, muscles et formations de soutien. CANE ReesS-vRanis- Moy-THomas, J. A., AND R. S. MILEs. 1971. fishes. W. B. Saunders Company, Philadelphia. NELSON, G. J. 1969. Gill arches and the phylogeny of fishes with notes on the classification of vertebrates. Bull. Am. Mus. Nat. Hist. 141:475-552. 1973. Relationships of clupeomorphs, with remarks on the structure of the lower jaw in fishes. Pages 333-349 in P. H. Greenwood, R. S. Miles, and C. Patterson, eds., Interrelationships of fishes. Academic Press, London. Oxtsson, R. 1968. Evolutionary significance of the *‘prolac- tin’’ cells in teleostomean fishes. Pages 455-472 in T. Orvig, ed., Nobel symposium 4. Current problems of lower ver- tebrate phylogeny. Interscience Publishers, New York. PATTERSON, C. 1977. Cartilage bones, dermal bones and membrane bones, or the exoskeleton versus the endoskel- eton. Pages 77-121 in S. M. Andrews, R. S. Miles, and A. D. Walker, eds., Problems in vertebrate paleontology. Ac- ademic Press, London. Romer, A. S. 1966. Vertebrate paleontology. The University of Chicago Press, Chicago. SCHAEFFER, B. 1968. The origin and basic radiation of the Osteichthyes. Pages 207-222 in T. Orvig, ed., Current prob- lems of lower vertebrate phylogeny. Proceeding of 4th No- bel Symposium. Interscience Publishers, New York. SCHULTZE, H.-P. 1969. Griphognathus Gross, ein lang- schnauziger Dipnoer aus dem Oberdevon von Bergish-Glad- bach (Rheinisches schiefergebirge) und von Lettland. Geol. Palaeontol. 3:21-61. THOMSON, K. S. 1967. Mechanisms of intracranial kinetics in fossil rhipidistian fishes (Crossopterygia) and their rela- tives. J. Linn. Soc. (Zool.) 46:223-253. . 1969. The biology of the lobe-finned fishes. Biol. Rev. Cambridge Philos. Soc. 44:91-154. VON WAHLERT, G. 1968. Latimeria und die Geschichte der Wirbeltiere. Eine evolutions biologische Untersuchung. Gustav Fischer Verlag, Stuttgart. WesToLlL, T. S. 1949. On the evolution of the Dipnoi. Pages 121-184 in G. L. Jepsen, G. G. Simpson, and E. Mayr, eds., Genetics, paleontology and evolution. Princeton Uni- versity Press. WiLey, E. O. 1975. Karl R. Popper, Systematics and clas- sification: a reply to Walter Bock and other evolutionary taxonomists. Syst. Zool. 24:233-243. . 1976. The systematics and biogeography of fossil and Recent gars (Actinopterygii: Lepisosteidae). Misc. Publ. Mus. Nat. Hist., Univ. Kans., No. 64:1-111. . 1979. Ventral gill arch muscles and gnathostome phy- logeny, with a new classification of vertebrates. J. Linn. Soc. London, in press. WINGSTRAND, K. G. 1966. Comparative anatomy and evo- lution of the hypophysis. Pages 58-126 in G. W. Harris and B. T. Donovan, eds., The pituitary gland. Butterworths, London. Palaeozoic OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES The Biology and Physiology of the Living Coelacanth No. 134, 11 pages December 22, 1979 OBSERVATIONS ON THE STRUCTURE OF MINERALIZED TISSUES OF THE COELACANTH, INCLUDING THE SCALES AND THEIR ASSOCIATED ODONTODES By William A. Miller Department of Oral Biology, State University of New York at Buffalo, Buffalo, New York 14226 ABSTRACT Pieces of dentary, premaxilla and pterygoid from five specimens of Latimeria cha- lumnae have been examined radiographically and histologically. Scales from seven, in addition to an embryo specimen (No. 1 from American Museum of Natural History), were examined radiographically and grossly; histological examination was made of scales from three of them. Our original findings on bone structure are confirmed. Bone remodeling in some bones results in primary and secondary haversian systems. Jaws are mainly composed of outer dense plates or laminae supported by interlaminar struts, while the teeth sit in shallow sockets in the bone and are attached by radiating smaller struts. Examination of scale rings (annuli) indicates that their number is variable on the same specimen, and even on the same scale depending along which axis the count is made. This is true of both minor and major scale rings. Embryo scales show that the initial ornamentation is by melanocytes, not odontodes, and that their orientation is related to the fine radially oriented ridges on the scale surface. Increase in thickness of scale is brought about by the addition of nonminer- alized isopedine layers. Adult scale growth may be divided into true increase in size of the scale at the periphery and modifications to the ornamentation both at the periphery (posteriorly) and on the anterior margin of the ornamental part where adjacent scales overlap. In either case melanocytes appear to be established prior to the development of odontodes. 68 MILLER: COELACANTH MINERALIZED TISSUES INTRODUCTION Brief descriptions of the histological structure of the mineralized tissues of Latimeria chalum- nae Smith have appeared since the first speci- men was caught nearly forty years ago (Bern- hauser 1961; Miller and Hobdell 1968; Peyer 1968; Miller 1969; Hobdell and Miller 1969; Iso- kawa, Toda and Kubota 1968; Grady 1970) al- though the topographical anatomy of these structures had been described in detail by Smith (1939, 1940) and Millot and Anthony (1958). Schaeffer (1977) has recently reviewed these structures in relationship to the skeletal tissues of other groups of fishes and Andrews (1977) has discussed the axial skeleton in great detail. The scales and their odontodes were de- scribed initially by Smith (1940) and by Roux (1942) and in greater detail by Smith, Hobdell and Miller (1972). Orvig (1977) has described the formation of successive generations of odon- todes in Latimeria scales and Schaeffer (1977) has discussed the tissues which make up the odontodes. MATERIALS AND METHODS Jaw material was obtained for histological ex- amination from the following specimens: Birmingham (U.K.) University, dentary. British Museum Natural History II, premaxilla and dentary. Peabody Museum, premaxilla and dentary and pterygoid. Field Museum II, premaxilla, pterygoid. Royal Society of London, premaxilla, pterygoid. Adult scales have been examined from: British Museum Natural History II. Field Museum II. Peabody Museum. Smithsonian Museum. Steinhart Aquarium, California Academy of Sci- ences. Miscellaneous scales from specimens: Nos. 84 and 85 USNM 205871. Photographs from Smith (1940). Embryo scales from embryo number 1 (A.M.N.H.) were also studied (Smith et al. 1975) Bone and tooth specimens and the whole head of BMNH II were examined by X-ray micros- copy (Hilger and Watts Intercol X-ray micro- 69 scope) at 50 or 60 KV 4 mA (Saunders and Ely 1965). Preliminary results were reported previ- ously (Hobdell and Miller 1969). Some material was recorded as stereo-pair radiographs for three dimensional representation of the bony structures. Other radiographs were taken of dis- sected material and the scales using fine-grain photographic film and a modified dermatological X-ray source (target film distance 75 cm 70 KVP 9 mA) (Miller and Radnor 1970). Selected specimens were subsequently de- mineralized, with radiographic control, in 10% formic acid and processed either through alco- hols and paraffin or through Cellosolve and ester wax. Serial or semiserial 8 um sections were stained with a wide variety of histological stains. Scales For the major investigations two sets of scales were taken from each specimen: A circumfer- ential set around the left side of the body just anterior to the root of the anterior dorsal fin; a longitudinal (horizontal) set, one scale below the lateral line (Smith, Hobdell and Miller 1972). Similar sets were obtained from embryo no. I and also other scales from along the lateral line system including nos. 1, 3 and 5. Adult scales were examined by X-ray micros- copy, microradiography, histological techniques and direct observation of cleared specimens by transmitted and reflected light (Smith, Hobdell and Miller 1972). Embryo scales were examined by contact radiography, clearing and mounting and histological techniques. RESULTS AND DISCUSSION Bone and Cartilage The radiographic appearance of the jaws has been briefly described previously (Hobdell and Miller 1969). There is a single bony element in each lower jaw although the dentaries are not directly connected in the midline (Fig. 1). These bones are composed of dense plates of mainly acellular bone on their lateral and medial as- pects, each made up of parallel lamellae. Large struts pass between these bony plates. The larg- er teeth are ankylosed by the bone of attachment to the tooth bearing bone beneath them (Hobdell and Miller 1969) which in turn is attached to the jaw bone proper by a firm fibrous ligament in- serted by Sharpey fibres into each bony com- ponent (Miller and Hobdell 1968). At the labial 70 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, No. 134 X-ray micrograph (Hilger and Watts) of den- FIGURE 1. taries of Latimeria chalumnae (BMNH II) prior to dissection. The overall trabecular pattern is evident as are the paired- tooth positions. The gular plates are visible below. aspect a separate set of small bony elements may be observed with one or two small teeth in each (Fig. 2). These also are attached by fibrous bands to the main bone of the jaw. The teeth of the premaxilla are held in a way similar to that of the other bony elements. Fine radiating struts may be observed between tooth and tooth bearing bone (Figs. 3 and 4) which is connected to jaw bone proper by fibres. In turn this bone is attached to large but apparently un- mineralized cartilaginous elements by fibrous bands (Fig. 5). The bone of the pterygoid is much lighter in structure with minor condensa- tions at the surface and around the larger tooth pairs. Teeth are ankylosed in a way similar to the other bones. Fine radiating struts were seen where the bone articulated with adjacent bones. Much of the bone studied histologically was relatively acellular in the larger bony elements. FiGu! X-ray micrograph of left dentary prior to dis- section to show the tooth-bearing bone blocks at the periphery of the major osseous component. Pe X-ray micrograph of premaxilla prior to dissec- # 8 FIGURE 3. tion. The strutted construction/arrangement is clearly visible. Tooth positions with different stages of tooth development may be observed (arrows). However the development of primary haversian systems to varying degrees, was often observed within the spaces of the trabeculae (Fig. 6). An irregular arrangement of osteocytes could be seen along with many, presumably unmineral- ized Sharpey fibres with which they might be confused at first sight. Bone tissue was lined with a tinctorially different osteoid which was particularly wide in the trabecular bone (Fig. 5). In some of the jaw bones secondary haversian systems were observed in which the outer bor- der of the bone unit was bounded by a scalloped reversal line (Fig. 7), which indicated prior bone resorption. Most bone resorption was associated with the attachment of new, developing teeth where it was frequently extensive even though adjacent to functioning teeth. Large craters were eroded FIGURE 4. from premaxilla. There is ankylosis of the tooth with a fine surrounding trabecular pattern. The coarser trabecular pattern of the rest of the bone is clearly visible. Flattened odontodes can be seen on the external/facial side of the bone (arrows). X-ray micrograph of slab cut parasagittally MILLER: COELACANTH MINERALIZED TISSUES FIGURE 5. Photomicrograph of parasagittal section of pre- maxilla. Major and minor trabecuale are present. Wide osteoid seams can be seen (arrows). The fibrous attachments between the bones and to the cartilage (C) are clearly visible. H & E x 40. prior to the attachment of the developing tooth which was observed some distance above the bone, solely within soft tissue (see below). Gross examination of jaw pieces showed that on occasion a large crater or socket could be seen on the oral surface of the bone which implied that much of the bone of attachment and pos- sibly several teeth in a group along with their tooth bearing bone had been shed. Considerable epithelial proliferation may accompany such tooth and bone loss (Miller and Hobdell 1968). Resorption was accompanied by the presence of multinculeate and mononucleate osteoclasts, smaller than, though similar to, those seen in mammalian bone remodeling. FIGURE 6. illary bone. Careful examination of the circular bodies around the haversian system reveals most of them as Sharpey fibres related to fibrous attachment. A few osteocytes are visible (arrows). van Gieson x 100. Primary haversian system (osteon) in premax- 71 FIGURE 7. A peripheral scalloped outline is visible in each (arrows). La- mellae and occasional osteocytes can be seen within each sys- tem. H & E x100. Secondary haversian system in pterygoid bone. On the whole, bone remodeling was minimal so that frequently there was incremental en- largement of bone processes in one direction in particular, with only slight accompanying re- sorption on the opposite side of the bone. Dis- tinct and asymmetrical resting lines were ob- served not only histologically (Fig. 8) but were clearly visible in microradiographs (Hobdell and Miller 1968). Some of the illustrations of Smith (1940) show similar but perhaps grosser rings on the surface of bones. The subopercular bone is such an example which in the first described specimen of Latimeria chalumnae had 16 dis- tinct major annuli and 40-44 minor ones. The gular plate that I have examined showed similar series of annular lines. Spicule of premaxilla (BMNH-II) with nine FIGURE 8. major growth rings and several minor ones. Some resorption has occurred at the lower edge. Lison x 100. 72 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, No. 134 FiGurRE 9. Tooth of dentary in situ. Dentine, cementum and bone of attachment are clearly visible. The tinctorial change associated with mesodermal enamel is visible at the outer, coronal part of the dentine. It will be noticed that the epithelium around the tooth stands away from the dentine. This space in life would be partly occupied by highly miner- alized true enamel. H & E x40. The cartilage of the jaws of Latimeria cha- lumnae shows an irregular arrangement of chon- drocytes with irregular shape and staining of the pericellular zones. There is an outer zone of pro- liferation but this is far less regular than that seen in mammalian cartilage. No areas of min- eralization were observed as is frequently seen in elasmobranch cartilage. Tooth Structure The teeth (Fig. 9) are made up of pseudo- dentine (sensu Bertin 1958) rather than ortho- dentine (Bernhauser 1961) with adjacent tubules connecting at all levels. Incremental lines are evident (Fig. 10); see also Miller and Hobdell (1968). The outer surface is covered with enamel which is distinct from the underlying dentine (Miller and Hobdell 1968; Miller 1969). This was confirmed by Grady (1970). Scanning electron microscopy also shows this tissue as a distinct FiGure 10. Photomicrograph of unerupted tooth. Bone of attachment and cementum may be seen. Incremental lines are visible with every fifth one more defined. These may be daily or possibly lunar increments. The space above the dentine is partly due to shrinkage and partly due to loss of highly min- eralized true enamel during histological processing. Gomori reticulin x40. structure (Smith 1978). The outer layer of den- tine has many characteristics of mesodermal enamel (sensu Kvam 1946) (Miller 1969). Iso- kawa et al. (1968) did not find enamel on gill- arch teeth but it should be remembered that gill- arch epithelium is endodermal in origin while dentary epithelium is ectodermal. Gillarch teeth may not have ectodermal enamel or perhaps it was lost in preparation. Radicular dentine blend- ed into a sheath of cementum-like tissue which was less well calcified than the surrounding tis- sues (Hobdell and Miller 1969). This was sepa- rated from the true bone of attachment by a PAS-positive and Gomori-negative line (Miller and Hobdell 1968). The tooth bearing bone is frequently observed to have several incremental layers on its lower surface (Fig. 11). These are asymmetrically ar- ranged to one end of each tooth/bone unit im- plying that there is a gradual tilting of this bone unit as initial teeth become worn and others are added at the end where incremental bone is thickest. The tooth bearing bones are held to the bone of the jaws by dense fibrous ligaments. Fine fibres are seen entering each bony component, the fibres joining to become a coarse central complex between the bones (Fig. 12). MILLER: COELACANTH MINERALIZED TISSUES ‘ Base of tooth bearing bone of dentary; several FiGure 11. incremental lines are visible with resulting areal increase. This seems to be associated with continued tooth development and subsequent tilting of the bone. H & E x40. Growth of Teeth We may interpret the various stages of tooth development in the following schema: 1. Initially all teeth develop on the underside of the mucous membranes at first as a thickening of the basement membrane when viewed after PAS staining. Presumably this is related to neural crest cells in the underlying mesenchyme although the first obvious mesenchymal change was early differentiation of odontoblasts. Mes- enchymal condensation, so marked an early change in mammals, did not occur until the ini- tial mineralized tissue had formed. The cap stage is formed by proliferation of the epithelium, but a well marked dental lamina is not seen. At this stage, bone well below the developing tooth starts to resorb to form a cup. 2. There is rapid growth in length of the tooth with little increase in thickness. This phenome- non has been observed in various nonmammal- ian teeth: Pristis cuspidatus (rostral teeth) (Mil- ler 1974), Amia calva (Miller and Radnor 1970) and Necturus maculosus (Miller and Rowe 1972). 3. Eruption of the tooth occurs and it is sub- sequently held in place intraorally by soft tis- sues. As a consequence, teeth may be observed in the mouth which are approaching full height but which are loose to the touch. Details of fibrous attachment between tooth Fine fibres attach to the bones per se which in turn form a dense coarser network between the bones. van Gieson x40. FiGuReE 12. bearing bone (above) and basal bone of jaw (below). 4. There is a subsequent further thickening of the dentine and development of cementum and bone of attachment as the tooth finally becomes ankylosed and fully functional. 5. The tooth bearing bone may show several incremental lines at its lower border where it is attached to the subdontal ligament (Fig. 12). This probably indicates further eruptive move- ment of the tooth and bone of attachment after ankylosis. 6. Other teeth may be added, usually on the inner or lingual side to the same tooth bearing bone. The older labial teeth may show signs of wear or fracture. Some new teeth may attach to the bone of attachment of previous teeth. 7. The presence of relatively large sockets in various parts of the mouth leads one to suggest that eventually a whole group of small teeth or odontodes may be lost including their bone of attachment and tooth bearing bone. Epithelial proliferation reminiscent of mammalian socket healing may then be seen (Miller and Hobdell 1968) which fills in the surface defect. 74 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, No. 134 Central portion of adult scale, stained in aliz- arin Red S and cleared. Note two odontodes only have stained and that the bone attachment of other odontodes has taken up the stain to varying degrees (arrows). The overlap that might be expected from Orvig’s theories (1977) is not observed. x6. FiGuRE 13. Scales The basic structure of the adult Latimeria scale has been described previously (Smith, Hobdell and Miller 1972). Each consists of a posterior third which is exposed and ornament- ed with odontodes (Smith 1940; Roux 1942; Or- vig 1977) and an anterior two-thirds overlapped by neighboring scales. Many annular rings are apparent at irregular intervals. These are pro- duced by slight variations in regular smaller ridges which radiate from the apex and are ap- proximately perpendicular to the annuli. It has been suggested that one may age indi- vidual Latimeria by scale ring counts as is pos- sible in various teleosts. However the number of annuli is extremely variable. The counting of minor annuli is very difficult and can be found to vary depending on whether one starts to count from the centre or from the periphery of the scale. There is also variation between the long and short axes of the scale which is accen- tuated when the scale is elongate as compared to oval. Even adjacent scales may vary, the centres of the scales in particular having very different numbers of annuli and yet the scales being of similar size and form. It has been sug- gested that this variation may be due to in vivo FiGure 14. Central portion of embryo scale. Ridges can be seen radiating out, in relation to a V-shaped line, in three groups. Melanocytes of varying size are arranged in rows which seem to be related to the scale ridges. An additional small colony of small melanocytes is visible on the leading edge of the ornamented part of the scale (arrow). Cleared scale x25. scale loss (possibly due to trauma) and subse- quent replacement, but adjacent scales in em- bryo No. | showed similar variation and these could not have been subjected to trauma. The scales of adult specimens are large and relatively difficult to remove and I have never seen a short scale which might be interpreted as a more rap- idly growing, replacing scale in any of the spec- imens so far examined, or in any illustrations of Latimeria scales in the literature. One is struck by the regular scale pattern although pigmenta- tion does vary and many scales are missing post- mortem. The major annuli show similar but less pro- nounced variation to the minor annuli. For in- stance those scales illustrated in Smith (1940) vary between I1 (scale 65) and 16 (scales 64 + 71). Specimen No. 84 varies between 7-10; No. 85, 10-17. The Peabody specimen to which Thomson (1966) gives a scale age of about eight years varied between 8 and 14. Odontodes At first sight odontodes appear very regular in their arrangement. However, their orientation MILLER: COELACANTH MINERALIZED TISSUES varies from generally radial to sometimes cir- cumferential. Adjacent groups may be at sharply varying angles. When viewed radiographically the bone of attachment can be seen to radiate diffusely around each odontode and the radio- graphic density of each is quite constant. Each odontode is discrete and there is not the overlap of image that one might expect from the inter- pretation of Orvig (1977) of odontode regenera- tion. However near to the apex of the scale and occasionally elsewhere a residual image of the bone of attachment was seen. Grossly a small raised ring around a space was observed. Stain- ing of several whole scales with alizarin red S prior to clearing resulted in little staining of odontodes—mature enamel and dentine do not stain readily with this technique (Miller 1959). However a few discrete odontodes had stained (Fig. 13) indicating immature calcified tissues and presumably a recently developed odontode which may either have been replacement or as is equally likely, to have been an adjustment for growth of the animal and its scales (see below). Bone of attachment also showed variations in its staining density. Vertical canals, which may be readily ob- served, open as pores on the underside of the ornamented part of the scale (Smith, Hobdell and Miller 1972). These seem to be extremely variable in their distribution being dense in some parts and missing in others. This variation may be either near the scale apex or occur periph- erally but does not appear to be related to the position or density of the odontodes on the ex- ternal surface of the scale. Neither were the ca- nals related to the major or minor annuli. It might be assumed that there are considerable but short-lived metabolic demands by the cells forming the odontodes and that consequently there would be a regular relationship between canals and odontodes. This appears not to be SO. As pointed out previously, the mineralization of this highly modified cosmoid scale is only found in the odontodes and most superficial part of the isopedine layer (Smith, Hobdell and Mil- ler 1972). However this layer of mineralization is thickened around the lateral canal to form a lightly mineralized and presumably rigid tube. In sections of embryo scales the layers of iso- pedine were observed to divide around the canal although they did not appear to be mineralized as yet when studied radiographically. The meth- 75 od of enlargement with growth would be an in- teresting study. Initially the isopedine beneath the ornamen- tation is not mineralized although its layers are laid down with the same dimensions of thickness as in the adult scale. Increase in thickness of the scale takes place by the increase in the number of isopedine layers and not by their interstitial growth. The adult scale shows variation in pigmenta- tion, indeed melanocytes may be observed in association with the external and internal sur- face of odontodes in the oral cavity as well as those on the scales. While in the adult the me- lanocytes seem to be secondarily arranged to the mineralized tissues, examination of embryo scales shows that the initial “‘ornamentation”’ is by melanocytes not by odontodes. Clearly dis- tinguishable melanocytes may be seen radiating in rows from the apex of the scale, usually as- sociated with the ridges which themselves ra- diate in three relatively distinct groups—two in the covered, unornamented, part of the scale and the third in the exposed ornamented portion (Fig. 14). That the melanocytes are the initial ornamentation is also apparent in the growth se- quences of the adult scale (see below). The me- lanocytes show great variation of size and ex- pansion. Whether this is functional, related to the fixation process or due to poor fixation per se is difficult to determine from this specimen. There is clearly an initial relationship between the melanocytes and ridges. There is also a later relationship between odontodes and melano- cytes although a relationship does not apparent- ly exist between odontodes and ridges (Smith, Hobdell and Miller 1972). To an experimental embryologist this is an extremely intriguing question; what inductive processes are involved in the formation and orientation of melano- cytes—of neural crest origin—and also cells which form the dentine of the odontodes—which may be presumed to also be of neural crest or- igin? Odontoblasts are certainly of neural crest origin in the oral teeth of some of the higher nonmammalian vertebrates (Horstadius 1950). Scale Growth Once the scale has been established it would appear that most of the areal growth takes place at the periphery of the scale and that odontodes are added at the edge so that the oldest ones are at the centre (apex) of the scale and the youngest 76 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, No. 134 Leading edge of ornamented portion of adult FiGure 1S. scale. A discrete colony of melanocytes is visible with no associated odontodes. Cleared scale x6. at the periphery. This is also probably true of the melanocytes. Orvig (1977) has drawn atten- tion to the replacement activity that may be ob- served at the apex. Odontodes are shed centrally leaving the bone of attachment as a flat shell. New odontodes sometimes develop over these remnants. Radiographs indicate that the dentine crowns are always discrete and that the coronal dentine of the first odontode does not become involved in the attachment of its successor and form an ‘‘odonto-complex’’ (sensu Orvig 1977). However usually the new odontodes are distinct from their predecessors (Fig. 13). There are clearly two phases in scale growth: first its enlargement and second, modifications of the ornamentation related to that growth and adjustments that occur between adjacent scales as the animal grows overall. Along the anterior edge of the ornamented portion one may observe occasional young and partially developed odontodes with no signs of previous bone of attachment related to other odontodes (Fig. 13). Many small melanocytes may be seen in this region. It might be tempting to suggest that as they were not exposed to the external environment they were not expanded and sc smaller than their neighbors. However in Ornamented portion of adult scale denuded of FiGure 16. odontodes presumably due to trauma. A central group of firm- ly attached odontodes can be seen. The small odontodes around the traumatized area may be new structures “‘recolon- izing’ the damaged scale. Cleared specimen x6. fish living at any depth this light related phenom- enon is unlikely. Similar observations on em- bryo scales (Fig. 14) strengthen this conclusion. In one specimen a relatively large “‘colony’’ of melanocytes was observed expanding onto the normally unornamented part of an adult scale. Their orientation was possibly related to the ridges of the unornamented portion (Fig. 15). It would seem that new odontodes are preceded by melanocyte proliferation in both initial de- velopment (see above and Fig. 14) and in later normal increase in area of the ornamented parts of the scale. Repair of the ornamented portion may also occur. Figure 16 illustrates a new set of odon- todes developing in the centre of a bare area. The scale came from the more bulbous part of the fish and may have been scraped against the substrate with consequent loss of odontodes. The colony was firmly attached and consisted of mostly small odontodes, particularly visible in the comparable radiograph. Around the trau- matized area small odontodes were observed. These may also have been recolonizing the area centripetally. We now pass onto the vexing question of how MILLER: COELACANTH MINERALIZED TISSUES does one age Latimeria. From the previous dis- cussion it is evident that the counting of scale annular rings is extremely uncertain and that although one may be able to count rings they are variable in number even between adjacent scales. There is also variation between the long axis and the short axis of the scale and this is accentuated when one compares a round central body scale with a more oval or elongated tail or posterior scale or a belly scale. Even though one can count with some degree of certainty major scale rings, it is apparent that even this param- eter is variable. Dentine is laid down in increments but the teeth are constantly being replaced. Are these increments daily, or are they perhaps a reflec- tion of the lunar migration that has been hinted at by McCosker (this volume) in his description of the possible lunar related movements of this fish? Teeth per se are not useful for aging. For accurate aging one needs an organ that shows incremental changes but with minimal variabil- ity. If one examines the various easily observed bones such as the subopercula (Smith 1940) or the gular plates (Fig. 17) one is struck by the regular incremental lines which are visible on the surface. Radiographs should be even more clear and the gulars are particularly appropriate. One would suggest that such material would be more appropriate than scale ring counts. Ed- ward Brothers (1977, personal communication) has suggested that the annuli of ear ossicles might be most appropriate for assessing the age of Latimeria specimens. However the techni- calities may be too complicated for routine use and cannot be used in the field. A gular or sub- opercula bone may be readily examined unaid- ed: It is suggested that scale ring counts are not appropriate for age determination but that the major resting lines in certain easily investigated bones such as the subopercula or gular plate may be more suitable. Lagios (1977) has suggested that the coel- acanth and Chondrichthyes are sister groups. As far as the skin and scales are concerned the great morphological similarity between odontodes on the scales of Latimeria and the placoid scales of selachians is very striking. Smith, Hobdell and Miller (1972) suggested however that there were close similarities between Latimeria scales and those of modern lungfish (Kerr 1955). While Wd, FiGuRE 17. annular rings. Field Museum II. Internal surface of gular plate showing 9-10 his data (Lagios 1972, 1974, 1975) are very per- suasive, the data on scales and skin are not very helpful. Smith (1978) has pointed out that true dental enamel is found in mammals, reptiles, amphibians, as well as actinistians (coelacanth) while elasmobranchs and actinopterygians have a distinctive enameloid or mesodermal enamel. Thus the outward similarity between elasmo- branch dermal denticles and coelacanth scale odontodes is not found in their structure and possibly their development. The oral teeth of both groups are likewise dissimilar. ACKNOWLEDGMENTS I wish to thank my previous coauthors, Martin H. Hobdell and Moya Meredith Smith, for past cooperation in preparing some of the material presented here; also the latter for making more recent information on Latimeria enamel avail- able to me while in manuscript form. I also wish to thank K. Thomson (Peabody Museum, Yale University), K. Liem (Field Museum, Chicago). S. H. Weitzman (Smithsonian Institution, Washington, D.C.) and most of all P. H. Green- wood, who greatly encouraged these studies at their beginning, for making material available for study. The more recent help and cooperation of George Brown, John McCosker and Mike La- gios has been a great pleasure. Dr. Lagios is to be thanked for making the scales of embryo no. 1 (A.M.N.H.) available for study. The technical assistance of Mona Everett, Christopher Miller and Glenn Denis was invalu- able as has been the photographic expertise of Malcolm McQuaig and Hector Velasco. The manuscript was typed by Mary Parkhill and Rose Parkhill. LITERATURE CITED ANDREWS, S. M. 1977. The axial skeleton of the coelacanth, Latimeria. Pages 271-288 in S. M. Andrews, R. S. Miles, 78 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, No. 134 and A. D. Walker, eds., Problems in vertebrate evolution. Academic Press, London. BERNHAUSER, A. 1961. Zur Knocken- und Zahnhistologie von Latimeria chalumnae Smith und einiger Fossil formen. S. B. Ost Akad. Wiss., mat.-naturw. 170:119-137. BERTIN, L. 1958. Tissues squelettiques. Pages 532-55 in P. P. Grassé ed., Traité de zoologie, Vol. 13. Masson & Cie., Paris. Hospe.l, M. H., AND W. A. MILLER. 1969. Radiographic anatomy of the teeth and tooth supporting tissues of Lati- meria chalumnae. Arch. Oral Biol. 14:855-858. Horstapius, S. 1950. The neural crest, its properties and derivatives in the light of experimental research. Oxford Univ. Press, London and New York. IsoKAWA, S., Y. TODA, AND K. KuBoTa. 1968. A histolog- ical observation of a coelacanth (Latimeria chalumnae). J. Nihon Univ. Sch. Dent. 14:102-114. Kerr, T. 1955. The scales of modern lungfish. Proc. Zool. Soc. Lond. 125:335-345. Kvam, T. 1946. Comparative study of the ontogenetic and phylogenetic development of dental enamel. Norski Tan- dloege Foren. Tid. 56 (Suppl.): 1-129. Lacios, M. D. 1972. Evidence for a hypothalamo-hypophys- ial portal vascular system in the Coelacanth Latimeria cha- lumnae Smith. Gen. Comp. Endocrinol. 18 (1):73-82. . 1974. Granular epithelioid (Juxtaglomerular) cell and renovascular morphology of the Coelacanth Latimeria cha- lumnae Smith (Crossopterygii) compared with that of other fishes. Gen. Comp. Endocrinol. 2296-307. 1975. The pituitary gland of the Coelacanth Lati- meria chalumnae Smith. Gen. Comp. Endocrinol. 25:126— 146. Locket, A. 70:456—458. McCosker, J. E. 1979. Inferred natural history of the living Coelacanth. This volume. MILLER, W. A. 1959. Layering in dentin caries as demon- strated by localization of dyes. M.S. Thesis, Univ. of IIli- nois. 1976. A future for the coelacanth? New Sci. . 1968. Preliminary report on the histology of the den- tal and paradental tissues of Latimeria chalumnae (Smith) with a note on tooth replacement. Arch. Oral Biol. 13:1289- 1291. 1974. Observations on the developing rostrum and rostral teeth of sawfish: Pristis perottetei and P. cuspida- tus. Copeia 1974 (2):311-318. , AND C. J. P. RADNorR. 1970. Tooth replacement pat- terns in young Caiman sclerops. J. Morphol. 130:501—510. , AND D. J. Rowe. 1973. Variations in tooth replace- ment patterns in adult Necturus maculosus. J. Morphol. 140:63-76. MILLoT, J., AND J. ANTHONY. 1958. Latimeria chalumnae, dernier des Crossopterygiens. Pages 2553-3597 in P. P. Grasse, ed., Traite de zoologie, Vol. 13 (3). Masson et Cie., Paris. OrviGc, T. 1967. Phylogeny of tooth tissues: evolution of some calcified tissues in early vertebrates. Pages 45-110 in A. E. W. Miles, ed., Structural and chemical Organization of teeth. Academic Press, New York and London. 1977. A survey of odontodes (*“‘dermal teeth’) from developmental, structural, functional and phyletic points of view. Pages 53-75 in S. M. Andrews, R. S. Miles, and A. D. Walker, eds., Problems in vertebrate evolution. Aca- demic Press, London. Peyer, B. 1968. Comparative odontology. Univ. of Chicago Press, Chicago and London. 347 pp. and pls. Romer, A. S. 1946. The early evolution of fishes. Q. Rev. Biol. 21:33-69. Roux, G. H. 1942. The microscopic anatomy of the Lati- meria scale. S. Afr. J. Med. Sci., Suppl. 7, pp. 1-18. SAUNDERS, R. L. DE C. H., AND R. V. Ety. 1965. Optique des rayons X et microanalyse. R. Castaing, P. Deschamps, and J. Philibert, eds., Herman, Paris. SCHAEFFER, B. 1953. Latimeria and the history of coelacanth fishes. Trans. N.Y. Acad. Sci. 15:170-178. . 1977. The dermal skeleton in fishes. Pages 25—S2 in S. M. Andrews, R. S. Miles, and A. D. Walker, eds., Prob- lems in vertebrate evolution. Academic Press, London. Situ, C. L., C. S. RAND, B. SCHAEFFER, AND J. W. ATZz. 1975. Latimeria, the living coelacanth, is ovoviviparous. Science 90:1105—1106. SmitH, J. L. B. 1939. A surviving fish of the order Actinistia. Trans. R. Soc. S. Afr. 27:47-S0. 1940. A living coelacanthid fish from South Africa. Trans. R. Soc. S. Afr. 28:1-106. SmitH, M. M. 1978. Enamel in the oral teeth of Latimeria chalumnae: a scanning electronmicroscope study. J. Zool Lond. 185:355-369. , M. H. HoBDELL, AND W. A. MILLER. 1972. The structure of the scales of Latimeria chalumnae. J. Zool. Lond. 167:501—509. THomson, K. S. 1966. Intracranial mobility in the coel- acanth. Science 153:999-1000. OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES The Biology and Physiology of the Living Coelacanth No. 134 15 pages December 22, 1979 MECHANISMS OF OSMOREGULATION IN THE COELACANTH: EVOLUTIONARY IMPLICATIONS By Robert W. Griffith Department of Biology, Southeastern Massachusetts University, North Dartmouth, Massachusetts 02747 and Peter K. T. Pang Department of Pharmacology and Therapeutics, Texas Tech University School of Medicine, Lubbock, Texas 79409 ABSTRACT Latimeria maintains a blood osmolarity near that of its sea water environment by retaining urea and trimethylamine oxide in concentrations similar to those of chon- drichthians. Ion levels are fairly low, as in teleosts, and evidence is presented suggesting that the rectal gland is involved as a sodium secreting organ. Bladder urine of Latimeria is isouremic with the serum, contrasting with the elasmobranchs which reabsorb most of the filtered urea. The evolutionary implications of the similarity between the methods of osmoregulation in Latimeria and in the chondrichthians are discussed. Because of (1) the lack of the characteristic chondrichthian mechanism of renal urea reabsorption, (2) the use of different mechanisms for internal fertilization, a necessary concomitant of the urea retention habitus in fishes and (3) the independent occurrence of urea retention in the euryhaline amphibian, Rana cancrivora it is suggested that urea reten- tion in Latimeria evolved independently of its occurrence in the cartilaginous fishes. The factors which might favor the appearance of urea retention as an adaptation to the marine environment rather than the hyposmotic mechanism used by teleost fishes are discussed. 79 80 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, No. 134 INTRODUCTION According to traditional views on the phylog- eny of vertebrates, the living coelacanth, Lati- meria chalumnae, occupies a crucial position. Although controversy surrounds the exact po- sition of the coelacanths relative to other cros- sopterygians and the Dipnoi (Miles 1975; Gar- diner 1973), few students of the subject doubt that the coelacanths branched off close to the origin of the tetrapods. This fact, plus the ro- mance surrounding the discovery of the living coelacanth by J. L. B. Smith in 1938 have made biologists in general and physiologists in partic- ular keenly interested in the coelacanths as a potential index of the biology of the prototetra- pod. Hence, it was considerably surprising when Pickford and Grant (1967) and Brown and Brown (1967) reported that the coelacanth pos- sessed an osmoregularoty mechanism apparent- ly identical with that of sharks and their kin and unlike that of other fishes and tetrapods [see also Lutz and Robertson (1971); Griffith et al. (1974)]. That mechanism is the possession of a blood osmolarity similar to that of seawater by virtue of the presence of very high urea levels. This feature was long thought to be a unique (and ubiquitous) adaptation of the elasmobranch fishes and represented such a characteristic ele- ment of their biology that the eminent renal physiologist Homer Smith suggested (perhaps not facetiously) that the sharks and their kin would more aptly be known by a taxonomic ap- pellation which reflected the urea retention hab- itus rather than by the less inclusive and exclu- sive term Elasmobranchii (Smith 1953). From this arises the question: is the posses- sion of a common feature such as urea retention an indication of some special evolutionary link between the sharks and the coelacanth; could it be an ancestral feature possessed by all early fishes and subsequently lost in groups other than the sharks and coelacanth; or is the linkage a spurious One due to the independent acquisition of the feature in the two groups? More recently, a variety of soft tissue features have been de- scribed unique to elasmobranchs and the coel- acanth (Lagios 1975; Epple and Brinn 1975; Chavin 1976; see also Lagios, this volume) which lends some credence to the notion that the elasmobranchs and Latimeria share a com- mon and unique ancestry. In this report we intend to analyze the os- moregulatory mechanisms in Latimeria and sharks in search of clues to determine if the fea- ture was independently derived in the two groups or is truly homologous. We also intend to see whether any of the other features that have been proposed as unique specializations of the two groups might not be consequences of urea retention. Surely such a major biochemical adaptation must affect other features of the bi- ology and physiology of Latimeria and the sharks. Finally, we would like to suggest an ex- planation for the adoption of the urea retention method of osmoregulating in some fishes and the adoption of hyposmotic mechanisms in other groups of fishes. Most of the actual data pre- sented here have been previously published else- where (Griffith et al. 1974, 1975, 1976; Griffith and Burdick 1976), but the format of this sym- posium permits a synthesis and a much deeper consideration of the evolutionary implications of this work. PRESENTATION OF DATA Blood Chemistry A comparison of the principal osmotically-ac- tive solute concentrations in the blood serum of Latimeria with those of representatives of other marine fish groups is given in Table 1. An in depth and specific comparison of different groups of fishes for each blood component would be beyond the scope of this discussion since a variety of factors including stress, sex, environmental salinity, diet and species vari- ability can influence levels considerably. The reader is referred to Griffith et al. (1974) for a discussion of these problems as they relate to the blood composition of Latimeria. Some very general conclusions can be safely drawn, how- ever. For inorganic ions (Nat, Kt, Catt, Mgtt, Ch. HCO; :SO,-; 20.3) Latimeria most closely resembles Fundulus and other teleost fishes (see review by Homes and Donaldson, 1969). This is especially true of the monovalent ions Na‘ and Cl which are near sea water levels in Myxine (ca. 500 mM) around 300 mM in hol- ocephalans and elasmobranchs and under 200 mM in teleosts and Latimeria. In contrast, the total blood osmolarity of Latimeria places it with Myxine, holocephalans and elasmobranchs with levels around the osmolarity of sea water GRIFFITH & PANG: COELACANTH OSMOREGULATION TABLE 1. BLOoD CHEMISTRY OF SOME SELECTED Ma- RINE FISHES. Rat- Coel- Hag- fish! Shark? acanth* Teleost* fish! Chi- Squa- Lati- Fundu- Myxine —maera lus meria lus glu- mon- acan- cha- hetero- tinosa strosa thias lumnae clitus Nat 487 338 296 197 183 Kt 8.4 11.7 Ted 5.8 4.8 (Caceh 4.8 4.3 310 4.8 2.3 Mg** 9.3 6.1 3.5 553 75) cl” 500 353 276 187 146 HCO, ee 2.6 — 9.6 13.3 SO, S7) S22 3.1 4.8 — PO, 0.4 — 2.4@ 551 523) Urea 2ASe 332 308 377 4.0** TMAO 38 0.0 72 122 ae AA’s “ — 11.6 16.0 Sens Osmol. 969 1,046 998 932 363 ' After Robertson (1976 and 1966). * After Robertson (1975). 3 After Griffith et al. (1974). 4 After Pickford et al. (1969). * Total non-protein nitrogen = 38 mg N/kg (TMAO prob- ably not included). ** Total non-protein nitrogen = 46.3 mg N/l TMAO prob- ably not included). (@ Total phosphate. (ca. 1,000 mOsm/l) and distinguishes it from te- leosts, euryhaline lampreys in sea water and [with one exception, the crab-eating frog (Gor- don et al. 1961)], from marine tetrapods. Note, however, that osmolarity in Latimeria is ac- tually slightly less than that of environmental water (932 vs. 1,035 mOsm/l). As is now well known, high blood osmolarity is obtained by virtue of high blood levels of nitrogenous com- pounds, principally urea and trimethylamine ox- ide (TMAO), in the coelacanth, elasmobranchs and holocephalans but by having high ion levels in Myxine. Hence, the blood chemistry of Latimeria shows some points of similarity to that of acti- nopterygian fishes (blood ion levels), and others to chondrichthian fishes (high osmolarity, high urea and TMAO). With respect to the nitroge- nous compounds, we suspect that high urea is the more meaningful feature: trimethylamine ox- ide levels are apparently strongly influenced by the diet (Groninger 1959), the feature is not ubiquitous in the chondrichthians since holo- cephalans lack it or have low levels (Read 1971; 81 TABLE 2. SODIUM-POTASSIUM ACTIVATED ATPASE IN Latimeria AND ELASMOBRANCH TISSUES. (Values in «M P;,/ 60!/mg protein.) Coel- Dogfish Skate acanth Mustelus Raja Latimeria canis erinacea chalumnae Rectal gland Sel! 4.5 5.1 Kidney 2.3 ite Ded Rostral organ — — 3.0 Muscle 1S) 1.0 0.6 Gills 0.7 0.9 = Liver — — 0.6 After Griffith and Burdick (1976). Robertson 1976); it is an important intracellular osmotic component in many teleosts (Norris and Benoit 1945) and it is found in substantial quan- tities in the blood serum of some deep sea and inshore marine teleosts (Griffith, unpublished). Rectal Gland Function In elasmobranchs, a substantial portion (but not all) of the excess sodium and chloride that enters the fish via passive routes is secreted by a specialized organ, the rectal gland (Burger and Hess 1960). Present evidence suggests that al- ternative routes of salt secretion can compen- sate for the rectal gland in its absence (Haywood 1975a; Chan et al. 1967). The coelacanth pos- sesses a rectal gland in the same anatomical po- TABLE 3. URINE COMPOSITION IN SOME SELECTED Ma- RINE FISHES. (Values are in mM/I: numbers in parentheses are urine/plasma ratios.) Hagfish Coel- Teleost Epta- Shark acanth Lophius tretus Squalus Latimeria pisca- stouti acanthias — chalumnae torius Nat 553 (1.0) 240 (1.0) 184 (0.9) 11 (0.1) Kt 11 (1.6) 2 (0.5) 9 (1.5) 2 (0.4) Cais 4 (0.8) 3 (0.9) 2 (0.4) UX CXS) Mg** 15 (1.2) 40 (33.3) 30 (5.7) 137 (54.5) Cl 548 (1.0) 240 (1.0) 15 (0.1) 132 (0.9) PO, 9 (3.9) 33 (34.0) 38 (5.5) 2 (0.3) SO, 7 (8.6) 70 (140) 104 (21.6) 42 (35.0) Urea 9 (1.8) 100 (0.3) 384 (1.0) 0.6 (1.9) TMAO — 10 (0.1) 94 (0.8) 13 (1.5) Osmol. — 800 (0.8) 962 (1.0) 406 (0.9) Modified after Griffith et al. (1976). Original data from Munz and McFarland (1964); Burger (1967); Brull and Nizet (1953); Brull and Cuypers (1954) and Griffith et al. (1976). 82 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, No. 134 sition (Millot and Anthony 1972) and with iden- tical ultrastructural features (Lemire 1976) as that of elasmobranchs. Investigations of the bio- chemistry of the rectal gland of Latimeria (Grif- fith and Burdick 1976) have demonstrated levels of (Na+, K*)-ATPase which show a very close parallel to the high levels found in the rectal glands of representative elasmobranchs (Table 2). (Na*, K*)-ATPase is, of course, an enzyme closely implicated in active salt transport in the rectal gland of elasmobranchs and in the salt se- creting organs of other animals (Bonting 1966). It must be pointed out that a homology between the rectal glands of elasmobranchs and Lati- meria has yet to be demonstrated; the rectal gland of the holocephalans differs considerably in structure from that of the elasmobranchs proper (Fange and Fugelli 1963; Read 1971; Stahl 1967). Kidney Function in Latmeria One of the most crucial physiological mech- anisms permitting the elasmobranch fishes to maintain high internal urea concentrations is the renal reabsorption of most of the urea that is filtered through the glomerulus (see reviews by Smith 1936; Hickman and Trump 1969; Schmidt- Nielsen 1972). This is also true of the holoceph- alans (Read 1971), but apparently not of the crab-eating frog which also manages to maintain high blood levels of urea (Schmidt-Nielsen and Lee 1962). In elasmobranchs, urea reabsorption appears to involve a countercurrent exchange mechanism in the kidney (Boylan 1972; Deetjen et al. 1972; Stolte et al. 1977), and also to be linked with sodium reabsorption (Schmidt-Niel- sen et al. 1972). Data on the solute composition of coelacanth bladder urine are compared with that of a selec- tion of other marine fishes in Table 3. We have also included values for the ratio of levels in urine to serum which may be used as an index of which substances are conserved and which eliminated. Although the coelacanth urine was admittedly from a severely stressed fish (Locket and Griffith 1972), many features are as antici- pated: high polyvalent ion (Mgt+, SO,=, PO,=>) U/S ratios are found as is characteristic of the urine of most marine animals, and monovalent ion levels (Na* and Cl) are below serum levels as is true of most fishes. Osmolarity is identical to that of serum, a condition approached in many healthy marine fishes of various groups (Hickman and Trump 1969). Rather unexpect- edly, the coelacanth appears to be unable to reabsorb urea and the urine is characterized by levels of TMAO only slightly lower than blood serum. That the urine urea data in Latimeria are not a pathological consequence of stress is suggest- ed by the calculation of Phosphate/Nitrogen ra- tios (Smith 193la) which come out close to those of normal carnivorous vertebrates (Griffith et al. 1976) and contrast with those of elasmobranchs which have high urinary P/N ratios since the bulk of the urea is lost through extrarenal routes (Smith 193la, 1931b; Chan and Wong 1977). Also, even severe stress never elicits isouremia in elasmobranchs (Forster et al. 1972). The U/P ratio of 0.8 for TMAO in Latimeria may (1) be indicative of some reabsorption of this com- pound, (2) may be a consequence of limited glo- merular permeability to TMAO or (3) more like- ly may be attributable to elevated blood levels during stress not being fully reflected in the blad- der urine which presumably included some urine formed prior to severe stress. Muscle levels of TMAO are much higher in Latimeria than blood levels as is also true of elasmobranchs, holoce- phalans and teleosts (Lutz and Robertson 1971; review by Lombardini et al., this volume), and could act as a source of blood TMAO during severe muscular exercise. DISCUSSION How the Coelacanth Osmoregulates Although based upon a single moribund spec- imen (Locket and Griffith 1972) the data we have presented here give us considerable insight into the mechanisms the coelacanth uses to regulate solutes and water. If the blood osmolarity is in- deed lower than environmental values as our data suggest, the coelacanth must be losing water to its environment and presumably drinks sea water to replace the lost fluid as do teleosts (Smith 1930a; Conte 1969). Preliminary data on the composition of gut water (Griffith, unpub- lished) suggests that this is so; sodium and chlo- ride levels drop from the stomach to the large intestine and rectum whereas osmolarity rises, possibly reflecting an ion-linked water transport in the gut comparable to that of teleosts. How- ever, it has been recently suggested that the true habitat of Latimeria may be in subterranean caves with a dilution of the media by aquifers GRIFFITH & PANG: COELACANTH OSMOREGULATION (McCosker, this volume). The hyposmotic blood of Latimeria might be regarded as circumstan- tial evidence favoring this hypothesis. Latimeria regulates its blood osmolarity at levels near that of its environment while main- taining total electrolytes at levels only 40% those of sea water. We have seen that this is accom- plished by building up high blood concentrations of organic compounds, principally urea and TMAO. Although the mechanisms involved in the maintenance of these high concentrations of nitrogenous products remain a topic for specu- lation, it appears that renal reabsorption is not a factor. Rather, we suspect that the mainte- nance of high blood urea and TMAO is a con- sequence of one or more of the following fac- tors: (1) low glomerular filtration rate, (2) extremely low branchial and epithelial perme- ability to urea and (3) high production rates of urea from hepatic ornithine-urea cycle enzy- matic pathways. Although Latimeria kidneys appear to be well supplied with moderately large glomeruli (Millot and Anthony 1973: Lagios 1974), their somewhat hyposmotic condition places them in negative water balance which should lead to a low GFR. Although conclusive demonstrations of an important role of the renin- angiotensin system in osmorregulation of other fishes is lacking, the presence of a renin-angio- tensin system in Latimeria (Nishimura et al. 1973) might provide a mechanism for control of renal function (Lagios 1974). The gill area of Latimeria relative to body weight is the lowest reported for any fish (Hughes 1972, 1976; Hughes and Morgan 1973) and Latimeria is a large animal offering, of course, a low surface area for urea loss relative to the volume of urea synthesizing tissue. The activity levels of en- zymes of the ornithine-urea cycle determined in slowly frozen Latimeria liver are comparable to those of elasmobranchs (Brown and Brown 1967; Goldstein et al. 1973), but it is feasible that even higher levels might characterize coelacanth tissue if it had been frozen more promptly or assayed when fresh. The regulation of electrolytes in the coel- acanth would seem to involve the kidneys, the rectal gland and, possibly, the gills. The kidneys function primarily in the excretion of polyvalent electrolytes such as Mg**, SO, and PO, and in the selective elimination of certain organic substances such as glucuronides and creatine and the conservation of others (glucose, lactate, 83 protein etc.). Although there is no direct evi- dence that the rectal gland functions in the se- cretion of NaCl, this would seem to be a likely conclusion in view of the high (Na*, K*)-ATP- ase levels (Griffith anc» Burdick 1976), the ultra- structural similarity to the elasmobranch rectal gland which functions in this way (Lemire 1976) and the apparent inability of the kidney to con- centrate NaCl. There is no evidence for or against a role of the gills in ion secretion in Lati- meria, although the gills do seem to play a sig- nicant role in ion transport in agnathans, elas- mobranchs and teleosts (review by Conte 1969) and such a role would not be unexpected. ‘Chloride’ cells are sparse in the gills of Lati- meria as they are in elasmobranchs (Hughes 1976). Interpreting Shared Features In comparing the osmoregulatory mechanisms of Latimeria with those of other marine fishes, perhaps the most notable feature is the unique occurrence of high blood osmolarity by virtue of high blood urea levels in the coelacanth and in the elasmobranch fishes and related holoceph- alans. For a comparative biologist this obser- vation immediately invites the question of the evolutionary significance of urea retention. Like any other physiological or morphological fea- ture, there are three possible explanations for urea retention being a uniquely shared charac- ter: (1) that the feature is a primitive one that has been lost in other more advanced lines, (2) that it is a specialized feature developed inde- pendently in the two groups and tells us nothing of their phylogeny and (3) that it is a specialized feature which evolved only once and is evidence of a uniquely shared phylogenetic relationship between the two groups (see also Atz 1973; Grif- fith et al. 1974; Lagios, this volume). Vertebrate Phylogenies Traditional phylogenetic schemes (e.g. Romer 1966; Moy-Thomas and Miles 1971; Schaeffer 1968; Gardiner 1973; Miles 1973) portray the coelacanth as being close to the origin of the tetrapods, sharing a common ancestry with oth- er bony fishes including the dipnoans and acti- nopterygians and having only a remote relation- ship to the elasmobranchs and holocephalans. The restriction of urea retention to the coel- acanth and chondrichthians may be concordant with traditional views on vertebrate phylogeny 84 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, No. 134 ELASMOBRANCH TELEOST COELACANTH TETRAPOD LL ELASMOBRANCH TELEOST COELACANTH TETRAPOD | ia FIGURE lI. ELASMOBRANCH COELACANTH TELEOST TETRAPOD ae TELEOST TETRAPOD COELACANTH ELASMOBRANCH uae Urea retention (indicated by shading) and some alternate vertebrate phylogenies: Possibility 1. A traditional scheme with urea retention as an ancestral trait which Is lost in teleosts and tetrapods. Possibility 2. A traditional scheme with the independent acquisition of urea retention in elasmobranchs and coelacanths. Possibility 3a. A heretical scheme with urea retention being a unique specialization evolved in the common ancestor of coelacanths and elasmobranchs. Possibility 3b. Another heretical scheme with urea retention being a unique specialization which evolved in the common ancestor of coelacanths and elasmobranchs. if the aforementioned possibilities #1 (loss of it as a primitive feature in other groups) or #2 (in- dependent acquisition) were to be the case, but if #3 were true (urea retention is a unique spe- cialization evolved only once) then the tradi- tional scheme must be discarded. This may be visualized in Figure 1, where possibilities #1 and #2 are shown in terms of traditional phy- logenies and possibility #3 is shown with two alternative phyletic schemes. We might point out that although possibility #3 (unique special- ization) directly invalidates the traditional phy- logeny, possibilities #1 (loss in other groups) and #2 (independent acquisition) do not unique- ly support the traditional phylogeny since the “‘heretical’’ phylogenies will fit equally as well. It is unfortunately the case that the fossil record is ambiguous both in regards to the precise re- lationships of the coelacanths relative to the chondrichthians and other major groups of fish- es (cf. Bjerring 1973; Jarvik 1968a, 1968b) and in terms of their early environmental history rel- ative to the marine environment (Thomson 1969, 1971; Moy-Thomas and Miles 1971). How can we decide which of the three alter- native explanations for the appearance of urea retention in chondrichthians and Latimeria is correct? Although neither will provide incon- trovertible proof, two approaches may be taken. First, we can see whether the bulk of the other characters supports the “‘heretical’’ phylogeny based upon distribution of urea retention. This would provide circumstantial evidence for or against urea retention as a unique specialization. Secondly, we might analyze the mechanisms in- volved in urea retention and see whether they are the same in the two groups and also see whether logical concomitants of urea retention are accomplished by the same mechanism in the two groups. This approach might provide a more direct means of judging whether urea retention is truly homologous and, hence will provide evi- dence for or against urea retention as an inde- pendently derived character state. A Consideration of Features Other than Urea Retention Recent studies on the soft tissue biology of Latimeria have come up with an impressive number of features which Latimeria and the chondrichthians have uniquely in common (see also Lagios, this volume). Among them are (1) a duct-associated pancreatic islet tissue (Epple and Brinn 1975), (2) adenohypophysial anatomy GRIFFITH & PANG: COELACANTH OSMOREGULATION (Lagios 1975), (3) extra sulfonation in chondro- itin sulfate (Mathews 1966, 1967) and (4) a rectal gland of apparent ion secretory function (Millot and Anthony 1972; Griffith and Burdick 1976; Lemire 1976). A number of other features are shared by the coelacanth and elasmobranchs but not necessarily uniquely including high TMAO levels (Griffith et al. 1974), high levels of en- zymes of the ornithine-urea cycle and trimethy- lamine oxidase (Brown and Brown 1967; Gold- stein et al. 1973), an ovoviviparous breeding mode with extremely large eggs (Griffith and Thomson 1973; Smith et al. 1975), a single me- dian compact thyroid gland (Chavin 1976) and, not trivially, both the elasmobranchs and Lati- meria are large fish. On the other hand, there are a variety of features which characterize La- timeria and other bony fishes that do not occur in chondrichthians including (1) true scales (2) internal endochondral bone (3) an operculum (4) a gas-filled diverticulum off the gut, the lung or gas bladder and (5) blood levels of monovalent ions (Griffith et al. 1974), and there are a number of characters which link the coelacanth and fos- sil rhipidistians with the earliest tetrapods in- cluding the structure of the limbs and girdles, the arrangement of the bones of the skull, and the presence of a kinetic skull with a joint sys- tem between the orbitotemporal and otic regions of the endocranium. In fact, arguments may be made for essen- tially all of the above characters as being (1) independently acquired (2) primitive features lost in other lines or (3) unique specializations. A compilation of the number of features held in common between the two groups actually pro- vides us with little insight into true phylogenetic relationships at the level of major groups as we are dealing with here. Most structural and func- tional relationships will change much more rap- idly in evolutionarily labile groups such as te- leosts and tetrapods while remaining in a near- primitive state in more static groups such as the lungfishes, coelacanths and chondrosteans. On the other hand, one should avoid making the assumption that, since in other features two evo- lutionary lines remain static, the occurrence of what can only be regarded as a highly special- ized feature that is uniquely shared in the two groups is a priori evidence of a close relation- ship. Different character complexes may evolve at different rates depending upon environmental pressures (see DeBeer 1954). 85 Evidence for the Non-independence of Charac- ters Returning again to urea retention, is it possi- ble that any of the other features linking Lati- meria and the chondrichthians might be inti- mately related to urea retention rather than being truly independent characters? Unques- tionably, some are. The retention of high TMAO levels in the blood is probably related to the re- tention of urea, since a reduced loss of one com- pound across surface areas (and/or through the kidney) is probably due to a rather general re- duction in permeability towards solutes. Elas- mobranchs are characterized by a rather low permeability to ions (Haywood 1975; Payan and Maetz 1973; Carrier and Evans 1972; Pang, Grif- fith and Maetz, unpublished) as well as to urea. High activities of enzymes of the ornithine-urea cycle would seem to be a prerequisite for main- taining high internal urea concentrations in the face of some unavoidable loss through the kid- neys and across the body surfaces, and high titres of TMAOase (Goldstein et al. 1973) prob- ably have a similar cause. Provision of a pro- tected environment for the embryo during its early development by ovoviviparity or viviparity and large adult size are also closely associated with the urea retention habitus for reasons elab- orated below (see also Griffith and Thomson 1973). The presence of sulfonated uronate resi- dues in the cartilage appears to be related to uremia since the feature is absent from fresh- water elasmobranchs with low blood urea (Ma- thews 1962; Thorson et al. 1967; Griffith et al. 1973) and elasmobranch high sulfate cartilage chondroitin sulfate protein does not dissociate in the presence of very high urea concentrations as do the cartilage proteins of other vertebrates (Mathews 1967). It is possible that the appear- ance of an ion-secreting organ that empties into the posterior part of the gut is also closely tied in with urea retention although, admittedly, this is speculation with little hard data behind it. One might argue that the necessary reduction in the permeability of the gills to solutes which is a consequence of urea retention eliminates them as a major site for the active secretion of ions and, in the absence of such *‘preadaptations’’ as tear, salivary or sweat glands, which are avail- able to amniotes as ion secreting organs, a urea- retaining fish must secrete excess salt into the terminal part of the gut or urogenital system. 86 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, No. 134 The appearance of such ion secreting organs in plotosid catfish (Lennep 1968; Kowarsky 1973) and in the lungfish Protopterus (Lagios and McCosker 1977) are suggestive if not particular- ly enlightening. Although we have suggested some bases for the interdependence of many of the characters shared by the chondrichthians and Latimeria, it is obvious that other features such as pituitary morphology and pancreatic islet associations cannot readily be regarded as consequences of urea retention. It is apparent that the analysis of other characters provides little insight into whether urea retention is independently ac- quired or homologous in Latimeria and the elas- mobranchs, and we should instead analyze the mechanisms involved in urea retention and see whether they are homologous in the two groups. Analysis of Mechanisms Involved in Urea Re- tention and of its Concomitants We believe that there are three lines of evi- dence which suggest that urea retention in chon- drichthians and Latimeria was independently derived. One is the absence of renal reab- sorption of urea in Latimeria, the second is the utilization of different modes of internal fertil- ization in the two groups, and the third is the independent appearance of urea retention as an osmoregulatory device in vertebrate groups oth- er than Latimeria and the chondrichthians. We have already considered the difference between Latimeria and the elasmobranchs in kidney function. In elasmobranchs up to 95% of the filtered urea is reabsorbed and U/P ratios under 0.30 are usual (reviews by Schmidt-Niel- sen 1972; Pang et al. 1977). This renal conser- vation of urea is of such obvious selective ad- vantage to a urea retaining fish that we were greatly surprised when we found that Latimeria lacked the ability to reabsorb urea. Urea reab- sorption is such a ubiquitous feature in chon- drichthians (including the holocephalans with a long separate evolutionary history, Read 1971) and involves such complex morphological (Boy- lan 1972; Stolte 1977) and physiological (Schmidt- Nielsen 1972; Hays et al. 1977) adaptations that it is difficult for us to conceive of it developing under conditions other than a severe challenge such as the initial movement of the ancestor of the elasmobranchs and holocephalans from fresh water to the marine environment with the concomitant development of urea retention. We regard the absence of renal urea reabsorption in Latimeria as evidence that it was never present in coelacanths but rather that they adopted dif- ferent mechanisms for conserving urea and evolved the urea retention habitus independent- ly from the chondnrichthians. The second point of evidence, the mechanism for internal fertilization, involves somewhat complex reasoning but is an equally strong ar- gument for the independent acquisition of urea retention. We have already suggested that inter- nal fertilization and an ovoviviparous or vivip- arous reproductive mode are obligatorily linked with urea retention which in fishes is a specific adaptation to osmoregulation in marine habitats, but some discussion may be necessary before this is obvious. The converse argument that urea retention arose secondarily to the evolution of a protected fetal environment as a means of dealing with the toxic buildup of ammonium in an analogous manner to the amniotes would seem to be discredited by the elimination of urea retention in freshwater stingrays without a change in their reproductive mode. The loss of urea in elasmobranchs occurs principally across the branchial and other sur- face epithelia (Smith 1931b; Payan et al. 1971, 1973; Boylan 1967; Chan and Wong 1977); whereas its synthesis takes place within the liver which, of course, represents a portion of the body volume. In very small animals such as ear- ly embryos the surface area will be very large relative to the volume of the liver and urea syn- thesis will not be able to keep up with urea loss. For this reason, urea retention is not economical in very small embryos and we observe that all chondrichthians (Wourms 1977) either maintain their young internally to an advanced stage of development or secrete a case around the egg which acts as a barrier to urea loss [albeit not a complete one, see Read (1968), Price and Dai- ber (1967)]. Such developmental patterns re- quire internal fertilization; the intromittant or- gans being the claspers, modified pelvic fins, in all chondrichthians (Wourms 1977) and presum- ably being erectile caruncules surrounding the ‘cloaca’ in the coelacanth (Griffith and Thom- son 1973; Millot and Anthony 1960). If a pro- tected embryonic environment with internal fer- tilization is a necessary prerequisite for urea retention as we suspect that it is, then the dif- ference between coelacanths and chondrichthi- ans in the method of internal fertilization is evi- GRIFFITH & PANG: COELACANTH OSMOREGULATION dence that both the reproductive mode and urea retention were independently acquired in the two groups. A common ancestor of coelacanths and chondrichthians possessing urea retention but not internal fertilization would be a contra- diction to our argument and an intermediate state between claspers and erectile cloacal ca- runcules is difficult to imagine. The final argument for independent acquisi- tion both provides direct evidence that urea re- tention can arise more than once and also pro- vides some support for the validity of the two preceding arguments. Like Latimeria and the chondrichthians, the crab-eating frog, Rana cancrivora, builds up high levels of urea in its blood to maintain blood osmolarity near that of its marine environment (Gordon et al. 1961). All evidence is that the ranid frogs are of relatively recent origin (Estes and Reig 1973; Lynch 1973) and, like other amphibia, are originally of fresh- water habitat and lack the urea retention habitus per se, although some amphibians do possess the capacity to build up urea (Funkhouser and Goldstein 1973; Schoffeniels and Tercefs 1966; Schlisio et al. 1973; Jungreis 1974). Unquestion- ably, Rana cancrivora did acheive the urea re- tention mechanism independently of that of Lat- imeria and the chondrichthians. Like Latimeria, and unlike the chondrichthians, Rana cancriv- ora lacks the ability to reabsorb urea in the renal tubules (Schmidt-Neilsen and Lee 1962) but, in apparent contrast to Latimeria, it possesses the capacity to reabsorb urea across the urinary bladder (Schmidt-Nielsen and Lee 1962; Gordon and Tucker 1968; Balinsky et al. 1972; Hew et al. 1972; Dicker and Elliot 1973). Hence, al- though it has reached the same solution as the elasmobranchs and Latimeria in terms of blood urea levels, Rana cancrivora uses an entirely different mechanism from either of them to do so. The problem of the loss of urea in very small embryos also has a unique solution in Rana can- crivora (Gordon and Tucker 1965). Like most amphibians, Rana cancrivora lays a large num- ber of small, unprotected eggs but does breed in saline waters. Up to the point of metamorphosis the tadpoles osmoregulate like marine teleost fishes by being hyposmotic to the environment and do not build up urea. Although this solution to the problem is probably dictated by the nor- mal switchover from an ammonotelic to a ureo- telic excretion during metamorphosis (reviews by Campbell 1973; Goldstein 1972), we think it 87 also supports our view that urea retention, in- ternal fertilization and a protected embryonic milieu are obligatorily coupled in fishes where metamorphosis does not occur. Choice of Strategy: The Energetics of Urea Re- tention and Hyposmotic Regulation Having concluded that urea retention proba- bly arose at least three separate times in verte- brates when freshwater animals entered a ma- rine environment, the question naturally arises as to why these animals chose urea retention whereas all actinopterygian fishes and the petro- myzontid agnathans opted for an hyposmotic mechanism for osmoregulating when they en- tered the marine environment. First of all, we should consider the energy ex- penditure that might be involved in the different mechanisms for osmoregulating. The teleost method of osmoregulation, maintaining the blood hyposmotic to the environment, involves active, energy-requiring sodium transport at two sites; an inward transport of sodium in the gut which is coupled with the uptake of water to replace that lost across the epithelia, and an out- ward transport across the gills to get rid of the sodium that enters through the gut and also that which enters passively across the epithelial sur- faces (Smith 1929; reviews by Potts 1968; Conte 1969). Elasmobranchs are hyperosmotic or isos- motic and do not drink sea water or actively move sodium across the gut, but they must ex- pend energy synthesizing urea to replace that lost across epithelial surfaces and also must ex- pend energy in eliminating excess sodium that enters passively by secreting it through the rec- tal gland or across the gills (review by Pang et al. 1977). Very crude estimates of the osmo- regulatory energy expenditure in terms of ATP which is hydrolyzed both in active sodium trans- port (ca. 0.4 mM ATP/mM Na’*) and in urea syn- thesis through the ornithine urea cycle (the equivalent of 4 mM ATP/mM urea) may be made for the elasmobranch Poroderma africanum based on the data of Haywood (1974, 1975b) and on representative teleosts based on the summary of Maetz and Bornancin (1975). In Poroderma, urea turnover is 0.12 mM/kg-hr and sodium ex- change 1s 0.185 mM/kg-hr of which about 0.050 mM/kg-hr may be via rectal gland secretion. Somewhat higher urea and sodium exchanges have been reported for other elasmobranchs (Burger and Tosteson 1966; Chan et al. 1967; 88 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, No. 134 Carrier and Evans 1972; Goldstein and Forster 1971; Payan and Maetz 1971). In teleosts bran- chial sodium excretion ranges from 3.8 mEq/Kg- hr in Salmo gairdneri to 78 in Mugil captio with 26 mEq/Kg-hr being a more-or-less typical value for marine species. Thus, the total energy ex- pended on osmoregulation may be estimated for our ‘“‘typical’’ teleost at 10.4 mM ATP/Kg-hr based solely on sodium excretion and for Po- roderma at 0.55 mM ATP/Kg-hr by adding the 0.48 mM used for urea synthesis and the 0.07 mM used for overall sodium excretion. From these very rough calculations it would seem that the elasmobranch method of osmoregulating is more economical although, as we will point out, many other factors will determine whether urea retention is both feasible and economical. Factors Predisposing Fishes Towards the De- velopment of Urea Retention Among the factors influencing the develop- ment of urea retention, we can suggest a number of predisposing features (prerequisites or pre- adaptations), four features which must be de- veloped simultaneously with urea retention but are not necessarily prerequisites, and one con- cept that may act to prevent the ready switch- over from a hyposmotic mechanism to urea re- tention and vice versa. The first prerequisite for urea retention is the presence of a complete ornithine-urea cycle. Most authors now agree that a complete O-U cycle was found in the ancestral gnathostome (see reviews by Campbell 1973; Goldstein 1972). However, it is absent in living agnathans (Read 1968, 1970, 1975) and, probably secondarily, is low or incomplete in many teleosts (Huggins et al. 1969; Wilson 1973; Gregory 1977). Groups lacking a complete O-U cycle may be prevented a priori from developing urea retention unless alternative pathways for urea synthesis, as through the metabolism of purines, can be used. A second predisposing feature is large size. Since very small fish have large surface/volume ratios, the loss of urea across the epithelia will exceed the capacity of the liver to synthesize it or, at any rate, will make the mechanism of urea retention uneconomical. Hyposmotic regulation will be no more expensive for small fish than for large fish since both sodium intake and sodium extrusion will take place across the surface ep- ithelia. A related predisposing feature, at least in fishes which adopt urea retention as a per- manent habitus, is internal fertilization since the early embryonic stages would be unable to syn- thesize enough urea to replace that lost across their relatively large surface areas. Another fac- tor which might influence the development of urea retention is the life history of the animal and two aspects of this are worth considering; activity patterns and the exposure to changing salinities. A very active fish requires a large gill area and a close contact between the branchial blood supply and the environmental water for the rapid exchange of gases (review by Hughes and Morgan 1973). It follows that these adap- tations would interfere with the development of a urea impermeable epithelium. Hence, it would be more likely for a sluggish than for an active fish to develop urea retention, particularly with- in smaller body sizes. We think it is something more than a coincidence that the more active elasmobranchs such as the lamnid sharks are in- variably large both at birth and as adults where- as sluggish elasmobranchs such as the batoids may be small in size. Air breathing, as in adult Rana cancrivora, might be a helpful preadap- tation to urea retention. We have earlier sug- gested (Griffith et al. 1973) that the teleostean method of osmoregulating might be more effi- cient for survival in conditions of rapidly chang- ing salinity since the high osmolarity inherent in urea retention becomes increasingly disadvan- tageous as salinities drop below sea water levels and in elasmobranchs physiological modifica- tions of blood solute levels are both slow and incomplete (Smith 193la; Price and Creaser 1967; Urist 1962; Thorson 1967; Thorson et al. 1973). Although there are some euryhaline elas- mobranchs (Smith 1936; Boeseman 1964; Thor- son and Watson 1975; Thorson 1976), these are exceptions and most euryhaline fishes are te- leosts. The Evolution of the Process of Urea Retention Other than the exceptional occurrence of urea retention in estivating lungfishes (Smith 1930b), we cannot conceive of any evolutionary path- way for the development of urea retention in fishes other than the movement from fresh water into the marine environment. The very fact that urea retaining fishes have blood ion levels lower than the environment is incontestible evidence of their freshwater ancestry. Virtually all marine invertebrates and certainly all that are primary marine groups, have extracellular fluid electro- GRIFFITH & PANG: COELACANTH OSMOREGULATION lyte levels that differ only in minor points from sea water and this holds true even for such high- ly organized groups as the crustaceans and cephalopod molluscs. In going from a freshwater to a marine en- vironment, a fish which becomes a urea retainer must undergo several modifications of his basic physiology as well as possess the proper com- bination of prerequisites for urea retention as described in the preceding section. Specifically, if not already present, there must be (1) devel- opment of surface epithelia impermeable to urea, (2) modification of renal function so that urea is conserved in the urine, (3) elevation of activities of enzymes of the O-U cycle to pro- duce sufficient amounts of urea, and (4) devel- opment of a tissue tolerance to elevated urea levels. We should point out that tissue tolerance to high urea is probably not a serious impedi- ment to the development of urea retention: urea levels comparable to those in the tissues of elas- mobranchs are not invariably toxic to teleost fishes (Hugoneng and Florence 1921; Rasmus- sen and Rasmussen 1977; Griffith et al., in press) although biochemical and osmoregulatory prob- lems often ensue. The Non-interconvertibility of Hyposmotic Reg- ulation and Urea Retention Although there is no a priori reason to believe that if is impossible for a marine teleost to de- velop urea retention or for a marine elasmo- branch to relinquish urea retention and become a hyposmotic regulator, we know empirically that this has not occurred. With the incredible diversity of bony fishes, one might think that unless there were some strong opposing force, that certain large, sluggish teleosts with internal fertilization (e.g. the zoarcids or brotulids) de- velop urea retention and one could also imagine that there must be some ecological niche for a small active shark for which urea retention might be an energetic disadvantage. We are con- vinced that the reason these switchovers have not occurred is a phenomenon we might desig- nate as evolutionary channelization. Simply ex- pressed, this means that once an evolutionary line has adopted a complex solution to an en- vironmental problem and has successfully adapted to it, that line must retain that solution and all concomitant features that go with it until a new and different environmental challenge presents itself that would make the original so- 89 lution disadvantageous. In addition to urea re- tention, we might suggest that other physiolog- ical adaptations such as low blood electrolytes in vertebrates and homeothermy in birds and mammals are good examples of the operation of the phenomenon. The only environmental change that has re- sulted in dechannelization of urea retention is the return to fresh water of such groups as the Amazon River rays of the family Potamotrygon- idae. Not only are blood urea levels low (Thor- son et al. 1967; Junqueira et al. 1968; Griffith et al. 1973) but enzymes of the ornithine urea cycle have declined (Goldstein et al. 1971; Gerst and Thorson 1977), the rectal gland has become non- functional (Griffith et al. 1973; Gerst and Thor- son 1977), renal reabsorption of urea is absent (Gerst and Thorson 1977), the cartilage has lost the extra sulfonation characteristic of elasmo- branchs (Mathews 1962) and the ability to tol- erate sea water is no longer present (Thorson 1970; Griffith et al. 1973; Gerst and Thorson 1977). It is interesting to speculate about the mechanisms that might be adopted were the po- tamotrygonids ever to become euryhaline again and reinvade the marine environment. SUMMARY The coelacanth, Latimeria chalumnae, dis- plays an osmoregulatory strategy similar in many respects to that of the elasmobranch and chimaerid fishes. Total serum osmolarity ap- proaches that of the marine environment, while concentrations of most major ions are consid- erably lower, the bulk of the remaining solutes consisting of urea and trimethylamine oxide. As do elasmobranchs, the coelacanth possesses a rectal gland of presumptive ion secreting func- tion. However, in contrast to the elasmobranchs and holocephalans, Latimeria lacks the ability to reabsorb urea across the renal tubules and also appears to have a serum osmolarity some- what below that of the environment, putting it in negative water balance. Based upon three lines of evidence, we sug- gest that the urea retention habitus in Latimeria and the cartilaginous fishes was acquired inde- pendently. (1) The absence of the characteristic chondrichthian renal reabsorption of urea in the coelacanth indicates that the urea retention hab- itus is only superficially similar in the two lin- eages and is not homologous. (2) Fishes that use urea retention as an osmoregulatory strategy 90 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, No. 134 must make some provision to prevent excessive loss of urea in small embryos whose large sur- face/volume ratios preclude a free existence. Although coelacanths, like many sharks, main- tain the young internally to an advanced state of development, the mechanisms for internal fer- tilization are radically different indicating that the reproductive mode, and by inference urea retention, were independently derived. (3) The euryhaline amphibian, Rana cancrivora, builds up urea like Latimeria and the chondrichthians, but both the mechanisms of urea retention in the urinary system (bladder but not tubular reab- sorption) and the reproductive strategy (larvae osmoregulate like marine teleosts) differ from both lineages, demonstrating both that urea re- tention is not a unique strategy of the coelacanth and cartilaginous fishes and that the other cri- teria we used as evidence for independent de- velopment of the mechanism in the two groups of fishes are valid. A number of features predispose fishes to- wards the urea retention habitus rather than hy- posmotic regulation when they move from fresh water to marine environments. Among them are: (1) the presence of a complete ornithine-urea cycle, (2) internal fertilization, (3) large size, (4) sluggish life style, and (5) invasion of stable ma- rine environments rather than the variable salin- ity of estuaries. Concomitant with the develop- ment of high blood urea must be: (1) urea impermeable epithelia, (2) some mechanism for reduction of urea loss through the urine, and (3) tolerance of elevated blood and tissue urea levels which would seem to include a number of biochemical adaptations. We suggest that once either urea retention of hyposmotic regulation has been adopted as the strategy for osmoregu- lating in a marine environment, the alternative strategy is no longer feasible unless a major en- vironmental reversal occurs. LITERATURE CITED Artz, J. W. 1973. Comparative endocrinology and systemat- ics. Am. Zool. 13:933-936. BALINSKY, J. B., S. E. DicKER, AND A. B. ELLiot. 1972. The effects of long-term adaptation to different levels of salinity on urea synthesis and tissue amino acid concentra- tions in Rana cancrivora. Comp. Biochem. Physiol. 43B:71-82. BJERRING, H. C. 1973. Relationships of coelacanthiforms. Pages 179-205 in P. H. Greenwood, R. S. Miles, and C. Patterson, eds., Interrelationships of fishes. Academic Press, New York. 536 pp. BoESEMAN, M. 1964. Notes on the fishes of Western New Guinea III. The fresh water sharks of Jamoer lake. Zool. Nededeelingen 40:9-22. BONTING, S. 1966. Studies on sodium-potassium activated adenosinetriphosphatase XV. The rectal gland of the elas- mobranch. Comp. Biochem. Physiol. 7: 953-966. BoyLANn, J. W. 1967. Gill permeability in Squalus acanthias. Pages 197-206, in P. W. Gilbert, R. F. Mathewson, and D. P. Rall, eds., Sharks, skates and rays. Johns Hopkins Press, Baltimore. 624 pp. 1972. A model for passive urea reabsorption in the elasmobranch kidney. Comp. Biochem. Physiol. 42A:27— 30. Brown, G. W. AND S. G. Brown. 1967. Urea and its for- mation in coelacanth liver. Science 155:570-573. BRULL, L., AND Y. Cuypers. 1954. Quelques caracteris- tiques biologiques de Lophius piscatorius L. Arch. Int. Physiol. 62:70—75. , AND E. NizsBert. 1953. Blood and urine constituents of Lophius piscatorius. J. Mar. Biol. Assoc. U.K. 32:321- 328. BurGER, J. W. 1967. Problems in the electrolyte economy of the spiny dogfish, Squalus acanthias. Pages 177-185, in P. W. Gilbert, R. F. Mathewson, and D. P. Rall, eds., Sharks, skates, and rays. Johns Hopkins Press, Baltimore. 624 pp. , AND W. N. Hess. 1960. Function of the rectal gland in the spiny dogfish. Science 131:670-671. CAMPBELL, J. W. 1973. Nitrogen excretion. Pages 279-316 in C. L. Prosser, ed., Comparative animal physiology. W. B. Saunders Co., Philadelphia. 966 pp. CARRIER, J. C., AND D. H. Evans. 1972. Ion, water and urea turnover rates in the nurse shark, Ginglymostoma cirratum. Comp. Biochem. Physiol. 41A:761-764. CHAN, D. K. O., J. G. PHILLIPS, AND I. CHESTER JONES. 1967. Studies on electrolyte changes in the lip-shark Hemi- scyllium plagiosum (Bennett), with special reference to the hormonal influence on the rectal gland. Comp. Biochem. Physiol. 23: 185-198. , AND T. M. WonG. 1977. Physiological adjustments to dilution of the external medium in the lip-shark Hemi- scyllium plagiosum (Bennett) II. Branchial, renal and rectal gland function. J. Exp. Zool. 200:85—96. CHAVIN, W. 1976. The thyroid of the sarcopterygian fishes (Dipnoi and Crossopterygii) and the origin of the tetrapod thyroid. Gen. Comp. Endocrinol. 30:142-155. CHEW, M. M., A. B. ELLIoTT, AND H. Y. Wona. 1972. Permeability of urinary bladder of Rana cancrivora to urea in the presence of oxytocin. J. Physiol. (London) 223:757- 22. Conte, F. P. 1969. Salt secretion. Pages 241-292, in W.S. Hoar and D. J. Randall, eds., Fish physiology, Vol. 1. Ac- ademic Press, New York. 465 pp. DeBEER, G. H. 1954. Archaeopteryx and evolution. Adv. Sci. 42:1-11. DEETJEN, P., D. ANTKOWIAK, AND J. W. BoYLAN. 1972. Urea reabsorption by the skate nephron, micropuncture of collecting ducts in Raja erinacea. Bull. M. Desert Is]. Biol. Lab. 12:28-29. Dicker, S. E., AND A. B. Ettior. 1973. Neurohypophysial hormones and homeostasis in the crab-eating frog, Rana cancrivora. Horm. Res. 4:224—260. Eppte, A., AND J. E. BRINN, JR. 1975. Islet histophysiology: evolutionary correlations. Gen. Comp. Endocrinol. 27:320— 349. GRIFFITH & PANG: COELACANTH OSMOREGULATION Estes, R., AND O. A. REIG. 1973. The early fossil record of frogs: a review of the evidence. Pages 11-64 in J. L. Vial, ed., Evolutionary biology of the Anurans. University of Missouri Press, Columbia, Missouri. 470 pp. FANGE, R., AND K. FUGELLI. 1963. The rectal salt gland of elasmobranchs, and osmoregulation in chimaeroid fishes. Sarsia 19:27-34. Forster, R. P., L. GOLDSTEIN, AND J. K. Rosen. 1972. Intrarenal control of urea reabsorption by renal tubules of the marine elasmobranch, Squalus acanthias. Comp. Bio- chem. Physiol. 42A:3-12. FUNKHOUSER, D., AND L. GOLDSTEIN. 1973. Urea response to pure osmotic stress in the aquatic toad Xenopus laevis. Am. J. Physiol. 224:524-529. GARDINER, B. G. 1973. Interrelationships of teleostomes. Pages 10S—131 in P. H. Greenwood, R. S. Miles, and C. Patterson, eds., Interrelationships of fishes. Academic Press, New York. 535 pp. Gerst, J. B., AND T. B. THoRSON. 1977. Effects of saline acclimation on plasma electrolytes, urea excretion, and he- patic urea biosynthesis in a freshwater stingray Potamotry- gon sp. Garman, 1877. Comp. Biochem. Physiol. 56A:87— 93% GOLDSTEIN, L. 1972. Adaptation of urea metabolism in aquatic vertebrates. Pages 55-77 in J. W. Campbell and L. Goldstein, eds., Nitrogen metabolism and the environment. Academic Press, New York. 318 pp. . AND R. P. Forster. 1971. Urea biosynthesis and excretion in freshwater and marine elasmobranchs. Comp. Biochem. Physiol. 39B:415-421. , S. Harley-DeWitt, and R. P. Forster. 1973. Activities of ornithine-urea cycle enzymes and of trimethylamine ox- idase in the coelacanth, Latimeria chalumnae. Comp. Bio- chem. Physiol. 44B:357-362. Gorpon, M. S., K. SCHMIDT-NIELSEN, AND H. M. KELLY. 1961. Osmotic regulation in the crab-eating frog (Rana can- crivora). J. Exp. Biol. 38:659-678. , AND V. E. TUCKER. 1965. Osmotic regulation in tad- poles of the crab-eating frog (Rana cancrivora). J. Exp. Biol. 42:437-44S. , AND . 1968. Further observations on the phys- iology of salinity adaptation in the crab-eating frog (Rana cancrivora). J. Exp. Biol. 49:185-193. Grecory, R. B. 1977. Synthesis and total excretion of waste nitrogen by fish of the Periophthalmus (Mudskipper) and Scartelaas families. Comp. Biochem. Physiol. 57A:33—36. GRIFFITH, R. W., AND C. J. BuRDICK. 1976. Sodium-potas- sium activated adenosine triphosphatase in coelacanth tis- sues: high activity in rectal gland. Comp. Biochem. Physiol. 54B:557-559. , M. B. MatHews, B. L. UMMINGeER, B. F. GRANT, P. K. T. PANG, K. S. THOMSON, AND G. E. PICKFORD. 1975. Composition of fluid from the notochordal canal of the coelacanth, Latimeria chalumnae. J. Exp. Zool. 192: 165-172. , P. K. T. PANG, AND L. A. BENEDETTO. 1978. Urea tolerance in the killifish, Fundulus heteroclitus. Comp. Bio- chem. Physiol., in press. 5 , A. K. SRIVASTAVA, AND G. E. PICKFORD. 1973. Serum composition of freshwater stingrays (Pota- motrygonidae) adapted to fresh and dilute sea water. Biol. Bull. 144:304—320. , AND K. S. THOMSON. 1973. Latimeria chalumnae: Reproduction and conservation. Nature 242:617-618. 9 , B. L. UMMINGER, B. F. GRANT, P. K. T. PANG, L. GOLDSTEIN, AND G. E. PickForD. 1976. Composition of bladder urine of the coelacanth, Latimeria chalumnae. J. Exp. Zool. 196:371-380. ; ; , AND G. E. PICKFORD. 1974. Serum composition of the coelacanth, Latimeria chalumnae Smith. J. Exp. Zool. 187:87-102. GRONINGER, H. S. 1959. Occurrence and significance of tri- methylamine oxide in marine animals. U.S. Fish Wildl. Serv. Spec. Sci. Rep., No. 333. Hays, R. M., S. D. Levine, J.D. Myers, H. O. HEINEMAN, M. A. KAPLAN, N. FRANKI, AND H. BERLINER. 1977. Urea transport in the dogfish kidney. J. Exp. Zool. 199:309-315. HaAywoop, G. P. 1974. The exchangeable ionic space, and salinity effects upon ion, water, and urea turnover rates in the dogfish Poroderma africanum. Mar. Biol. 26:69-75. 197S5a. A preliminary investigation into the roles played by the rectal gland and kidneys in the osmoregula- tion of the striped dogfish Poroderma africanum. J. Exp. Zool. 193:167-175. 1975b. Indications of sodium, chloride, and water exchange across the gills of the striped dogfish, Poroderma africanum. Mar. Biol. 29:267-276. HICKMAN, C. P., AND B. F. TRUMP. 1969. The kidney. Pages 91-239 in W. S. Hoar and D. J. Randall, eds., Fish physi- ology, vol. 1. Academic Press, New York. 465 pp. Hoimes, W. N., AND E. M. DONALDSON. 1969. The body compartments and the distribution of electrolytes. Pages 1- 89 in W. S. Hoar and D. J. Randall, eds., Fish physiology, vol. 1. Academic Press, New York. 465 pp. HuaGains, A. K., G. SkUTCH, AND E. BALDWIN. 1969. Or- nithine-urea cycle enzymes in teleost fishes. Comp. Bio- chem. Physiol. 28:587—602. HuGues, G. M. 1972. Gills of a living coelacanth, Latimeria chalumnae. Experientia 28:1301—1302. 1976. On the respiration of Latimeria chalumnae. Zool. J. Linn. Soc., London 59:195—208. . AND M. MorGANn. 1973. The structure of fish gills in relation to their respiratory function. Biol. Rev. 48:419-475. HUGONENQ, L., AND G. FLORENCE. 1921. Experience du cours se reapportant a l’azotemie. Bull. Soc. Chim. Biol., Paris 3:174-175. JARVIK, E. 1968a. Aspects of vertebrate phylogeny. Pages 497-523 in T. Orvig, ed., Current problems of lower ver- tebrate phylogeny. Almqvist and Wiksell, Stockholm. 539 Pp. 1968b. The systematic position of the Dipnoi. Pages 223-245 in T. Orvig, ed., Current problems of lower ver- tebrate phylogeny. Almqvist and Wiksell, Stockholm. 539 pp. JuNGREIS, A. M. 1974. Seasonal effects of dehydration in air on urea production in the frog, Rana pipiens. Comp. Bio- chem. Physiol. 47A:39-SO. JUNQUEIRA, L. C. V., G. HoxTerR, AND D. ZAGO. 1968. Ob- servations on the biochemistry of fresh water rays and dol- phin blood serum. Rev. Bras. Pesqui. Med. Biol. 1:225-226. Kowarsky, J. 1973. Extra-branchial pathways of salt ex- change in a teleost fish. Comp. Biochem. Physiol 46A:477— 486. Laaios, M. D. cell and renovascular morphology of the coelacanth, Larti- meria chalumnae Smith (Crossopterygii) compared with that of other fishes. Gen. Comp. Endocrinol. 22:296—307. 1974. Granular epithelioid (juxtaglomerular) 92 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, No. 134 . 1975. The pituitary gland of the coelacanth Latimeria chalumnae Smith. Gen. Comp. Endocrinol. 25:126-146. . 1979. The Coelacanth and the Chondrichthyes as sis- ter groups. A review of shared apomorph characters and a cladistic analysis and reinterpretation. This volume. , AND J. E. McCosker. 1977. A cloacal excretory gland in the lungfish Protopterus. Copeia 1977:176—178. Lemire, R. 1976. Caractéristiques ultrastructural des cellules acineuses de la gland post-anale du coelacanth (Latimeria chalumnae Smith). C. R. Acad. Sci., Ser D, 282:641-644. LeENNEP, E. W. 1968. Electromicroscopic histochemical studies of salt excreting glands in elasmobranchs and marine catfish. J. Ultrastruct. Res. 25:94-108. Locket, N. A., AND R. W. GRIFFITH. on a living coelacanth. Nature 237:175. LOMBARDINI, J. B., P. K. T. PANG, AND R. W. GRIFFITH. 1978. Tissue amino acid and taurine levels in some aquatic animals. This volume. Lutz, P. L., AND J. D. RoBeERTSON. 1971. Osmotic constit- uents of the coelacanth Latimeria chalumnae Smith. Biol. Bull. 41:553—560. Lyncu, J. B. 1973. The transition from archaic to advanced frogs. Pages 133-182 in J. F. Vial, ed., The evolutionary biology of the anurans. University of Missouri Press, Co- lumbia, Missouri. 470 pp. Maetz, J., AND M. BORNANCIN. 1975. Biochemical and bio- physical aspects of salt excretion by chloride cells in te- leosts. Fortsch. Zool. 23:322-362. MATHEWS, M. B. 1962. Sodium chondroitin sulfate-protein complexes of cartilage III. Preparations from shark. Bio- chim. Biophys. Acta 58:92-101. . 1966. The molecular evolution of cartilage. Clin. Or- thop. 48:267-283. 1967. Macromolecular evolution of connective tis- sues. Biol. Rev. 42:499_-SS51. McCosker, J. 1979. Inferred natural history of the living Coelacanth. This volume. Mies, R. 1973. Relationships of acanthodians. Pages 63-99 in P. H. Greenwood, R. S. Miles, and C. Patterson, eds., Interrelationships of fishes. Academic Press, New York. 535 pp. . 1975. The relationships of the Dipnoi. Colloq. Int. C. N. R. S. 218:133-148. MILLoT, J., AND J. ANTHONY. 1960. Le cloaque chez les coelacanthes. Bull. Mus. Nat. Hist. Nat., Paris 32:287—289. , AND . 1972. La glande post-anale de Latimer- ia. Ann. Sci. Nat. Zool., Ser. 14, Paris 12:305-318. , AND . 1973. L’appareil excréteur de Latimeria chalumnae Smith (Poisson, Coelacanthidé). Ann. Sci. Nat. Zool., Ser. 15, Paris 12:293-328. Moss, M. 1977. Skeletal tissues in sharks. Am. Zool. 17:335— 342. Moy-THomas, J. A. 1971. Paleozoic fishes. Extensively re- vised by R. S. Miles. W. B. Saunders, Philadelphia. 259 pp. Munz, F. W., AND W. N. MACFARLAND. 1964. Regulatory function of a primitive vertebrate kidney. Comp. Biochem. Physiol. 13:381—400. NISHIMURA, H., M. OGAWA, AND W. H. Sawyer. 1973. Renin-angiotensin system in primitive bony fishes and a hol- ocephalian. Am. J. Physiol. 224:950-956. Norris, E. R., AND G. J. BeNorr. 1945. Studies on tri- methylene oxide II. The origin of trimethylene oxide in young salmon. J. Biol. Chem. 158:439-442. 1972. Observations PANG, P. K. T., R. W. GRIFFITH, AND J. W. Arz. 1977. Osmoregulation in elasmobranchs. Am. Zool. 17:365—377. PAYAN, P., AND J. MAETZ. 1971. Balance hydrique chez les elasmobranches: Arguments en faveur d’un controle endo- crinien. Gen. Comp. Endocrinol. 16:535—554. , AND . 1973. Branchial sodium transport mech- anisms in Scyliorhinus canicula: Evidence for Na*t/NH,* and Na*/H* exchanges and for a role of carbonic anhydrase. J. Exp. Biol. 58:487—502. PicKFoRD, G. E., AND F. B. GRANT. 1967. Serum osmolarity in the coelacanth, Latimeria chalumnae: urea retention and ion regulation. Science 155:568—5S70. ; , AND B. L. UMMINGER. 1969. Studies on the blood serum of the euryhaline cyprinodont fish, Fundulus heteroclitus, adapted to fresh or to salt water. Trans. Conn. Acad. Arts Sci. 43:25-70. Potts, W. T. W. 1968. Osmotic and ionic regulation. Annu. Rev. Physiol. 30:73-104. Price, K. S., AND E. P. CREASER. 1967. Fluctuations in two osmoregulatory components, urea and sodium chloride, of the clearnose skate Raja eglanteria Bosc. 1802. I. Upon laboratory modification of external salinities. Comp. Bio- chem. Physiol. 23:65—76. , AND F.C. DAIBER. 1967. Osmotic environments dur- ing development of dogfish, Mustelus canis (Mitchell) and Squalus acanthias, and some comparison with skates and rays. Physiol. Zool. 40:248—260. RASMUSSEN, L., AND R. A. RASMUSSEN. 1977. Exogenous 4C-urea distribution in selected marine fishes—especially two species of salmon, and resultant alteration in sera and gill enzyme levels. Comp. Biochem. Physiol. 56A:403—409. READ, L. J. 1968a. A study of ammonia and urea production and excretion in the freshwater-adapted form of Pacific lam- prey Entosphenus tridentatus. Comp. Biochem. Physiol. 26:455—466. . 1968b. Urea and trimethylamine oxide levels in elas- mobranch embryos. Biol. Bull. 135:537—547. 1970. Nitrogen metabolism in chondrichthian and agnathan fishes. Pages 23-34, in B. Schmidt-Nielsen, ed., Urea and the kidney. Excerpta Medica Foundation, Am- sterdam. 495 pp. 1971. Chemical constituents of the body fluids and urine of the holocephalian Hydrolagus colliei. Comp. Bio- chem. Physiol. 39A:185—192. . 1975. Absence of ureogenic pathways in liver of the hagfish Bdellostoma cirrhatum. Comp. Biochem. Physiol. 51B:139-141. ROBERTSON, J. D. 1966. Osmotic constituents of the blood plasma and parietal muscle of Myxine glutinosa L. Pages 631-644 in H. Barnes, ed., Some contemporary studies in marine science. Allen and Unwin, London. 716 pp. 1975. Osmotic constituents of the blood plasma and parietal muscle of Squalus acanthias L. Biol. Bull. 48:303- 319. 1976. Chemical composition of the body fluids and muscle of the hagfish Myxine glutinosa and the rabbit-fish Chimaera monstrosa. J. Zool. London 178:261—277. Romer, A. S. 1966. Vertebrate paleontology, 3rd ed. The University of Chicago Press, Chicago. 468 pp. SCHAEFFER, B. 1968. The origin and basic radiation of the Osteichthyes. Pages 207-221 in T. Orvig, ed., Current prob- lems of lower vertebrate phylogeny. Almqvist and Wiksell, Stockholm. 539 pp. GRIFFITH & PANG: COELACANTH OSMOREGULATION ScHLisio, W., K. JUERSs, AND L. SPANNHOF. 1973. Osmo- und ionenregulation von Xenopus laevis Daud nach adap- tation in verschiedenen osmotisch wirksamen Lésungen I. Toleranz und Wasserhaushalt. Zool. Jahrb. Abt. Allg. Zool. Physiol. Tiere 7:275—290. SCHMIDT-NIELSEN, B. 1972. Mechanisms of urea excretion by the vertebrate kidney. Pages 79-103, in J. W. Campbell and L. Goldstein, eds., Nitrogen metabolism and the envi- ronment. Academic Press. 318 pp. , B. TRUNIGER, AND L. RABINOWITZ. 1972. Sodium- linked urea transport by the renal tubule of the spiny dog- fish, Squalus acanthias. Comp. Biochem. Physiol. 42A:13- D5): SCHMIDT-NIELSEN, K., AND P. LEE. 1962. Kidney function in the crab-eating frog (Rana cancrivora). J. Exp. Biol. 39: 167-177. SCHOFFENIELS, E., AND R. R. TERCEFS. 1966. L’os- moregulation chez les batraciens. Ann. Soc. R. Zool. Belg. 96:23-39. SmitH, C. L., C. S. RAND, B. SCHAEFFER, AND J. W. ATz. 1975. Latimeria, the living coelacanth, is ovoviviparous. Science 190:1105—-1106. SmitH, H. W. 1930a. The absorption and excretion of water and salts by marine teleosts. Am. J. Physiol. 93:480—S0S. 1930b. Metabolism of the lungfish, Protopterus aethiopicus. J. Biol. Chem. 86:97—130. 193la. The absorption and excretion of water and salts by the elasmobranch fishes I. Freshwater elasmo- branchs. Am. J. Physiol. 98:279-295. 1931b. The absorption and excretion of water and salts by the elasmobranch fishes II. Marine elasmobranchs. Am. J. Physiol. 98:296-310. . 1936. The retention and physiological role of urea in the Elasmobranchi. Biol. Rev. 11:49-82. 1953. From fish to philosspher. Little, Brown and Co., Boston. 264 pp. STAHL, B. J. 1967. Morphology and relationships of the Hol- ocephali with special reference to the venous system. Bull. Mus. Comp. Zool. Harv. Univ. 135:141-195. 93 STOLTE, H., R. G. GALAskE, G. M. EISENBACH, C. LEK- HENE, B. SCHMIDT-NIELSEN, AND J. W. BoYLAN. 1977. Renal tubule ion transport and collecting duct function in the elasmobranch little skate Raja erinacea. J. Exp. Zool. 199:403-410. THOMSON, K. S. 1969. The environment and distribution of paleozoic sarcopterygian fishes. Am. J. Sci. 267:457-464. 1971. The adaptation and evolution of early fishes. Q. Rev. Biol. 46: 139-166. THORSON, T. B. 1967. Osmoregulation in fresh water elas- mobranchs. Pages 265-270 in P. W. Gilbert, R. F. Mathew- son, and D. P. Rall, eds., Sharks, skates, and rays. The Johns Hopkins Press, Baltimore. 624 pp. . 1970. Freshwater stingrays, Potamotrygon spp.: fail- ure to concentrate urea when exposed to saline medium. Life Sci. 9(Part 2):893-900. 1976. The status of the lake Nicaragua shark: an updated appraisal. Pages 561-574, in T. B. Thorson, ed., Investigations of the Ichthyofauna of Nicaraguan Lakes. University of Nebraska Press, Lincoln. 663 pp. , C. M. Cowan, AND D. E. Watson. 1967. Pota- motrygon spp.: Elasmobranchs with low urea content. Sci- ence 158:375-377. 5 , AND 1973. Body fluid solutes of ju- veniles and adults of the euryhaline bull shark Carcharhinus leucas from freshwater and saline environments. Physiol. Zool. 46:29-42. , AND D. E. Watson. 1975. Reassignment of the Af- rican freshwater stingray, Potamotrygon garouaensis, to the genus Dasyatis, on physiologic and morphologic grounds. Copeia 1975:701-712. Urist, M. R. 1962. Calcium and other ions in blood and skeleton of Nicaraguan freshwater shark. Science 137:984— 986. WILson, R. P. 1973. Nitrogen metabolism in channel catfish II. Evidence for an apparent incomplete ornithine-urea cycle. Comp. Biochem. Physiol. 46B:625-634. WourMms, J. P. 1977. Reproduction and development in chon- drichthyan fishes. Am. Zool. 17:379-410. No. 134, 17 pages OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES The Biology and Physiology of the Living Coelacanth SOME BIOCHEMICAL PARAMETERS IN THE COELACANTH: VENTRICULAR AND NOTOCHORDAL FLUIDS By Lois E. Rasmussen Oregon Graduate Center, 19600 NW Walker Road, Beaverton, Oregon 97005 ABSTRACT The coelacanth, Latimeria chalumnae, has a distinct evolutionary appearance. Knowledge of its biochemical constituents may clarify this position and at the same time aid in the understanding of its physiological functioning. Therefore, utilizing a variety of microtechniques, a spectrum of biochemical parameters was measured in the notochordal fluid, and, for the first time, in the ventricular fluid of the coelacanth. These included inorganice and selected organics—urea, lipids, proteins, and enzymes, and involved both total assays and separations—electrophoretic and thin layer chro- matography. The results were repetitively reproducible despite conditions prior to assay. Although only one specimen was available for study, the results were precise, without appreciable replicate variation, allowing valid comparisons between previous coelacanth notochordal fluid and serum data and other marine fish brain fluids. These comparisons allow some speculation on the phylogenetic place of the coelacanth and its unique and/or unusual biochemical physiology. December 22, 1979 INTRODUCTION Certain, but not all, of the more primitive ma- rine, or partially marine, fishes are among the small coterie of animals which possess high (by mammalian standards) urea tissue and fluid levels. The investigation in these hyperuremic animals of two distinct urea functions, the well- documented one of osmoregulation, and a pos- 94 sible additional function in the maintenance of protein and/or enzyme stability seemed of rele- vance. Pursuing this goal, a spectrum of biochemical constituents was measured in the ventricular brain fluid (for the first time) and in notochordal fluid of the coelacanth, Latimeria chalumnae; the assay of urea was given special attention. RASMUSSEN: COELACANTH VENTRICULAR AND NOTOCHORDAL FLUIDS 95 Previous investigations established certain nor- mal urea ranges in a chimaera—Hydrolagus col- liei, several elasmobranchs and selected Isopon- dyli (Rasmussen and Rasmussen 1977), and included tracer experiments involving experi- mentally induced transitory hyperuremia (Ras- mussen 1971, 1974b; Rasmussen and Rasmussen 1977). The second purpose was a comparison of our data with data for another coelacanth notochor- dal fluid (Griffith et al. 1975) to give an indication of the range of variation between individual coelacanths, recognizing that the data compari- son was numerically limited as our values were obtained from only a single fish. Fishes are known to have recognizable individual varia- tions, including species variations, differences among individuals within a given species (Mus- telus canis, Rasmussen and Rasmussen 1967b), and even individual variations dependent on en- vironmental factors such as salinity and temper- ature (M. canis, Rasmussen and Rasmussen 1967a and 1967b), and/or physiological factors such as estivation states (Protopterus, Lock- wood 1963), reproductive states (salmon, Urist and Van de Putte 1967), and/or stress (salmon and sharks, Idler et al. 1959; Rasmussen and Rasmussen 1967b). Nevertheless the data are valuable in comparison to the previously re- ported notochordal fluid values (Griffith et al. 1975), and provide a comparison of these no- tochordal fluid values with the first time mea- surements in the ventricular brain fluid. METHODS Coelacanth #79 can be regarded as a “‘sec- ond-best’’ specimen (biochemically). Within hours after capture, the fish was frozen whole, and stored at —10 to (0)°C for 18 months. After transport to the United States the still frozen fish was thawed and then dissected (May 1975) at San Francisco (see Acknowledgments). The question of the degree of stress experienced dur- ing capture or the holding period was discussed by Dr. Lagios during the June 1977 AAAS meet- ing and is briefly discussed at the end of this paper. Subsequent tissue conditions were as fol- lows. After removal, the fluids were rapidly re- frozen. Unlike the procedures followed for other fish fluids no NADH was added (although sub- sequently a few additional unfrozen, refrozen aliquots did have this stabilizer). The material received was subdivided into aliquots at the time of the assay of the most labile substances. By the time of assay or actual separative analyses, the fluids were thawed 2-3 times. All of these separative analyses were on different aliquots from a single sample of both notochordal and ventricular fluids from the same coelacanth. Multiple assays were done on replicate samples, a factor to keep in mind during the data analysis. However, despite its age (3+ years) and these conditions, the excellence of the physiological condition of at least the ventricular fluid is evi- denced by the lack of visible hemolysis, the lack of very high Ca, Mg, or K levels, and reason- ableness of the assayed enzyme levels. Ex- tremely high levels might have been indicative of possible cell damage. Inorganics measured included the following: Total osmolarity by Precision Systems os- mometer, vapor pressure OSsmometer, and melt- ing point osomometer, Cl using Hg* plus the indicator diphenyl carbazone (Hawk et al. 1954), Na by Stone and Goldzieher (1949), and Ca by a modification of Yanagisawa (1955) and Schales and Schales (1941). Organics measured for their total values in- cluded the following lipids: Total /ipid after lipid extraction (Folch et al. 1957) determined by the dry weight of the lipid chloroform-methanol ex- tract (Zollner et al. 1966) and/or the colorimetric determination using sulfo-phospho-vanillin re- agent (Fring and Dunn 1970); total lipid phos- phorus by the method of Lowry and Tinsley 1974: and triglycerides assayed by the dienzyme method of Wakayama and Swanson 1977 with Calbiochem reagents (glycerolkinase, glycerol- 3-phosphate dehydrogenase). Cholesterol was assayed both by a slight modification of Peder- sen, 1973 CSF method and a precise tri-enzy- matic method utilizing cholesterol ester hydro- lase, cholesterol oxidase, and horse radish peroxidase (Witte et al. 1974). Urea was assayed by Archibald, 1945 method using phenyl 1,2 pro- panedione 2 oxime. Total glycolipids were as- sayed by method of Jatzkewitz and Mehl 1960. Proteins were measured by method of Lowry et al. 1951). Data involving more than two repli- cates were expressed as the mean + SE (num- ber). Differences where P was less than 0.05 were considered significant. Separation Methods included thin layer chro- matography (TLC) for lipids and several vari- eties of electrophoretic separation techniques for the proteins. Three types of proteins were 96 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, No. 134 examined—soluble proteins, lipoproteins, and glycoproteins, by 3—4 types of electrophoresis. Cellulose acetate electrophoresis (CAE), agar gel electrophoresis (AGE), starch gel electro- phoresis (SGE), and polyacrylamide gel electro- phoresis (PAGE) of at least 3 and up to 10 rep- licates of coelacanth notochordal and spinal fluids soluble proteins both ‘‘au naturale’’ and concentrated between 2-10 fold by pressure di- alysis in special collodion tubes were run. Both the visibly clear supernatant and the gently cen- trifuged more precipitate, fiber containing frac- tions were separated. The CAE technique uti- lized was the Microzone System (Beckman Bulletin #7033), SGE (Smithies 1959), AGE (Uriel 1964), 7% w/v PAGE (Pratt and Danger- field 1963). Stains for soluble proteins were Fast Green and Amido Black (Davis 1964). Lipoproteins were stained after AGE or PAGE with lipid crimson, with Sudan Black in ethanoldiol (McDonald and Ribeiro 1959), or by Sudan Black pre-incubated for |—2 hours before electrophoresis (Fring et al. 1971), as appropri- ate. Glycoproteins were separated by AGE (Uriel 1964) and stained by azure A producing a purple- red color due to an asymmetric dimethyl thioin (Uriel 1964), or the periodic acid (specifically oxidizing 1,2 glycol groups)—Schiff (less specif- ic, reacting with slow lipoproteins, ketones, and unsaturated groups) reaction (PAS) until a pink color appeared (Zacharius et al. 1969). This PAS technique demonstrated both glyco- and lipo- proteins. Groups of lipids were separated using thin lay- er chromatography (TLC). Lipids were removed from the lipoproteins by ethanol diethyl ether and acetone was used to separate glycolipids from phospholipids. The primary solvent system used was CCl, 85, MeOH 15, HAc 10, and H,O 4 v/v with a somewhat different solvent system used for specific cholesterol separations. Cho- lesterol was visualized on silicic acid coated plates, after separation by 50% H.SO, spray- stain, followed by heating (Jatzkewitz and Mehl 1960), or H,SO, with dichromate, or 2% FeCl.,. The highly specific spray-stain, diphenylamine, was used to resolve the glycolipids (Christie 1973; Freeman and West 1966). Several coelacanth tissues, especially the liv- er, have been analyzed for enzyme levels. The results of Solomon and Brown (1975) with fro- zen coelacanth liver, demonstrating high levels of the following enzymes: peptidase, diaphorase, acid phosphatase, esterase, LDH, and peroxi- dase (the latter four having residual activity after 8 years), were used as a guide toward choosing the enzymes assayed in this investigation. Ac- cordingly, two of these lactic dehydrogenase (LDH), and acid phosphatase (AcP), and a third enzyme—alpha-amylase were assayed. For each enzyme assay, 5 replicates per assay were used, and the assays carried out at least three times. The methods used were: amylase—(Cseska et al. 1959); AcP—(Modder 1973); LDH—(Amador et al. 1965). The rate reaction was employed be- cause isozymes of LDH from different species and tissues may have different affinities for py- ruvate (Hochachka 1965), and because of its su- perior precision. A unit of LDH activity equals an increase in absorbance at 340 my of 0.001 per minute per milliliter of serum or fluid at 25°C. Additional LDH assays were carried out at varying urea concentrations. RESULTS !. Total Osmolarity and Inorganics A background spectrum of serum data in rel- evant vertebrate species for total osmolarity and inorganic ions, especially chloride is listed in Table 1. For comparative purposes, values were averaged somewhat in Table 1. The total serum osmolarity mOsm/I of marine hagfishes (1,152), chimaeras (942), coelacanth (1,181) (Griffith et al. 1974), and elasmobranchs (973) was 2-3 fold that of lampreys (317), marine teleosts (500) (Schmidt-Nielsen 1976), and human serum (300). The sodium levels in the elasmobranchs and chi- maeras were noticeably higher than coelacanth and marine teleost levels (Table 1); specifically the chimaera value was 268, the elasmobranchs 255, but the coelacanth was 197 and the teleosts 180. Chlorides showed a similar species pattern. Mg coelacanth levels were of the elasmobranch magnitude, and Ca levels were similar to chi- maera and elasmobranch levels, but 2 fold higher than teleosts. In the nervous system associated fluids (Grif- fith et al. 1975 and coelacanth #79 Rasmussen) the total osmolarity was highest in the coel- acanth, slightly lower in elasmobranchs, notice- ably lower in chimaeras, and significantly lower in the teleosts. The total osmolarity was greater in the CSF than the serum, and the total os- molarity was greater in the fluids than seawater RASMUSSEN: COELACANTH VENTRICULAR AND NOTOCHORDAL FLUIDS Oy TaBLe 1. Electrolytes (meq/I). Serum Fluids OsMo Na Cl Mg Ca OsMo Na Cl Mg Ca Ventricular brain fluid 1,181 197 187 523 4.8 Coelacanth 1,048 185 125 Notochordal fluid 250 ilo? 1,000 190? 2047 ES eZ Ventricular brain fluid 300! 306! 7.8 4.8 Chimaera 1,670! 307! 3141 3.88! 3.03! 942 268 272 801 285 283 3.9 Ventricular brain fluid 973 255 239 3.0 5.0 Elasmobranch 9904 270 264 1 3 1,0008 Extra brain fluid 270 244 1 3 Ventricular brain fluid 500 i80 160 1.4 2.6 Teleost low 180 120 3 Amphibian 830° R. cancrivora? 220 109 71 Dal R. exculenta® Seawater 930° ' Read, L. 1971. * Griffith et al. 1975. ’ Schmidt-Nielsen 1976. 4 Total osmolarity slightly higher in ventricular fluid than serum. > Total osmolarity greater in fluids than seawater and slightly higher than serum. ® Potts and Parry 1964. (Table 1). The coelacanth notochordal chlorides Table 2 presents a comparison of previously were higher than teleost or coelacanth CSF reported coelacanth notochordal fluid data (Grif- chlorides, but lower than elasmobranchs and fith et al. 1975) and the current Coelacanth #79 chimaera (Table 1). data. Critical values for P (less than 0.05) were TABLE 2. mOsm/] mM/] Total osmolarity Na Cl K Mg Coelacanth notochordal fluid 1,058 + 2 (3)! 188.7 + .2 (3)! 204.2 + 3.8 (3)! Mleteh se 3 iD: = 2022 1,062 + 4 (5) 190.1 + .3 (10) ele S610) aI) Se of! 1.48 + .04 ventricular fluid 1,048 + 2 (4) 185.2 + THS (10) 125.9 + 4.3 (10) Wie} Se"=5) 1235-02 Chimaera ventricular fluid 801 + 3 (8) 240 + 3.2 (20) Elasmobranch ventricular fluid 990 I+ 5 (7) 264 I+ 5.1 (20) ' Griffith et al. 1975. * Mean + S.E. (number of replicate samples). Significant variation at p < 0.05 level. 98 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, No. 134 TABLE 3. UREA mM/I. Fluids Extra O-U Serum Ventricular brain Notochordal Spinal cord Cycle S/T) se 7 420 + 13 (3)! Coelacanth CSF > Serum + 540 + 8 (6) 410 + 7 (10) (Higher totals than elasmobranchs) 3034 325 + 5 (20) 425 + 6 (20) 400 + 7 (20) Chimaera CSF > Serum + 300-50 380 — 380 420 wal ie) (equivalent or lower totals than elasmobranchs) Elasmobranch CSF = Serum + — S. acanthias + — R. eglanteria CSF > Serum + — Teleost? Serum > CSF low ' Griffith et al. 1975. * Bentley 1971. 3 Forster and Bergland 1956. 4 Urist and Van de Putte 1967. noted and demonstrated both the lack of varia- tion and the precision of the methods. Especially to be noticed were the lower chloride ventricular fluid values contrasted to notochordal fluid levels. Ten replicate samples of both notochor- dal and ventricular fluids demonstrated this close correlation and precision. 2. Organics Urea Historically hemolyzed serum was one of the first coelacanth tissues available for biochemical study; analysis yielded urea values of 2,264 mg% (377 mM/l) (Pickford and Grant 1967). In sub- sequent measurements of unhemolyzed serum, urea values were, not surprisingly, close (Grif- fith et al. 1974). Many other vertebrate serum values are about 10 mM/I, including teleosts at about 5 mM/I (Bentley 1971). However other marine fishes have similar values, ratfishes, 303— 350 mM/I, and marine elasmobranchs (250-420, ave. 350 mM/l). Relevantly noted on Table 3 was the general presence or absence of the ornithine- urea cycle and its enzymes (Prosser 1972). The CSF/serum urea ratios are also listed on Table 3. In elasmobranchs the ratios were equiv- alent, whereas in the ratfish CSF the values were greater than the serum with the absolute values equivalent to or lower than elasmobranch levels: whereas in the coelacanth, the ventricular levels were greater than the serum and also higher than elasmobranchs numerical values (Table 3). Urea numerical values for coelacanth fluids included: intraocular—303 mM/I (Cole 1973); notochordal—410 mM/I (Griffith et al. 1975) and #79 420 mM/I ; ventricular fluid—S40 mM/I (Ta- ble 3). In the teleosts and elasmobranchs brain fluid and serum, urea values were equivalent to each other (although different between species) but in both the coelacanth and chimaera the levels were higher in the fluids than in the serum. Total brain fluid levels were higher in the coelacanth than in the chimaeras and elasmobranchs (Table SE Table 3 delineates differences and similarities in various types of brain fluids. In elasmo- branchs the urea levels were identical in the CSF and EDF. In both the brain and notochordal fluids of the coelacanth the actual values were higher than elasmobranch fluids. Chimaera CSF values (325 mM/l) were lower (by 33%) than EDF (425 mM/l). Spinal cord fluid (SpC.F.) chi- maera values (400 mM/Il) were similar to EDF levels. In the coelacanth, notochordal fluid val- ues were 24% lower than ventricular fluid val- ues. Coelacanth notochordal fluid urea values were higher than chimaera CSF, lower than EDF, and about the same as chimaera spinal cord fluid values (Table 3). RASMUSSEN: COELACANTH VENTRICULAR AND NOTOCHORDAL FLUIDS 99 TABLE 4. Lipips (mg%). Coelacanth #79 Chimaera Elasmobranchs Human Noto- Ven- Ven- Ven- Teleost Ven- chordal tricular tricular tricular tricular Serum fluid fluid Serum fluid Serum fluid Serum Serum fluid Total 500+7 5.0 755) 7008 700 Triglycerides 322 32 42-200 10-100 % Total 64% 140-215 Cholesterol 151 0.06 0.01 320 0.6 150 0.35 210! 140-215 0.25-0.6 4743 86-214! 350° 975 7004 662 % Total 1.2% 0.2% *Spleen *Spleen 13% ! Pacific Cod, Gadus macrocephalus. * Ling Cod, Ophiodon elongatus. 8 Carp, Cyprinus carpio, Field et al. 1943. 4 Larsson and Fange 1977. ° Urist and Van de Putte 1967. ® Nevenzel et al. 1966. 7 Estimated from reported values, Griffith et al. 1974. 3. Lipids Total lipid content was 2.5 mg% in the ven- tricular fluid and about double—S mg% in the notochordal fluid (Table 4). This compared with the following other measured biological fluids: human serum 700 mg%, elephant serum 260 mg%, marine teleosts serum 700 mg%, elephant temporal gland fluid 75 mg% (Buss et al. 1976), coelacanth serum 151 mg% (Pickford and Grant 1967), and human CSF (Griffith et al. 1975) 0.1 mg% (Table 4). Triglycerides in the notochordal fluid were 3.2 mg%; low levels compared to human serum (10- 100 mg%), elephant serum (30 mg%), ratfish se- rum (32 mg%), elasmobranch serum (42-200 mg%), and even elephant temporal gland fluid (5-10 mg%) (Table 4). In the coelacanth the per- cent of triglycerides/total lipids was 63%, com- pared with 6% values for spleen and 78% for liver (Table 4). Human serum triglycerides were only 30%. Another lipid component, cholesterol was de- tectable at low levels—O.01 mg% in the ventric- ular fluid, and 0.06 mg% in the notochordal fluid. The possible loss of some of the cholesterol and/ or its ester due to the initial handling proce- dures, subsequent freezings and thawings, and its storage in plastic containers are unknown factors. The accuracy and sensitivity were noted by the closeness of the two assay methods. Ta- ble 5 depicts the detailed values of notochordal fluid cholesterol levels, which were low (0.06 TABLE 5. CHOLESTEROL (mg@%). Ventricular brain fluid Average Range Coelacanth 0.006 0.1—0.02 Chimaera 0.6 0.8—0.4 Elasmobranch 0.36 0.42—0.30 Teleost Human! 0.44 0.29-0.59 ' Cevallos et al. 1595. 100 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, No. 134 Cholesterol Glycolipids --------- x 15 Ee) sCHOL- STD) © (PTC-sTD) a : Se GUE are ORIGIN====—=+—>-=— FeClo Diphenylamine spray spray THIN LAYER CHROMATOGRAPHY Ficure 1. Diagram depicting TLC of two types of lipids: cholesterol spots on left and glycolipids on the right. Migration distances of standards are depicted in the center. Spray-stains are indicated. mg%) even compared to human CSF (average 0.44 mg%; range 0.29-0.59 mg%, Cevallos et al. 1959), elasmobranchs CSF (average 0.36, range 0.2-0.5 mg%), and chimaera CSF (0.5 mg%; range 0.8—0.4 mg%) (Table 5). The cholesterol/ total lipid content was compared in various species and tissues; contrasting to the 13% in the coelacanth spleen (Nevenzel et al. 1966). To be noted with interest were the high serum levels in some teleosts, Gadus macrocephalus and Ophiodon elongatus. The high chimaera values (Table 4) are contrasted to the relatively low elasmobranch levels (Table 4). In addition to the total cholesterol levels mea- sured, by TLC two spots, specifically stained for cholesterol, migrated the same distance as purified, commercial cholesterol (Fig. 1). Other several general lipid components migrated less far. The two spray stains relatively specific for cholesterol, 2% FeCl, and 50% H.SO, (at times with dichromate) and charring, demonstrated two cholesterol spots, the slower migrating, quantitatively greater one probably cholesterol, and the one migrating at the solvent front, its ester (Fig. 1). Several other lipids were analyzed by TLC. 17- ketosteriods, demonstrable in the ratfish se- rum and fluids (Rasmussen, unpublished) were not detected in the coelacanth fluids. Glycolipids were readily detectable by TLC at low fluid con- centrations. Staining with diphenylamine dem- TABLE 6. PROTEINS OF NERVOUS SYSTEM ASSOCIATED FLUIDS. Mg% Ratios Ven- Noto- tricular chordal Ven- Noto- fluid fluid tricular chordal fluid fluid serum serum Chimaera 20-50 Elasmobranch 50! 1:30 Coelacanth 30) s2 2(G)) ite) se OS) 1g 1:16 210? Teleost Sculpin 300 Human 15—40 1:300 ' Certain species up to 200 mg%. > Griffith et al. 1975. onstrated 2 glycolipid components in the clear notochordal fluid, with an 18 cm advancing sol- vent front, two glycolipid components migrated 3 cm and 6 cm from the origin. The faster mi- grating component migrated the same distance as the phosphatidyl choline standard. TLC of the fiber containing notochordal fluid fraction demonstrated two glycolipids at 5 and 2 cm. 4. Proteins Protein values, the mean of five replicates, are listed in Table 6. For coelacanth #79 ven- tricular brain fluid values were 130 + 2 mg% (5), and notochordal fluid levels were 180 + 9 (5). Contrasted to human CSF, the coelacanth fluid protein values were high and compared with other fish brain and nervous system associated fluids these values were also high. The ratio of protein in these fluids/serum is listed in the last column in Table 6. The ratios were: human CSF, 1:300; elasmobranch ventricular fluid, 1:30; coelacanth ventricular fluid, 1:21; and notochor- dal fluid, 1:16. a. Soluble Proteins.—Electrophoretic sepa- ration of the soluble proteins, lipoproteins, and glycoproteins revealed a variety of proteins. First, the soluble proteins in the notochordal fluid, by CAE separated into 3 main peak areas, 20% of the protein in peak 1, 75% in peak 2, and 5% in peak 3. The middle peak 2 corresponded approximately to the general extended region of the beta-globulins. This contrasted to human se- rum and CSF where 73% (peak 1) was albumin- like proteins plus a-globulins. Coelacanth dis- tribution was more similar to elasmobranchs. RASMUSSEN: COELACANTH VENTRICULAR AND NOTOCHORDAL FLUIDS 101 Figure 2. Soluble proteins (SP), separated by AGE, stained with Amido Black, of the notochordal fluid. Table 7 also lists the number of bands re- solved by AGE, SGE, and PAGE in the noto- chordal fluid. First, by AGE, stained with Amido Black, 4 bands were discernible in the notochordal fluid (Fig. 2). Compared to the spi- nal fluid, Fig. 3, the patterns were different. By SGE in the ventricular fluid Fig. 4 at least 6 clearly recognizable bands were separated; in the notochordal fluid, utilizing two stains a TABLE 7. FIGURE 3. Soluble proteins, separated by AGE, stained with Amido Black, of the notochordal fluid (right) and ventric- ular fluid (left). Origin is faint line in the middle. greater number (8—10) of faster migrating bands were resolved, stained with Amido Black (Fig. 5) and Fast Green (Fig. 6). Quantitatively, the most amount of protein and qualitatively the greatest number of bands was seen in the beta- globulin region. Multi-concentrated soluble proteins of coel- acanth notochordal fluid (middle), cat serum FLUID PROTEINS. % type of protein (CAE) Peak | Peak 2 Peak 3 Human (CSF) 13 17 10 Elasmobranchs (CSF) S. acanthias 25 67 8 C. taurus 34 aif 9 Coelacanth (Noto.) 20 75 5 # of bands resolved by AGE SGE PAGE 4-5 6 8 Soluble proteins — 2 5 Lipoproteins 3-4 2 4 Glycoproteins 102 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, No. 134 FiGure 4. Soluble proteins, separated by SGE, stained with Amido Black of the ventricular fluid. Origin is at bottom. (left), and marine elasmobranch ventricular fluid (right) were separated by SGE runs of short du- ration (Fig. 7). Both the coelacanth notochordal fluid and the elasmobranch ventricular fluid exhibited dominant bands in the beta-globulin region with some minor (1-2) bands similar in mobility to albumin (Fig. 7). PAGE of the soluble proteins of the spinal fluid demonstrated 7-8 bands stained with Fast Green or Amido Black Table 7, Fig. 8 (left), rep- resented diagrammatically in Fig. 9. PAGE of the notochordal fluid resolved into 6 bands in the 1x concentrated material stained with Amido Black (Fig. 9), into 8 bands in the 3x COELACANTH NOTOCHORD FLUID PROTEINS Amido Black Ficure 5. Soluble proteins, separated by SGE, stained with Amido Black of the notochordal fluid. Origin is at right. concentrate (Fig. 8—middle strip) and in the 6x concentrated material, stained with Fast Green (Fig. 8—right strip), and also in the 8x concen- trate stained with Amido Black (Fig. 9). In the ventricular fluid simultaneous gels of 8x con- centrated material stained with amido black also demonstrated 8 bands; the predominance of protein was in the beta-globulin region. Noto- chordal fluid also demonstrated predominantly beta-globulin migrating proteins with several ad- ditional definitive bands in the a and y-globulin regions (Fig. 9). Compared to published protein patterns in several marine fish sera (Drilhon 1959), the patterns were distinct, bearing a re- semblance to chimaera serum and spinal fluid patterns (Rasmussen, unpublished), and elas- mobranch fluids (Rasmussen and Rasmussen 1967a). b. Lipoproteins.—In human serum 3% of the total proteins are lipoproteins which migrate electrophoretically with the alpha-globulins, and 5% with the beta-globulins (Harper 1963). Ven- tricular fluid lipoproteins, somewhat concen- trated, were separated by PAGE after prestain- ing with Sudan Black in ethylene glycol; as diagrammed in Fig. 9, the ventricular fluid re- solved into more bands than the notochordal fluid, but the percent of the faster migrating ones was greater in the notochordal fluid. Fig. 10 (left strip) is the actual gel of the 4x concentrated +i &= —-. COELACANTH NOTOCHORD FLUID J FiGure 6. Soluble proteins, separated by SGE, stained with Fast Green, of the notochordal fluid. Origin is at right. RASMUSSEN: COELACANTH VENTRICULAR AND NOTOCHORDAL FLUIDS 103 FIGURE 7. time run), of cat serum (left), coelacanth ventricular fluid (middle), and elasmobranch CSF (right), stained with Amido Black. Soluble proteins, separated by SGE (shorter notochordal lipoproteins ‘stained with Sudan Black; Fig. 10 (right strip) is the same material stained with lipid crimson. c. Glycoproteins.—Glycoproteins are hexos- amines (less than 5%) containing proteins, mi- grating electrophoretically predominantly with the albumin and alpha,-globulin fractions in mammalian serum. In human CSF the glycopro- tein was 2.5—7 mg%, being somewhat less than 15% of the total protein (Harper 1963). In the coelacanth notochordal fluid the total glycopro- FIGURE 8. ventricular fluid (left), notochordal fluid (3 concentrated) (middle), and 8x concentrated (right), stained with amido black and fast green respectively. Origin is at bottom. Soluble proteins, separated by PAGE of the tein was 9 mg% (5% of the total protein), and in the ventricular fluid 1-2 mg% (1.5% of total pro- tein) (Table 8). Table 8 also compares published coelacanth notochordal fluid values (Griffith et al. 1975), and the current coelacanth #79, both the num- ber of bands resolved on PAGE, and the molec- ular weights measured by Griffith et al. 1975. By the various electrophoretic methods, glycopro- tein bands were visualized. By AGE and azure A staining 4 glycoprotein bands were resolved in the notochordal fluid, whereas AGE, followed by PAS staining of the ventricular fluid at the same concentration resolved 2—3 bands, proba- 104 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, No. 134 POLYACRYLAMIDE GEL ELECTROPHORESIS COELACANTH FLUIDS Ventricular brain fluid Notochordal fluid Soluble proteins Lipoproteins Soluble proteins Lipoproteins Glycoproteins (Ix) (8x) 8 6 4 2 CM Amido Sudan Amido black Sudan Azure A black black black FiGure 9. Composite diagram of the PAGE separation of the soluble proteins, lipoproteins, and glycoproteins of the notochordal and ventricular fluids. TABLE 8. GLYCOPROTEINS. Bands resolved AGE SGE PAGE Total % GP Se mg% total Azure A PAS Az.A FeCl, Az.A Human (CSF) =I 15 Coelacanth Ventricular fluid =? 1S fixe} 3 1 3 Notochordal fluid! 9 5 4 3-4 2 1 4 ' Molecular weights: 50,000—100,000 (Griffith et al. 1975). TABLE 9. COELACANTH. Enzymes Ventricular fluid Notochordal fluid Amylase Somogyi Units SSS 50 + 4(5 250 + 6(5 ( ara ) (5) (5) Acid phosphatase (wm/hr) 0.38 + 0.03 (5)! 0.42 + 0.02 (5) LDH Detectable?” Detectable (std. units) Alteration of LDH levels M Urea 0.5 0.6 0.4 0.3 0.1 Decreasing % of maximum values 100 61 59 45 15 % ' Elasmobranch, S$. acanthias 0.13 + 0.02 (5). * Slightly higher than notochordal fluid. 3 §. acanthias, 185 units; H. colliei, detectable. RASMUSSEN: COELACANTH VENTRICULAR AND NOTOCHORDAL FLUIDS 105 FiGure 10. Lipoproteins of the notochordal fluid separat- ed by PAGE stained with Sudan Black (left) and lipid crimson (right). Origin is at bottom. bly mixture of glyco- and lipoproteins (Table 8). AGE, by the latter method, of the notochordal proteins also demonstrated 3—4 lipo- and/or gly- coproteins (Fig. 11). SGE, followed by azure A stain of more concentrated samples of ventric- ular fluid also resolved several bands and in the notochordal fluid 2 bands (Table 8). FeCl, stain- ing after SGE demonstrated a predominant band near origin both in the ratfish and coelacanth fluids. PAGE resolved 3 bands in the ventricular fluid, and 4 in the notochordal (Table 8, Fig. 9, FiGure 11. rated by AGE PAS staining. Origin is faint line about mid- point. Glycoproteins of the notochordal fluid sepa- Fig. 12), with the notochordal glycoproteins generally being less negative in their migration pattern than those of the ventricular fluid. Enzymes The results from the three enzyme assays were as follows: acid phosphatase levels were 0.42 wM/hr/mg protein (1 uM = 0.139 mg nitro- phenol) in the notochordal fluid and 0.38 .M/hr/ mg protein in the ventricular fluid, levels not of significant difference; contrastingly elasmo- branch brain fluid levels were 0.13 and ratfish 4.0, whereas human values ranged from 0.8—2.3 (Table 9). Amylase levels in the notochordal fluid were considerably higher—250 Somgyi units—than the ventricular fluid levels—50 Som- gyi units, which were within the limits of the human serum values (Table 9). The third enzyme LDH, as measured by the rate of NADH, depletion per minute was de- tectable in both fluids at values less than 30 units. The notochordal fluid was somewhat low- er than the ventricular fluid, and low compared to other brain fluids such as S$. acanthias (185), and human (40) (Richterich 1969), but was of comparable values to chimaera brain fluids (Ta- ble 9). Comparative serum values were human serum (225 units), and the high salmon serum (4,000 units). The detectability of LDH in the spinal fluid samples frozen, without the stabiliz- 106 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, No. 134 FiGure 12. Glycoproteins, separated by PAGE of the notochordal fluid (right) and ventricular fluid (left) stained with azure A. Origin is at bottom. ing NADH, additive, and thawed three times prior to assay (Table 9) was remarkable. Observing the effects of varying urea concen- trations on the levels of LDH and taking 100% equal to the highest total assay value obtained at 0.5 M urea, progressive loss of LDH level was observed on either side of this urea opti- mum. Specifically at 0.4 M, the LDH level was only 59% of the total, and at 0.6 M only 61%; at 0.3 M only 45%, and at 0.1 M only 15%. Tak- ing into account the 3% precision of the rate reaction as utilized, the differences are signifi- cant (Table 9). DISCUSSION Several results from this biochemical investi- gation of coelacanth #79 notochordal and ven- tricular brain fluids are worthy of further em- phasis. First, among the inorganic components, chloride values in the coelacanth ventricular fluid were equivalent to the chloride levels in chimaera and elasmobranch ventricular fluids. As seen in Table 2, notochordal chloride values were relatively close, but significantly different from previously reported values (Griffith et al. 1975), and notochordal chloride values were def- initely significantly different from ventricular levels. Among the simple organic compounds, urea levels were of interest. The relative importance in the coelacanth of the ornithine-urea cycle and its enzymes (Brown and Brown 1967) and the probable functioning of the whole coelacanth organism as a urea space (Brown and DiJulio 1976) are relevant to a discussion of the high coelacanth urea levels. Coelacanth liver values were high—284 mM/I—(DiJulio and Brown 1976): high urea values were also found in the ventricular and notochordal associated fluids. The values for the coelacanth ventricular fluid were among the highest values reported for ver- tebrates, 540 mM/I, considerably higher than the notochordal fluid values or the 350 mM/Ifor the ventricular fluid of the hyperuremic marine elas- mobranchs; like the chimaeras, the values were higher in the ventricular fluid than the serum. Numerically, the values were the same order of magnitude as the chimaera extradural brain and spinal cord fluids. These high urea values are especially intriguing in view of some initial en- zyme results. The detection of the glycolipids in the CSF was of interest in view of the low amount of albumin in both this fluid and serum, and cor- related with the resolution of glycoproteins in the fluids by various electrophoretic methods. Of interest is the higher total soluble protein content in the notochordal fluid than in the ven- tricular fluid; not unexpected as the notochordal canal lining is composed of secretory type cells. Compared to fluids in other species, protein levels were numerically high. That the majority of the soluble proteins mi- grated with the mobility of the beta-globulins, RASMUSSEN: COELACANTH VENTRICULAR AND NOTOCHORDAL FLUIDS 107 and that the albumins detected were in very low concentration and had a different mobility from mammalian albumins was confirmed by all the electrophoretic methods. Obviously the proteins of the coelacanth brain fluids were different in type-proportions and within a given type, dif- fered in size and mobility from mammalian and other vertebrate forms. Compared to other fish species, the most similarity was noticed to the chimaera fluids and then next to the elasmo- branchs fluids. The demonstration of the variety of lipopro- teins and glycoproteins was of interest. The no- tochordal lipoproteins were distinctive in pat- tern when compared to the ventricular fluid proteins. Glycoproteins were also distinctive in the notochordal fluid, resolving into, at the same concentration, a greater number of bands than the ventricular fluid. These coelacanth glycopro- teins, were of different mobility and of higher molecular weight than the serum glycoproteins of certain Antarctic teleosts (DeVries et al. 1970). Data (Griffith et al. 1975) indicated some similarities in carbohydrate content to human CSF, namely, glucosamine, mannose, galactose, and sialic acid (Gottschalk 1966). Interestingly, the observations (Qureschi et al. 1977) on the more even distribution of various electropho- retic zones of Oncorhynchus nerka, sockeye salmon, than the cow was somewhat apparent by less quantitative methodology in the coel- acanth fluids, especially the notochordal. Other complex organics such as the total lip- ids were low in comparison to serum but higher than human CSF, notochordal fluid being twice that of ventricular fluid; triglycerides were rel- atively high, more than 50% of the total lipids, and cholesterol levels were between 0.2—1.2%. Both the assay of low levels of total cholesterol (0.06 mg%) and the detection by TLC of two cholesterol spots was possible because of our use of an ultrasensitive measurement technique; these levels may be influenced by the handling procedures or unknown degrees of stress to the animal. Our percent cholesterol results correlat- ed with the moderately high reported levels for coelacanth serum (Pickford and Grant 1967), and compared to high serum cholesterol in var- ious marine fishes (Larsson and Fange 1977), including the ling and Pacific cod (Table 4). The cholesterol content was comparable to that of marine elasmobranchs (Griffith et al. 1975). Serum cholesterol levels may have some correlation to FFA levels in turn related to both the triglyceride content (Nevenzel et al. 1967) and the rate of lipase activity. A link between these lipid levels and observations on LDH may have a corollary in the pulmonate land snails, which can survive long periods of anoxia by de- pleting glycogen and utilizing anaerobic glycol- ysis and have the resultant end products of FFA (Storey 1977). As a relationship exists between high FFA and triglycerides and like other marine fishes utilizing muscle and liver for triglyceride storage (Larrson and Fange 1977), 78% of the lipids in the coelacanth liver are in the form of triglycerides (Nevenzel et al. 1967), the inves- tigation of whether the coelacanth funnels its LDH reaction toward FFA end products rather than lactate, and the measurement of FFA is projected. The presence of a pyruvate inhibited H-type LDH partially accounts for this channelling in the snail. The habitat of the coelacanth is in question; it may live in oxygen deficient regions or in a freshwater cave habitat. Yet two fishes— the halibut and lamprey dwelling in rather an- aerobic environments possessed only M-type LDH (Wilson et al. 1964). It would be of rele- vance to determine the predominant LDH type (M or H), and the relative percent of isozymes. The measurement of total LDH levels, the determination of the optimal pyruvate concen- tration for maximum activity, and the establish- ment of the relative percent of the isozymic forms may provide information on this habitat and relevant biochemical information on the functioning of carbohydrate metabolism in the coelacanth. Total LDH was present at detect- able levels in the frozen-thawed coelacanth fluids (Rasmussen 1977), and in the liver (DiJulio and Brown 1976). The latter two aspects are cur- rently being investigated. The exciting detectability of measurable levels of enzymes in quite old frozen fish tissue pre- sents some interesting speculation regarding en- zyme stability. The enzymes were assayable in both the concentrated and unconcentrated fluids of notochordal and ventricular fluids; AcP levels were similar or slightly higher than levels re- ported for elasmobranchs; amylase levels were higher in the notochordal fluid than the spinal fluid; but LDH interestingly was higher in the spinal fluid than the notochordal. Because op- timal LDH levels appeared to be at high critical urea levels as seen in the supplemental urea ex- 108 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, No. 134 periments, it is possible that in these primitive fishes the urea may have a functional stabil- ization role in the preservation of certain en- zyme activity levels. The interplay that the high levels of trimethylamine oxide (TMAO) reported (Griffith et al. 1975) may have in correlation with the urea or with another aspect of cellular me- tabolism is under consideration. SUMMARY By a variety of techniques, including total as- says and electrophoretic and TLC separations, inorganic and organic constituents of the coel- acanth notochordal and ventricular fluids result- ed in some interesting observations both about the coelacanth itself and in comparison to other marine fishes. Because of numerically repro- ducible results with less than expected varia- tions from replicate samples and, despite a dif- ferent coelacanth specimen obtained, stored, dissected, and assayed under different condi- tions, valid comparisons could be made with the notochordal fluid from the other previous coel- acanth specimen, for with several exceptions the data from these two coelacanths, in parameters dually measured, were close. Measurements of special interest and rele- vance included: 1. Chloride measurements in the notochordal fluid were considerably higher than the ventric- ular fluid levels (and also were significantly higher than Griffith et al. 1975 values), and slightly lower than levels in marine elasmo- branch ventricular fluid. 2. The urea level value (540 mM/l) in the ven- tricular fluid was significantly higher than the 410 mM/I in the notochordal fluid, and the 350 mM/I in elasmobranch ventricular fluid. The nearest values to the high ventricular fluid val- ues were found in ratfish EDF (425 mM_/l). 3. Lipid fluid analysis showed respectively in the ventricular and notochordal fluids, total lip- ids 2.5 mg% and 5 mg%:; triglycerides less than 3.2 mg%, and cholesterol 0.01 mg% and 0.06 mg%. The total triglycerides were low, but the % per total lipids was high. Separation by sev- eral TLC techniques demonstrated two choles- terol components and two glycolipid compo- nents. 4. Total proteins were respectively in the no- tochordal and ventricular fluids, 180 and 130 mg%. The protein content was comparatively high, three fold than of elasmobranch brain fluid, which was higher than human CSF. This is demonstrated by the ventricular fiuid/serum ratio values respectively, coelacanth 1:16 and 1:21, shark 1:30, and human 1:300. 5. Soluble proteins were predominantly in the beta-globulin region as seen by CAE; again a similarity to elasmobranchs and a dissimilarity to humans. By four electrophoretic techniques up to 8 bands were resolved and differences not- ed between the notochordal and ventricular fluids, with a similarity noted to the proteins of the ratfish (unpublished). 6. The lipoproteins demonstrated more bands in the ventricular fluid (6) than the notochordal. The hexosamine containing glycoproteins were greater in the notochordal fluid (9 mg) compared to the ventricular fluid (2.5—7 mg). The data were in accord with Griffith et al. (1975). At least four bands were resolved by several techniques, in the notochordal fluid. 7. Enzyme level measurements demonstrated higher levels of amylase in the notochordal fluid compared to the ventricular fluid, AcP similar levels, and LDH slightly higher levels in the ventricular fluid. Data were presented showing the effect of urea as an additive. 8. Possibilities regarding the role of urea and its relation to coelacanth physiological function- ing in its natural environment are discussed. ACKNOWLEDGMENTS Numerous acknowledgements are necessary to give proper credit to this multi-supported ef- fort. Foremost, George Brown, College of Fish- eries, University of Washington is recognized as the prime mover in the obtaining of good phys- iological quality coelacanth tissue for biochem- ical analysis through the initial organization and laborious building up of SPOOF (Society for the Protection of Old Fishes). Prime financial sup- port for the particular coelacanth #79 used in this study was generously awarded by the Char- line H. Breeden Foundation through CAS 33111. Special logistics and the actual obtaining of the specimen was due to the efforts of M. Lagios and J. McCosker. I. O. Buss, L. Kirschner, and H. Went, Department of Zoology, Washington State University are thanked for their generous use of laboratory facilities. The expert, deft re- moval of the brain and notochordal fluids was performed by Susan Brown. RASMUSSEN: COELACANTH VENTRICULAR AND NOTOCHORDAL FLUIDS 109 LITERATURE CITED AMADOR, E., H. REINSTEIN, AND N. BENoTTI. 1965. Pre- cision and accuracy of lactic dehydrogenase assays. Am. J. Clin. Pathol. 44:62—-70. ARCHIBALD, R. 1945. Colorimetric determination of urea. J. Biol. Chem. 157:507-518. BENTLEY, P. 1971. Endocrines and osmoregulation. Spring- er-Verlag, New York. 300+ pp. Brown, G., AND S. Brown. 1967. Urea and its formation in coelacanth liver. Science 155:570—5S72. Buss, I., L. RASMUSSEN, AND G. SMuTs. 1976. Role of stress and individual recognition in the function of African ele- phants’ temporal gland. Mammalia 40:437-451. Ceska, M., K. BIRATH, AND R. BRom. 1959. A rapid method for the clinical determination of a-amylase activities in hu- man serum and urine. Clin. Chim. Acta 26:437-444. CEVALLos, W., N. PAPADOPOULOUS, AND W. Hess. 1959. Determination of cholesterol and phospholipid in CSF. Fed. Proc. 18:202. Curistic, W. 1973. Lipid analysis. Pergamon Press, Oxford, pp. 30-40; 152-197. Cote, D. 1973. Intraocular fluid composition in the coel- acanth, Latimeria chalumnae. Exp. Eye Res. 16:389-395. Davis, B. 1964. Disc electrophoresis II (Method and appli- cation to human serum proteins). Ann. N.Y. Acad. Sci. 121:404—-427. DiJuLio, D., AND G. BRown. 1975. Urea levels in tissues of the coelacanth, Latimeria chalumnae. Am. Zool. 15:801. DRILHON, A. 1959. Etude electrophoretique en gel d’amidon de serums de Poissons, Cyclostomes, Selaciens, Teleos- teens. C. R. Soc. Biol. 153: 1532-1535. DeVries, A., S. KOMATSU, AND R. FEENEY, 1970. Proper- ties of freezing-point-depressing glycoproteins. J. Biol. Chem. 245:2901-2908. FieLb, J., C. ELVEHJEM, AND C. JuDAy. 1943. A study of the blood constituents of carp and trout. J. Biol. Chem. 148:261-269. Fotcn, J., N. Lees, AND C. H. SLOANE-STANLEY. 1957. A simple method for isolation and purification of total lipids from animal tissues. J. Biol. Chem. 226:491—509. FREEMAN, C. P., AND D. West. 1966. Complete separation of lipid classes on a single thin layer plate. J. Lipid Res. 7:324-327. FrINGs, C. S., AND R. T. DUNN. 1970. A colorimetric meth- od for determination of total serum lipids based on the sulfo- phospho-vanillin reaction. Am. J. Clin. Pathol. 53:89-91. , L. P. Forster, AND P. S. COHEN. 1971. Electro- phoretic separation of serum lipoproteins on polyacrylam- ide gel. Clin. Chem. 17:111-114. GoTTSCHALK, A. 1966. Glycoproteins. Elsevier, New York, Ps622 GriFFITH, R., M. MATHEWS, B. UMMINGER, B. GRANT, P. PANG, K. THOMSON, AND G. PICKFORD. 1975. Composi- tion of fluid from the notochordal canal of the coelacanth, Latimeria chalumnae. J. Exp. Zool. 192:165—172. , B. UMMINGER, B. GRANT, P. PANG, AND G. PICk- FORD. 1974. Serum composition of the coelacanth, Lati- meria chalumnae. J. Exp. Zool. 187:87-102. Harper, H. 1963. Physiological chemistry. Lange Medical, Los Altos, California, p. 136. ‘a Hawk, P., B. OseER, AND W. SUMMERSON. 1954. Practical physiological chemistry. McGraw-Hill, New York. Hays, A., AND G. WATSON. 1976. The plasma transport of 25 hydroxy cholecalciferol in fish, amphibians, reptiles, and birds. Comp. Biochem. Physiol. 53B: 167-172. HOCHACHKA, P. 1965. Isoenzymes in metabolic adaption of a poikilotherm: Subunit relationships in lactic dehydroge- nases of goldfish. Arch. Biochem. Biophys. 111(1):103. IDLER, F., A. RONALD, AND P. SCHMIDT. 1959. Biochemical studies on sockeye salmon during spawning/migration VII. Can. J. Biochem. Physiol. 37:1227-1238. JATZKEWITZ, H., AND E. MEHL. 1960. Cholesterol identifi- cation by thin layer chromatography. Z. Physiol. Chem. 320:251. LARSSON, A., AND R. FANGE. 1977. Cholesterol and free fatty acids (FFA) in blood of marine fish. Comp. Biochem. Physiol. 57:3B:191-197. Lockwoop, A. P. 1963. Animal body fluids and their regu- lation. Harvard, p. 50. Lowry, O., N. ROSEBROUGH, A. FARR, AND R. RANDALL. 1951. Protein measurement with folin phenol reagent. J. Biol. Chem. 193:265. , AND I. TINSLEY. 1974. A simple sensitive method for lipid phosphorus. J. Lipids 9:491. McDoNnaLbD, H., AND L. RiBerRO. 1959. Ethylene and pro- pylene glycol in pre-staining of lipoproteins for electropho- resis. Clin. Chim. Acta 4:458. Mopper, C. P. 1973. Investigation on acid phosphatase ac- tivity in human plasma and serum. Clin. Chim. Acta 43:205. Musquera, S., J. MASEGY, AND J. PLANAS. 1976. Blood proteins in turtles. Comp. Biochem. Physiol. 55:225—230. NEVENZEL, J.. W. RODEGKER, J. MEAD, AND M. Gor- DON. 1966. Lipids of the living coelacanth Latimeria chal- umnae. Science 152:1753-175S. PEDERSEN, H. 1973. Cerebrospinal fluid cholesterol. Acta Neurol. Scand. 49:626-638. PicKFoRD, G., AND F. GRANT. 1967. Serum osmolarity in the coelacanth, Latimeria chalumnae, urea retention and ion regulation. Science 155:568—570. Potts, W., AND G. PARRY. 1964. Osomotic and ionic regu- lation in animals. Pergamon Press, Oxford. Pratt, J., AND W. DANGERFIELD. 1963. Polyacrylamide gel of increasing concentration gradient for electrophoresis of lipoproteins. Clin. Chim. Acta 23:189-201. Prosser, L. 1972. Comparative animal physiology, p. 303. QurescH!l, M., W. VANSTONE, AND P. ANASTASSIADIS. 1976. Resolution of glycoproteins of Pacific salmon blood by electrophoresis and fractional precipitation and chro- matography. Comp. Biochem. Physiol. 55B:245—248. RASMUSSEN, L. 1971. Organ distribution of exogenous '*C- urea in elasmobranch with special regard for nervous sys- tem. Comp. Biochem. Physiol. 40A:145—154. 1974a. '4C-urea distribution in chimaeras. Comp. Biochem. Physiol. 47A:729-743. 1974b. 'C-urea distribution in rajiformes and te- leosts. Comp. Biochem. Physiol. 47A:745-751. , AND R. RASMUSSEN. 1967a. Comparative protein and enzyme profiles of cerebrospinal fluid, extradural fluid, ner- vous tissue and serum of elasmobranchs. Pages 361-381 in P. Gilbert, R. Mathewson, and D. Rall, eds., Sharks, skates, and rays. Johns Hopkins Press, Baltimore. , AND 1967b. Some observations on the pro- teins and enzyme levels and fractions in normal and stressed elasmobranchs. Trans. N.Y. Acad. Sci. 29:397-414. 1977. Exogenous 'C-urea distribution , AND 110 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, No. 134 in two species of salmon and resultant alteration in sera and gill enzyme levels. Comp. Biochem. Physiol. 56A:403-409. Reap, L. 1971. Chemical constituents of body fluids and urine of the holocephalian, Hydrolagus colliei. Comp. Bio- chem. Physiol. 39:(2A): 185-193. RICHTERICH, R. 1969. Clinical chemistry, theory and prac- tice. Academic Press, New York. SCHALES, O., AND S. SCHALES. 1941. Simple and accurate method for the determination of chloride in biological fluids. J. Biol. Chem. 140:879. SCHMIDT-NIELSEN. 1976. Animal physiology, 4th ed. Pren- tice Hall, Inc. Englewood, N.J. SmitHies, O. 1959. An improved procedure for starch gel electrophoresis. Biochem. J. 71:585—587. SOLOMON, F., AND G. Brown. 1975. Enzymes of the coel- acanth, Latimeria chalumnae. Soc. Exp. Biol. Med. Ab- stract. STONE, G., AND J. GOLDZIEHER. 1949. A rapid colorimetric method for the determination of sodium in biological fluids and particularly in serum. J. Biol. Chem. 181:511—521. Storey, K. 1977. LDH in tissue extracts of the land snail, Helix aspersa. Comp. Biochem. Physiol. 56B:181. UriEL, J. 1964. Pages 30-81 in P. Grabar, and P. Burtin, eds., Immuno-electrophoretic analysis. Elsevier, New York. Urist, M., AND K. VAN DE PuTTE, 1967. Comparative bio- chemistry of blood of fishes. Pages 271-285 in P. Gilbert, R. Mathewson, and D. Rall, eds., Sharks, skates, and rays. Johns Hopkins Press, Baltimore. WAKAYAMA, J., AND J. SWANSON, 1977. Ultraviolet spec- trometry of serum triglycerides by a totally enzymic meth- od, adaption to centrifugal analyzer. Clin. Chem. 23(2):223- 228. WILSON, A., N. KAPLAN, L. LEVINE, A. Pesce, M. REICH- LIN, AND W. ALLISON. 1964. Evolution of lactic dehydro- genases. Fed. Proc. 23:1258—1265. Witte, D., D. BARRETT, AND D. Wycorr. 1974. Enzymatic determination of cholesterol. Clin. Chem. 20:1282. YANAGISAWA, F. 1955. New colorimetric determination of calcium and magnesium. J. Biochem. 42:3—-11. ZACHARIUS, R. M., T. E. Zett, J. MORRISON, AND J. J. Woop.ock. 1969. Glycoprotein staining following electro- phoresis on acrylamide gels. Anal. Biochem. 30:148-152. ZOLLNER, N., C. EBERHAGEN, AND G. WOLFRAM. 1966. Uber die Zusammensetung der Cholesterinstei bei einigen Leber-Krankheiten. Klin. Wochenschr. 39:817-819. OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES The Biology and Physiology of the Living Coelacanth No. 134, 17 pages December 22, 1979 CHORDATE CYTOGENETIC STUDIES: AN ANALYSIS OF THEIR PHYLOGENETIC IMPLICATIONS WITH PARTICULAR REFERENCE TO FISHES AND THE LIVING COELACANTH By Guido Dingerkus Department of Ichthyology, The American Museum of Natural History, New York, New York 10024 INTRODUCTION Since the early 1960's, cytogenetic studies have become a favored method of gaining insight into the evolutionary relationships of verte- brates. Although it is sometimes assumed that they provide more basic data than do osteolog- ical and other anatomic approaches, I doubt that cytogenetic analysis is intrinsically any more in- formative than the more traditional anatomic studies. Such studies, however, do permit in- vestigation of another independent set of char- acters which are unavailable to traditional ex- amination, and these may complement the contribution of the morphologic methods toward analysis of phylogenetic relationships. The first examination of fish chromosomes (by Retzius 1890, on Myxine glutinosa) and subse- quent studies until 1964 were limited to histo- logically sectioned and prepared material. Al- though microscopic sections provide a reasonably accurate chromosome number, ac- tual three-dimensional chromosome morphology or karyotype cannot be determined. It was not until the introduction of ‘squash’ preparations (Roberts 1964) and flame-dry methods (Denton 111 and Howell 1969) that both fish chromosome number and morphology could be documented. Flame-dry and similar air-dry techniques are to- day the most widely-used techniques to prepare fish chromosomes and permit the preparation of reliable chromosome karyotypes and idiograms. Although karyotypes have been described for perhaps only 400 of the estimated 20,000 species of living fishes, representative species of most of the major phyletic groups have been studied. These morphologic methods have been supple- mented significantly by the introduction of tech- niques such as fluorometric assay (Hinegardner 1971), which permit DNA-per-cell values to be determined and compared for most of the major phyletic groups of fishes. Latimeria has never been karyotyped, but its nucleoli, in interphase nuclei, have been exam- ined (Dingerkus 1977; Dingerkus et al., unpub- lished data) and its DNA/cell content has been measured (Vialli 1957; Pedersen 1971; Cimino and Bahr 1974). On the basis of these data from Latimeria, and the same information, as well as karyotypic data from other chordate groups, the cytogenet- 112 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, No. 134 ic relationships of Latimeria can be evaluated and a hypothetical Latimeria karyotype pro- posed. CHROMOSOME NUMBERS AND MORPHOLOGY The following terms will be used to describe chromosome morphology. Microchromosomes are small, usually less than 0.5 uw in the greatest dimension, and without discernible centromere or chromosome arms. Macrochromosomes are larger than 0.5 w with a discernible centromere and chromosome arms. On the basis of centro- mere position, macrochromosomes may be sep- arated into three groups: acrocentric, metacen- tric, and submetacentric, in which the centromere is located respectively at or extremely close to the terminal end of the chromosome, at or near the median of the chromosome, or in an inter- mediate position. Figure 1 shows comparative idiograms of chromosomes of all the principal chordate groups. An attempt has been made to draw these as accurately as possible to the scale indicated on the first idiogram so that size and morphology can be compared easily. However, because of their extremely large size the idiogram of Nec- turus maculosus has been drawn at half scale. Illustrated idiograms of representative species appear in parentheses in the text to facilitate re- ferral to the figures. Tunicates (represented by Ciona intestinalis) have a karyotype of 8 to 32 small microchro- mosomes and no macrochromosomes. Although reports of 4 to 40 chromosomes exist in the lit- erature, it is believed that these are due to the inaccuracies of older techniques. The true basic chromosome number of tunicates seems to be close to 24 (Taylor 1967; Colombera 1973). Am- phioxi (represented by Branchiostoma floridae) have a karyotype of 38 microchromosomes, but these are somewhat larger than those of tuni- cates (Howell and Boschung 1971). This would suggest that the primitive chordate karyotype may consist of microchromosomes. The hagfishes (Eptatretus stoutii) have a chro- mosomal makeup of 46 to 52 acrocentric chro- mosomes (Taylor 1967; Potter and Robinson 1973). Nygren and Jahnke (1972a) found an ex- tremely variable number of chromosomes in Myxine glutinosa. 1 believe that much of the latter variation arose from a loss of chromo- somes or from overlapping spreads as a result of the technique that was used. Lampreys of the northern hemisphere (/ch- thyomyzon gagei) have a much higher number of chromosomes, ranging from 142 to 174. In some of these, it is extremely difficult to detect the centromere, and they may be true micro- chromosomes. The rest are all acrocentric (Howell and Duckett 1971; Potter and Robinson 1973). The southern hemisphere lampreys (Mor- dacia mordax) have a chromosome number of 76 and all chromosomes are metacentrics (Pot- ter, Robinson and Walton 1968; Potter and Rob- inson 1973). In the holocephalians or chimaeras, chromo- some numbers of 58 and 86 have been found for Hydrolagus colliei and Chimaera monstrosa, respectively. Both have a mixture of acrocentric chromosomes and microchromosomes (Ohno et al. 1969b; Nygren and Jahnke 1972b). Among the skates and rays, composite chro- mosome numbers also exist. In the family Raj- idae (Raja clavata) a range of 98 to 104 chro- mosomes was found, consisting of a mixture of a few metacentric chromosomes, many acrocen- tric chromosomes, and many microchromo- somes ( Nygren and Jahnke 1972b). Stingrays, family Dasyatidae (Dasyatis sabina), have 68 to 84 chromosomes. Most of these are metacen- trics, the rest being acrocentric, or just barely submetacentric (Donahue 1974). Electric rays, family Torpedinidae (Narcine brasiliensis), have a very low number of 28, all chromosomes being metacentric or submetacentric and most of them large in size (Donahue 1974). Only two species of sharks have been reliably karyotyped, Squalus acanthias and Etmopterus spinax (Nygren and Jahnke 1972b). The latter has a karyotype of 86 chromosomes, made up of acrocentrics and microchromosomes. S. acanthias has a karyotype of 78 chromosomes, comprising a mixture of metacentrics and acro- centrics but no microchromosomes. In the Chondrostei, consisting of paddlefish- es, family Polyodontidae (Polyodon spathula), and the sturgeons, family Acipenseridae, a high number of chromosomes has been found, rang- ing from 112 to 240. These are a mixture of meta- centrics, submetacentrics, acrocentrics, and mi- crochromosomes (Ohno et al. 1969a; Fontana and Colombo 1974; Dingerkus and Howell 1976). Among the gars, family Lepisosteidae, only Lepisosteus oculatus (= productus) has been karyotyped. It has 68 chromosomes with a mix- DINGERKUS: CHORDATE CYTOGENETIC STUDIES ture of metacentrics, acrocentrics, and micro- chromosomes (Ohno et al. 1969). The bow fin, Amia calva, has 46 chromosomes, a mixture of submetacentrics to metacentrics and acrocen- trics but with no microchromosomes (Ohno et al. 1969). In the Teleostei, the most speciose of all the fish groups, a diverse array of chromosome numbers has been found (see the tabulation of chromosome numbers in Chiarelli and Capanna 1973; and Denton 1973). Among the teleosts, no species exhibits microchromosomes. In the os- teoglossids (Osteoglossum bicirrhosum), there is a variety of chromosomal compositions, rang- ing from 34 to 56, including various mixtures of metacentric and acrocentric chromosomes (Uyeno 1973). Clupeids (Alosa pseudoharengus ) have 48 acrocentric chromosomes (Mayers and Roberts 1969). The Euteleostei, or higher te- leosts, have chromosomes ranging from 14 to 140, with various assortments of metacentrics, submetacentrics and acrocentrics. The basic chromosome number has been suggested to be 48 by Ohno (1970a). The Brachiopterygii, comprised solely of the Polypteridae (Polypterus palmas), have 36 rath- er large, submetacentric to metacentric chro- mosomes (Denton and Howell 1973). The Dipnoi or lungfishes (Lepidosiren para- doxa) are similar to polypterids in having 38 sub- metacentric to metacentric chromosomes, but they are huge in size (Ohno and Atkin 1966). Older studies (Wickbom 1945) indicate this also holds true for Neoceratodus and Protopterus. The amphibians exhibit very divergent karyo- types among the three groups: Urodela (sala- manders), Anura (frogs and toads), and Apoda (caecilians). The Urodela (Necturus maculosus) have a chromosome number between 22 and 64, the majority of species having mostly metacen- tric to submetacentric chromosomes with some acrocentrics, although some species have no ac- rocentrics (Morescalchi 1973). Their chromo- somes are extremely large, and with the dip- noans share the distinction of the largest known animal chromosomes. Anurans (represented by Ascaphus truei) generally have a chromosome number that ranges from 14 to 48, but a few species have 104 chromosomes (Morescalchi 1973). The more primitive species (such as as- caphids and discoglossids) have a mixture of metacentrics, submetacentrics, acrocentrics or microchromosomes. In contrast, the advanced 113 species (such as bufonids, hylids, ranids and rhacophorids) have a mixture of metacentrics to submetacentrics. The Apoda (Geotrypetes seraphinii) have a range of 20 to 42 chromo- somes, with a mixture of metacentrics to sub- metacentrics and some very small acrocentrics or microchromosomes (Stingo 1974; Morescal- chi 1973). Reptiles can be divided into three karyotype groups. The tuatara, Sphenodon punctatus, has 36 chromosomes. These are mostly large meta- centrics to submetacentrics, with a few very small metacentrics and one pair of very small acrocentrics or microchromosomes (Wylie et al. 1968). The crocodilians (Alligator mississipien- sis), of which all species have been karyotyped (Cohen and Gans 1970), have from 30 to 42 chro- mosomes. In some species, presumably the more primitive ones, the karyotype is made up of a number of large metacentrics, fairly large acrocentrics, and very small metacentrics. In other species, the acrocentrics appear to have fused, resulting in a karyotype consisting entire- ly, or almost entirely, of metacentrics which range in size from large to very small. The liz- ards and snakes (the Squamata) and the turtles have very similar chromosomal makeups. The lizards (Agama pallida), have from 22 to 63 chromosomes (see review by Gorman 1973). The species with higher numbers have a mixture of mainly acrocentrics and microchromosomes; those with lower numbers have more metacen- trics, probably resulting from fusions of acro- centrics. This same general pattern is shown by the snakes (Drymarchon corais), with 24 to 50 chromosomes (see review by Gorman 1973), and again in the turtles (Chelydra serpentina) in which chromosome numbers from 26 to 68 are found (Killebrew 1977; Gorman 1973). In turtle karyotypes, however, there are more metacen- trics. The birds have a pattern very similar to the latter reptile groups, but with higher numbers of chromosomes, that is, from 56 to 90 (see review by Ray-Chauduri 1973). The loon, Gavia stel- lata, as can be seen from the idiogram, has a combination of mostly acrocentrics, one pair of submetacentrics, and many microchromosomes. Other birds with lower numbers have more metacentrics, again probably arising from fusion of acrocentrics. The monotremes (Tachyglossus aculeatus) have a chromosome number from 54 to 64, con- 114 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, No. 134 sisting of submetacentric and metacentric chro- mosomes. Marsupials (Didelphis marsupialis) have 14 to 22 chromosomes, the species with higher numbers having all acrocentrics and those with lower numbers proportionally more metacentrics (Sharman 1973). Among the eutherian mammals, between 6 and 84 chromosomes have been found (see re- view by Matthey 1973). Once again, the species with higher numbers have a karyotype consist- ing mostly of acrocentrics, and in the lower- numbered species there are more and larger metacentrics. Especially in the lemurs (Lemur catta), a series of very small chromosomes has been described. These have been considered to be very small acrocentrics, but their centro- meres are very difficult if not impossible to vis- ualize, and I refer to them as 8 microchromo- somes. In the fox, a few ‘“‘microchromosomes”’ have been observed (Low and Benirschke 1972). These differ in appearance and number among individuals of this species, however, and prob- ably are 8 chromosomes rather than true micro- chromosomes. Evidence indicates these 8 chro- mosomes represent deleted segments from a chromosomal rearrangement. They occur in varying low numbers and are not present in all individuals. Therefore, these 8 chromosomes are not to be confused with true microchromo- somes which are found in constant, and usually high, numbers in all individuals of a species. DNA PER CELL CONTENT It is well established that cells from the same individual at the same state of the cell cycle, and also cells from other individuals of the same species, contain the same amount of DNA. Be- cause this uniformity exists, DNA/cell values have become another basis for comparison be- tween different species and higher systematic categories. When used in conjunction with karyotypes, DNA/cell values can yield infor- mation as to whether karyotypic differences are due to minor changes, such as chromosomal rearrangement, or major changes such as tetra- ploidy. Moreover, since DNA is the basic ge- netic material, it is believed that these DNA/cell comparisons can give an estimate of the funda- mental similarity or difference between groups. In the following discussion and in Figure 2, the DNA values given are those per haploid cell. In the graph, each bar indicates the range of DNA values found within a systematic group but every value within a particular bar is not nec- essarily represented by a species possessing that value. Tunicates have the lowest DNA/cell of any chordate. These values range from 0.15 to 0.21 picograms (pcg) for various species. Amphiox- us, Branchiostoma lanceolatus, has 0.6 pcg of DNA per cell. Eptatretus stoutii, a hagfish, has a rather high value of 2.8 pcg. The northern hemisphere lampreys exhibit between 1.4 and 2.5 peg of DNA per cell. The only holocephalian that has been studied, Hydrolagus colliei, has 1.5 to 1.6 pcg. Sharks and rays range between 2.75 and 9.8 pcg. For sturgeons, values of 1.75 to 5.1 peg have been obtained. No evaluation has yet been made for either species of paddlefish. The gar, Lepi- sosteus oculatus (= productus) has a value of 1.35 pcg, the alligator gar, Atractosteus spathu- la (= Lepisosteus ferox), has 1.2 peg, and the bowfin, Amia calva, has 1.23 pcg. Teleosteans exhibit a wide range of DNA val- ues, viz. 0.4 to 4.4 pcg. In Latimeria, Vialli (1957) obtained a value of 2.8 pcg, Pedersen (1971) one of 3.5, and Cimino and Bahr (1974) one of 3.61 pcg. Cimino and Bahr (1974) have also reinterpreted Vialli and Pedersen’s data to values of 3.465 and 3.8 pcg, respectively. It seems valid to conclude that the DNA/cell of Latimeria lies somewhere between these values. In the brachiopterygians the DNA/cell value has been estimated at 4.7 to 5.8 pcg (Bachmann, Goin, and Goin 1972; Hinegardner 1976). The species of living dipnoans range between 80.2 and 140 pcg of DNA/cell. The lungfishes probably have the largest amount of DNA/cell of any living animal. Urodelans also have extremely high values ranging from 10 to 95 pcg per cell. Anurans have lower values, from 0.8 to 10.5 peg, and caecili- ans are similar with 3.7 to 13.9 pcg. No DNA/cell measurement has been made on Sphenodon, the tuatara. Snakes and lizards both range from 2.1 to 2.4 peg. Turtles and crocodil- ians share ranges from 2.8 to 3.0. Birds have surprisingly low amounts of DNA, ranging from 1.54 to 2.0 pcg—especially in view of their high number of chromosomes. Mammals have a fairly uniform range, with monotremes exhibiting from 3.2 to 3.4 pcg, marsupials from 3.4 to 3.5, and eutherian mammals from 3.1 to 3:5: DINGERKUS: CHORDATE CYTOGENETIC STUDIES NUCLEOLI AND NUCLEOLAR ORGANIZER REGIONS (NORs) Nucleolar organizer regions (NORs) are re- gions of certain chromosomes that are com- posed of highly repetitive ribosomal DNA (rDNA), which code for ribosomal RNA (rRNA) and in the interphase cell forms the nucleolus or nucleoli (Howell 1977). The NORs and nucleoli can be selectively stained using an ammoniacal silver technique (Denton et al. 1976; Howell and Denton 1974; Howell et al. 1975). Recent studies of nucleoli and NORs in fishes (Dingerkus 1977; Dingerkus et al., unpublished data) show that di- ploid species have 2 NORs, usually with two small nucleoli in the interphase nucleus; tetra- ploid species have four NORs, usually with four nucleoli in the interphase nucleus; and species with high amounts of tandemly duplicated genes have two NORs, usually with two very large nucleoli in a large interphase nucleus. Representative photographs of these three states are shown in Figure 3. The barb, Barbus tetrazona, paddlefish, Polyodon spathula, and lungfish, Lepidosiren paradoxa, represent, re- spectively, the diploid, tetraploid, and tandem gene-duplicated states. The interphase nucleus of Latimeria has been shown to contain two small nucleoli, like that of the barb, indicating on this basis that Latimeria is a typical diploid species. Recent studies on NORs in mammals have shown that different species have from one to ten NORs per cell (Goodpasture and Bloom 1975; Yates et al. 1976). Since there is no evi- dence of ploidy among mammals, this variation in NORs cannot be correlated with states of ploidy, as it can in fishes. The variability in NOR numbers among different species of mammals must be due either to a holdover from ploidies undergone in chordate evolution long ago, or to rearrangements in the mammalian genome. The fact that the number of NORs can vary from cell to cell in the same individual tends to indicate that the variable numbers of NORs in mammals results from genomic rearrangement. As more species are studied in this way, we may be able to further understand the significance and evo- lution of NORs. DISCUSSION The following discussion includes a brief re- view of the apparent trends in chromosomal 115 evolution. These have been described and dis- cussed by Ohno, Wolf, and Atkin (1968), Ohno (1970a, 1974), and Koulischer (1973). The primitive chromosome type appears to be the microchromosome. Some microchromo- somes later fuse into acrocentrics. These acro- centrics can then undergo robertsonian fusion to form metacentrics or submetacentrics. Which of these two types of chromosome is formed depends, respectively, on whether the fusing elements are of equal or unequal size. These changes do not necessarily occur in all the chromosomes of a given complement. Hence the karyotype of a species may be a composite of varying degrees of these changes, and may have a mixture of microchromosomes, acro- centrics, submetacentrics, and metacentrics. Since tunicates are the most primitive of living chordates, it is not surprising that they have a primitive karyotype consisting entirely of microchromosomes, and a low level of DNA/ cell. Some of the differences in chromosome number among tunicates can be explained by fusions to account for lower numbers. On the other hand, incidents of tetraploidy may ac- count for the higher ones. I believe that the basic tunicate chromosome number is 20 to 24. Amphioxus has almost twice as many chromo- somes, and one might conclude that a tetra- ploidy has occurred. If this is true, amphioxus should have twice as much DNA, but it actually has more than double the DNA/cell value of tunicates. Hagfish, with their chromosome range of 46 to 52, may have arisen from the basic amphioxus karyotype by the way of another tetraploidy. Their number is somewhat lower than expected, but some of their chromosomes are much larger than the others, and this would indicate the fusion of some of the original chromosomes. If this hypothesis is true, the DNA/cell of hagfish should be approximately twice that of amphiox- us. It is, however, much higher than that. The lampreys of the northern hemisphere have a very high number of chromosomes, and this could be due to a tetraploidy of the basic hagfish karyotype. Southern hemisphere lampreys, however, have approximately half this number, that is, in the range of the hagfish. Chromosome morphology reveals that hagfish have small ac- rocentrics to microchromosomes, while south- ern hemisphere lampreys have metacentrics that are roughly twice as large as the acrocentrics of 116 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, No. 134 their northern hemisphere relatives. The south- ern hemisphere lampreys probably arose from a karyotype like that of the northern hemisphere lampreys by fusion of all the acrocentrics into metacentrics. The DNA/cell of southern hemi- sphere lampreys has not been measured, but I predict that they will be found to have approx- imately the same number of DNA/cell as do the northern ones. The latter have a DNA/cell con- tent that is roughly four times that of amphioxus. One might speculate that the hagfish karyotype originally arose by tetraploidy from a basic am- phioxus karyotype, that this basic hagfish karyo- type then gave rise to a basic lamprey karyotype through another tetraploidy and that after this event, the hagfish lineage underwent consider- able tandem gene duplication. The common lam- prey karyotype eventually gave rise to the southern hemisphere lampreys by robertsonian fusion of the acrocentrics. A test of this hypoth- esis would be to determine whether hagfish ex- hibit a high level of ribosomal gene multiplicity, since this would indicate a high level of tandem gene duplication. Some indication that this is in- deed the case has been provided by the fact that Ohno and Morrison (1966) found that one species of hagfish has four independent gene loci for monomeric haemoglobin polypeptides, while Adinolfi et al. (1959) found only two in the lam- prey Lampetra planeri. Another possibility is that after the splitting of the lamprey lineage, hagfish underwent a secondary tetraploidy in- stead of tandem gene duplication. The low chro- mosome number of hagfish would tend to argue against this, but such a low number could be due to chromosomal fusions in the hagfish. The holocephalians, or chimaeras, exhibit a mixture of acrocentrics and microchromosomes, with approximately the same amount of DNA/ cell as in the lampreys. Such a karyotype could have been formed by a fusion of some of the small acrocentrics and microchromosomes of the basic lamprey karyotype. This hypothesis finds support in their high number of chromo- somes (which is, however, less than the number found in the lampreys), and in the size of their acrocentrics (which are proportionately larger than those of the lampreys). The holocephalian karyotype, however, could also have arisen from a basic hagfish karyotype by way of a tet- raploidy, followed by secondary chromosomal fusions. Further biochemical studies and chro- mosome banding patterns may shed light on this problem. The Elasmobranchia (skates, rays, and sharks) have approximately twice as much DNA/cell as do the chimaeras. This may indicate that a tet- raploidy occurred between the basic holoce- phalian and elasmobranch karyotypes. This is supported by the rajid karyotype which contains approximately twice as many chromosomes as in the chimaeras and consists mostly of acro- centrics and microchromosomes. The dasyatid karyotype, which has a lower number of chro- mosomes than rajids, but has a greater propor- tion of metacentric chromosomes seems to have arisen from a basic chondrichthyan karyotype through robertsonian fusion of some of the ac- rocentrics. The torpedinid lineage appears to have undergone a greater amount of chromo- somal fusion since they have a low number of very large, mostly metacentric-submetacentric chromosomes. Unfortunately there are karyo- type data on only two species of sharks. Etmop- terus has a karyotype very similar to that of a rajid but with fewer chromosomes. The karyo- type of Squalus has a low chromosome number with metacentrics, no true microchromosomes, and some very small acrocentrics. This karyo- type could have developed by the fusion of some of the elements of an Etmopterus karyotype, and Etmopterus probably more closely repre- sents the basic shark karyotype. Preliminary studies on the chromosomes of Ginglymostoma cirratum, Sphyrna tiburo, and Negaprion bre- virostris Support this (Dingerkus, unpublished data). I hypothesize that rajids and some sharks possess a karyotype most like the basic elas- mobranch karyotype. Chromosome banding patterns again may show which of these is most closely related to the holocephalian lineage. Some elasmobranchs have fairly high amounts of DNA/cell (up to 9.8 pcg). This would suggest that during their evolution these forms under- went secondary tandem gene duplication follow- ing their tetraploid ongin. The Actinopterygii (ray-finned fishes) may be divided into three main groups: Chondrostei, Holostei, and Teleostei. The chondrosteans have a very high chromosome number, including metacentrics, submetacentrics, acrocentrics, and a high number of microchromosomes. It has been argued that they have a tetraploid origin (Dingerkus and Howell 1976). This is upheld by DINGERKUS: CHORDATE CYTOGENETIC STUDIES the presence of four nucleoli per interphase cell in Polyodon (Dingerkus 1977). Since chondros- teans and elasmobranchs have approximately the same amount of DNA/cell, however, it can be argued that the basic chondrostean karyotype did not arise by tetraploidy from a chondrich- thyan karyotype. Instead, the lineage probably arose by tetraploidy from some sort of holoce- phalian-like karyotype. Since the chondrosteans already have a number of metacentrics among their large number of chromosomes, and these metacentrics occur in sets of four, it may be concluded that the karyotype from which the chondrostean karyotype arose already pos- sessed metacentrics and was already more ad- vanced than the basic holocephalian karyotype. In some species of sturgeons, a chromosome number is found that is twice that of other stur- geons. These are almost certainly due to a tet- raploidy of the already tetraploid sturgeon karyotype, and therefore they could be consid- ered as octoploids. The holostean gars, with a lower number of chromosomes than chondros- teans, originated from a karyotype in which chromosomes must have been lost and/or fused. The lowered DNA/cell level found in holosteans agrees with this hypothesis. Gars exhibit only two nucleoli per nucleus (Dingerkus 1977), and this together with their lower chromosome num- ber and DNA/cell level, indicates that a diploid- ization of the basic tetraploid chondrostean karyotype gave rise to the holostean karyotype. Amia, with a still lower number of chromo- somes, no true microchromosomes (the last pair of very small acrocentrics might be argued to be microchromosomes), and less DNA/cell, seems to have arisen from such a basic holostean karyotype and to have undergone even more chromosome losses and fusions than have the gars. At first glance, the teleosts appear to be ina chaotic state with regard to chromosome num- bers and DNA/cell values. With further obser- vation, however, distinct trends can be ob- served. Ohno (1970a) has attributed this variation as “‘a reflection of nature’s great experiment with gene duplication.”’ I tend to agree with Ohno’s statement, but would add gene deletion to duplication. Evidence from both karyotypes and DNA/cell values has shown that numerous species and lineages have either undergone or arisen through tetraploidy. Uyeno and Smith 117 (1972) showed that the catostomids arose, as a lineage, through tetraploidy from a cyprinid ancestor. Ohno et al. (1967) showed that nu- merous cyprinids have evolved by tetraploidy, most notably Carassius auratus. This is also the case in clupeid and salmonid fish (Ohno et al. 1969b). These cases are confirmed by observa- uions of nucleoli (Dingerkus 1977). Hinegardner and Rosen (1972), based on DNA/cell values, have suggested that ploidy has played an im- portant role in the evolution of various catfish, gobies, and parrotfish. They also have suggested a correlation between more specialized species and lowered levels of DNA/cell. I believe this lower level of DNA/cell is probably correlated with a reduction in number of chromosomes. Since the teleosts are such a diverse and spe- close group, it is not surprising to find that both of these mechanisms appear to have been used widely in evolving the present species of te- leosts, since we can account for species of te- leosts with high chromosome numbers and DNA/cell levels as well as species with low numbers of chromosomes and DNA/cell levels. Somewhere between the extremes must lie the basic teleost karyotype. No species of teleost possesses microchromosomes, and microchro- mosomes must have been lost in the transition between holosteans and teleosteans. Ohno (1970) argues that the basic teleost karyotype is 48 acrocentric chromosomes, mainly because this configuration is widely found among the te- leosts. If we look at the karyotypes of the os- teoglossids, which are currently viewed as the most primitive living teleost group (Patterson and Rosen 1977), we see that they have a higher chromosome number, with metacentrics present in some species (Uyeno 1973). Neither DNA/cell nor nucleoli studies indicate any tetraploidy (Hinegardner and Rosen 1972; Dingerkus 1977). I think that Osteoglossum with 56 chromo- somes, mostly acrocentrics, probably exhibits a karyotype that is most like the basic teleost karyotype. From such a karyotype, the other lower-numbered karyotypes, which in many cases include numerous metacentrics, probably arose by chromosomal rearrangements, fusions, and deletions. The anguilloid eels and clupeids, which have an apparent basic number of 48, probably underwent chromosomal fusion and deletion. This is indicated by similarities in DNA/cell values among osteoglossids, anguil- 118 AAAADAARAAAAAd ddd d ad Od aaa AEA eee oco ODO cic Badd ddddddddddddddaaa eee eee eee ee ee ee | eee eee ee ee ee ee | Bebb bdbdbdbddbdbddbrdbdbaa eee ee ee © ee ee ee eee eee ee ee ee ee ey Ce ee ee ee | Ichthyomyzon gagec, 2N = 164 on ee od | P< P< oe seed =e e+ ee mee

= wn 2 a oO = < 2 = xt = — - < - oO lu 00 = iS) oy 60 Q i) ~ a. va) oN y “So a. Coal cq ! ~N w . ' ™ ' a) N im _ = FIGURE 4. discussed in the text. tinguishable from the extant elasmobranchs, chondrosteans or polypterids. It is apparent that reliable descriptions of the banding patterns of fish chromosomes will be essential to enable us to reconstruct the evolu- tion of fish karyotypes. It is hoped that the nec- essary techniques will soon be developed. ACKNOWLEDGMENTS I would like to thank Drs. W. Mike Howell, Donn E. Rosen, James W. Atz, Gareth J. Nel- son, Lowell D. Uhler, Charles H. Uhl, Michael D. Lagios, William L. Brown, Jr., John E. McCosker, C. Lavett Smith, Howard E. Evans, and Perry W. Gilbert for assistance, advice, and enlightening discussions in the preparation of this paper. I would also like to thank Celeste Roman, Peter Spencer, Thomas Lindsay, Larry Herbst, Frank Bartalone, Bernice Churnetski, Tina Segalla, U. Erich Friese, Weysan Dun and my mother, Lieselotte Dingerkus for assistance in the preparation of this manuscript. 125 SARCOPTERYGI I oD 3) 0 * — Soe Gar mon 2 i=) iss —¢ wn x WwW = ol OV uu ACTINOPTERYGII S Va) a . io) i=) = D ae eee us =< D — = = 2 2 op) [) > a wo ve a io) = = uw Zz < On oe 22) O — = ay e> Q io) = o fo) w os z n z fa) “Zo =) ss = w S| a s) — x aie = WwW Oo fo) = ao we oO oO B. Amphibians SS To sol SS SESS. PRN O SESS X [a4 SSS NS Ss le St eae ~ Lizard a SESS w ie Ne ~S S = SOS rr if NSSISES chick ; faa) SS ASAA rekon C. Reptiles SO \ AS Turtle and ro) b \ sS Crocodile Birds = \ puck ray 7 30;-— =a) L (= Z L Ww S li Opossum tu -20;- RGH |_ Standard Sea Ox Gui Goat L b Horse 10;— ESS i | Se It l al a 2 8 32 125 500 2000 myg equivalent of RGH FIGURE 2. relatedness of GH in pituitary extracts from representatives Composite diagram showing immunochemical of various vertebrate classes based on radioimmunoassay with antiserum to RGH. (From Hayashida 1970, Gen. Comp. En- docrinol. 15:432, by permission of the editors.) (salmon and striped bass) gave no significant stimulation at comparable and higher doses. I]. COMPARATIVE IMMUNOCHEMICAL STUDIES Before going on to the discussion of results obtained in bioassays with PEs of other primi- tive fishes I would like to review the results of immunochemical studies which we were con- ducting at the same time using the same PEs employed in the above bioassays. In contrast to results obtained in our earlier investigations with the use of rabbit antisera to purified mammalian GHs in which the findings suggested the existence of a high degree of im- munochemical species specificity for GH even among mammals themselves (Hayashida and Li 1958a, 1958b and 1959), the subsequent use of antisera prepared in rhesus monkeys gave us a very useful tool for crossing species barriers (Hayashida and Contopoulos 1967; Hayashida 1970). A. Immunodiffusion Studies The results of the immunodiffusion studies employing monkey antiserum to purified rat GH 131 DaDDLetisH 6 ween SALMUN. FIGURE 3. Ouchterlony plate showing immunochemical relatedness of GH in PEs of primitive bony fishes to rat GH and the lack of detectable relatedness of GH in PEs of modern bony fishes to RGH. RPE indicates 0.3 mg equivalent of an adult male rat pituitary. The sturgeon and paddlefish PEs rep- resent 1.0 mg and 1.5 mg of tissue, respectively, while the salmon and mackerel PEs represent 2.0 mg of tissue each. Photographed at 24 hours. (From Hayashida 1970, Gen. Comp. Endocr. 15:432, by permission of the editors.) are shown in Figure |. These studies were car- ried out with crude extracts of pituitaries rep- resenting each of the major vertebrate classes. In the first agar plate we can see that PEs of species representing several different orders of mammals show reactions of identity with each other and with the purified rat GH reference. In the second plate the PEs of avian species gave reactions of identity with each other, but reac- tions of partial identity with respect to the rat GH indicated by the formation of spurs, sug- gesting that avian GHs do not share all of the immunoreactive determinants possessed by rat GH. In the next plate we see that an avian PE shows a reaction of identity with PEs of two different reptilian species (snapping turtle, Chel- ydra serpentina and a crocodilian, Caiman sclerops), and that an amphibian PE (mud pup- py, Necturus) shows a reaction of partial iden- tity compared to the avian and reptilian species as well as with respect to rat GH. In the fourth plate of this series we again see that the Nec- 132 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, No. 134 turus PE gives a reaction of partial identity with rat GH and that another amphibian PE (bullfrog, Rana catesbeiana) almost gives a reaction of identity with the Necturus PE. On the other hand, the lungfish PE (Protopterus aethiopicus ) as well as the teleost PEs (striped bass, Morone saxatilis, and carp, Cyprinus carpio) gave no detectable reaction with this technique although higher levels of the PEs were employed. B. Radioimmunoassay Studies In the radioimmunoassay (RIA) studies, high- ly purified rat GH was labeled with I'*° (Green- wood et al. 1963) and used as the labeled hor- mone as well as for standards. Preliminary tests had shown that unlabeled, purified rat GH would competitively inhibit the binding of the labeled hormone with the antibody to rat GH. The avail- able purified GHs (mammalian species only) as well as the GHs in the PEs from the various vertebrate classes were then employed to test their relative abilities to compete in this system (Hayashida 1969, 1970). The results showed (Fig. 2) that all mammalian GHs or PEs were the most efficient as indicated by the steepness of their inhibition slopes, and that PEs of reptiles and birds were the next most efficient while those of amphibians were still less efficient and those of fishes including the African lungfish and two teleost species showed the least ability to compete in this system. The PEs of two other teleost species including the mackerel (Scember scombrus) and the carp (Cyprinus carpio) were also found to be virtually inactive in this system. The GHs or PEs thus fell into major immuno- chemical groupings which corresponded to the results obtained by immunodiffusion and cor- related well with their major phylogenetic clas- sifications. C. Studies with Chondrostean PEs (Sturgeon and Paddlefish) Immunodiffusion studies employing the mon- key antiserum to rat GH (Fig. 3) showed that purified rat GH and a rat PE gave reactions of identity, while the PE of the sturgeon (Acipenser transmontanus) gave a Clear line of partial iden- tity with rat GH or PE. Paddlefish (Polyodon spathula) PE showed a slightly weaker reaction than that of the sturgeon, but a reaction of iden- tity with the latter. The two teleost PEs again were negative. The results of RIA were in ac- cord with those obtained by immunodiffusion. yg Pituitary Tissue (0632S 1.0 40 15.8 64.0 60 rae Mackerel Salmon 50 Paddlefish ro) Sturgeon ae O SAO uw O O 2 fa) FA Si0)r Para) — Zz Ww U a mm 20); ——- / SQUALUS & PIVSTELUS / FIGURE 5. skate SKaTe IL SQ Musrexus SQUALUS SQUALUS ste Ouchterlony plates demonstrating the reaction of GH in pituitary extracts (PEs) of elasmobranchs with monkey antiserum to rat GH. Antiserum (0.15 ml) is located in the center well. Each of the Squalus and Mustelus extracts represents the extract from 1.5 mg of tissue, and the skate extract represents 2.0 mg of tissue in each case. All extracts in 50 ul of phosphate-saline buffer. Photographed after 48 hours of immunodiffusion at room temperature. (A) RGH indicates highly purified rat GH (2.7 USP units/mg, 7.5 wg in SO wl volume). Note clear reactions of partial identity with RGH shown by the Squalus and Mustelus PEs (GHs). (B) Note reactions of qualitative identity between the PEs (GHs) of Mustelus and Squalus. (From Hayashida 1973, Gen. Comp. Endocrinol. 20:377, by permission of the editors.) daily injection of antiserum to rat GH, complete- ly blocked the tibial plate stimulating activity of both shark and skate PEs, indicating that this biological activity of the elasmobranch PEs was due to GH or a GH-like substance present in the extracts, which was immunochemically related to mammalian GH (rat). Immunodiffusion studies (Fig. 5A, B) re- vealed the presence of a substance in PEs of the sharks which was partially related to rat GH, although the skate PE did not show a detectable precipitin line. The results of RIA (Fig. 6) with these extracts indicated that the shark PEs con- tain a GH or GH-like substance which shows a low, but significant degree of relatedness to rat GH, while the skate PE showed the weakest cross-reaction as revealed by its even lesser slope. Thus the results of RIA again correlated well with those obtained by immunodiffusion. The results of all of the studies reported thus far have shown that PEs of all existing primitive fishes which we have studied to date including the African lungfish, elasmobranchs, and chon- drosteans, are capable of significantly stimulat- ing growth in the mammal on the basis of the rat tibia assay, whereas the PEs of several different species of teleosts were found to be incapable of stimulation. Moreover, immunochemical tests have indicated that the PEs (GHs) from the above primitive fishes show a significant degree of immunochemical relatedness to tetrapod GHs while the PEs (GHs) of the several species of teleosts studied, show little or no such related- ness at least on the basis of the methods em- ployed. B. Holosteans Holosteans would be of particular interest in these studies since they are generally considered to be intermediate in their level of development toward the teleosts. Extracts were prepared from the pituitaries of the bowfin (Amia calva) and the gar (Lepisosteus spp.) which included both long and short-nosed gars. Immunodiffu- sion tests showed that PEs (GHs) of both of these holosteans gave reactions of partial iden- tity with rat GH, with that of the gar being slight- ly weaker than the precipitin line due to the bowfin. Both PEs, however, gave reactions of identity with each other (Hayashida 1971). The results of the rat tibia assay indicated that the bowfin PE (GH) was capable of significant stim- ulation in this assay although, perhaps, slightly less active than that due to equivalent amounts of PEs of the shark or sturgeon. Again, RIA revealed that both holostean PEs (GHs) showed HAYASHIDA: GROWTH HORMONE STUDIES sg Pituitary Tissue 0 16 64 250 1000 lial 50t— x e@ O (B% 4 4—4 Skate = e e im 40 —~e Mustelus [=a} < = uw e Squalus O b SS Mustelus Squalus O 30}— Zz L {a) Za te ray — 7h. 20}— Ww U i= e ao jee) é I Xe 10 ee ° RGH @—»— 6 Standard JRE Ue Jay | 16 64 250 1000 4000 16006 mg RGH standard FiGuRE 6. Radioimmunoassay with monkey antiserum to rat GH (RGH) showing the relatedness of GH in pituitary extracts of Squalus, Mustelus, and Raja (skate) to each other and to RGH. Two different pituitary extracts are represented for each species of shark. Note that the skate extract (GH) shows the least cross-reaction indicated by its lower slope. (From Hayashida 1973, Gen. Comp. Endocrinol. 20:377, by permission of the editors.) low but significant relatedness to rat GH with their slopes approximating that of sturgeon PE observed in previous studies. IV. RECENT USE OF MONKEY ANTISERUM TO TURTLE GH In our more recent studies carried out in col- laboration with Drs. Susan Farmer and Harold Papkoff of the Hormone Research Laboratory (Hayashida et al. 1975; Farmer et al. 1976b) it was found that an antiserum prepared in the monkey against GH from the snapping turtle (Chelydra serpentina) representing a species phylogenetically lower than the mammal, con- siderably broadened the extent of cross-reactiv- ities observed with the GH antiserum. To illus- trate, in immunodiffusion studies GHs of reptiles and birds now gave reactions of identity 135 or near identity with those of mammals (Fig. 7A, B, C). However, GHs or PEs of species below the reptilian level (turtle), continued to give re- actions of partial identity in a “‘step-laddering”’ fashion as we had previously observed with the use of antiserum to rat GH. For example, we can see in Fig. 7D that sturgeon PE (GH) (Aci- penser transmontanus) gave a reaction of partial identity with turtle PE, and shark PE (Mustelus canis) or purified shark GH (Prionace glauca) (U. J. Lewis) in turn gave a reaction of partial identity with GH in sturgeon PE. Bowfin (Amia calva) PE also gave a reaction of partial identity with turtle PE, while the PE of the teleost rep- resented by the striped bass (Morone saxatilis) was again negative. Radioimmunoassay was performed to test these and other PEs employing the turtle GH serum and using rat GH as the labeled hormone and for standards. The results were in complete accord with those obtained by immunodiffusion. All mammalian, reptilian and avian GHs and PEs tested (Fig. 8A, B) gave slopes of inhibition which were parallel to the rat GH standard, while an amphibian GH preparation (bullfrog, Rana catesbeiana) (Fig. 8B) gave a curve with a lesser slope and PEs of the sturgeon and bow- fin (Fig. 8A) gave curves of even lesser slope, while the shark PE showed a still lesser slope. A highly purified shark GH preparation (U. J. Lewis) gave a slope that was essentially identi- cal to that shown by the shark PE. Both the monkey antiserum to snapping turtle GH as well as the antiserum to shark GH were found to be capable of completely neutralizing the rat tibial plate stimulating activity of shark PE (Hayashida and Lewis, unpublished obser- vations). V. STUDIES WITH COELACANTH AND RATFISH PEs In a recent investigation we had the oppor- tunity to test the PE of a coelacanth (Latimeria chalumnae Smith) in the rat tibia assay and de- termine whether or not it contains a substance immunochemically related to tetrapod GHs uti- lizing monkey antiserum to turtle GH (Hayashi- da 1977). The fish was a gift from the Comoran government to the California Academy of Sci- ence Coelacanth Expedition of 1975 and is iden- tified as Latimeria No. 79, CAS 33111. It was a relatively large male specimen which mea- sured 110 cm in length and weighed 30 kilos. 136 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, No. 134 ALLIGHIOR2 * @*~ FIGURE 7. } PGT !/ BGA e °o. o \ SOHO. PE POSS: PE | Immunochemical relatedness of purified somatotropins or somatotropins in pituitary extracts from various ver- tebrate species based on immunodiffusion tests. Monkey antiturtle somatotropin serum (0.15 ml) is in the center well of each agar plate. The amounts of pituitary somatotropin (GH) in (A) are 10 ug for snapping turtle (T); 10 wg each for bovine (B), porcine (P), and whale (W); 7.5 wg for rat (R). The amounts in (B) are 10 wg each for T, B, P. The amounts in (C) are 40 pg for monkey (M), 15 ug each for chicken and duck. The amount in (D) is 10 wg for shark (Prionace glauca) somatotropin. All pituitary extracts (PE) are the equivalent of 2.0 mg wet tissue weight per well in all agar plates. All peripheral wells contain antigen in a volume of 0.05 ml, except those in Fig. 7C, in which the volumes are 0.2 ml each. The plates were photographed at 43-46 hour incubation, except Fig. 7A, at 24 hr. (From Hayashida et al. 1975, Proc. Nat. Acad. Sci. 72:4322, by permission of the editors.) The whole gland weighed 60 mg not including the long, slender rostral extension described by Lagios (1975). The results of the rat tibia assay (Table 5) in- dicated that the coelacanth PE was indeed ca- pable of showing significant stimulation in this assay. Only one dose level of the extract was employed utilizing only 2 rats due to the short- age of the precious material. Because of this, the relative potency of the extract cannot be ex- pressed in terms of the ovine GH used for stan- dards. However, if we assume that the dose-re- sponse for this PE were parallel to that of the standard, had enough material been available, the degree of response shown by the 5.0 mg dai- ly dose of the PE would be equivalent to that induced by about 5 to 6 ywg/day of ovine GH. Immunodiffusion studies employing the mon- key antiserum to turtle GH (Fig. 9) showed that sturgeon PE gave a reaction of partial identity HAYASHIDA: GROWTH HORMONE STUDIES 137 Microgram pitutary tissue Os O66 25 O AO Ib @ 20 i T T T | yl | 60 = B A. 530) |= ac SS & 40h 2 S e . = S = 5 o 2 N Shark GH (Lewis) Oo (Prionace glauca) o) aN Zz oye ® Shark PE a [a) ® (Mustelus) Z Zz fos a be g \ : & O U +— Monotreme PE p a fs 2AdiF \ + (Echidna) Bowfin PE ne a 4 Amphibian GH = (Amia) 4 (Bullfrog) A Reptilian PE Ovine GH (Alligator) O Sturgeon PE + (Acipenser) Bovine GH Marsupial PE—> ‘\ ae (Opossum) ~~ Mammalian GH — <+— Mammalian GH (RGH Standard) (RGH Standard) ee Nl Mt ya | al | | J Jie Se ieee = | J OD5105alieeza 4 16 64 250 1000 4000 OWPHOS Ww 2! 16 64 250 1000 4000 Nanogram equivalent of RGH Nanograms of purified GH FiGuRE 8. Radioimmunoassay with monkey antiserum to turtle somatotropin (GH) showing immunochemical relatedness between purified somatotropins in pituitary extracts (PE) from various vertebrate species. Rat somatotropin (RGH), used as the reference standard, competitively inhibited the binding of '*'I-labeled RGH as shown by the standard curve. (A) Results obtained with pituitary extracts compared to the RGH standard. (B) results obtained with purified somatotropins compared to the RGH standard. Note that all mammalian, reptilian, and avian somatotropins or pituitary extracts gave the steepest inhibition slopes, which paralleled or nearly paralleled the RGH standard curve, and that somatotropins or pituitary extracts of an amphibian and relatively primitive fishes gave curves with lesser slopes, which showed significant relatedness to those of the phylogenetically higher species. Alligator pituitary extracts and shark somatotropin curves are represented by broken lines. These preparations were run in other assays in which the total binding of labeled RGH and the slope of the RGH standard curves were essentially identical to those shown here, and were therefore entered in this figure to illustrate their relative positions and slopes with respect to the standard curves. (From Hayashida et al. 1975, Proc. Nat. Acad. Sci. 72:4322, by permission of the editors.) with turtle GH and the shark PE in turn gave a__ by the more sensitive method of RIA (Fig. 10), similar type of reaction with sturgeon PE al- the coelacanth PE did show a low but significant though weaker, while both coelacanth and lung- degree of relatedness to tetrapod GHs and a fish PEs gave no detectable reaction. However, closer relatedness to sturgeon PE and a still TABLE 5. TiBiAL PLATE STIMULATING EFFECT OF COELACANTH PITUITARY EXTRACT.# Daily Animal Tibial plate width Group dosage (no.) (um + SE) Significance” 1. Hypophysectomized controls = 8 193235-= S259 2. Ovine GH 7.0 wg 6 244.8 + 6.20 I vs. 2 p < 0.001 3. Ovine GH 14.0 pg 4 268.1 + 1.74 4. Ovine GH 28.0 wg 4 302.6 + 5.11 5. Coelacanth PE 5.0 mg 2 23329 = 1B ALO I vs. 5p < 0.01 4 The assay was performed in female Sprague-Dawley rats, hypophysectomized at 28 days and used beginning 9 days postoperatively. » Significance of difference between mean values in terms of p values of Fisher. From Hayashida (1977), Gen. Comp. Endocrinol. 32:221, with permission of editors. 138 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, No. 134 closer degree of relatedness to shark and lung- fish PEs. Striped bass and ratfish (Hydrolagus colliei) PEs showed only a very smali degree of cross-reaction based on the relative flatness of their curves with respect to those of tetrapods. Ratfish PE has shown certain significant dif- ferences from shark PE by both bioassay and by immunochemical studies. When tested in the rat tibia assay ratfish PE showed only a slight, al- though statistically significant degree of stimu- lation (Table 6), whereas in the case of shark PEs (Squalus acanthias and Mustelus canis) it was previously demonstrated that they are ca- pable of showing greater stimulation at compa- rable dosage levels of PE (refer to Table 4). In immunodiffusion studies the ratfish PE did not show a precipitin line with monkey antiserum to rat GH or turtle GH, whereas the shark PEs readily did so (Fig. 5). This correlates well with the above findings based on RIA, in which rat- fish PE showed only a small degree of immu- nochemical relatedness to tetrapod GHs, much less than that exhibited by shark PEs or purified shark GH (Fig. 10). GENERAL SUMMARY 1. In confirmation of earlier reports we have found that PEs of several teleost species are not capable of eliciting significant biological activity in the mammal on the basis of the rat tibia assay which is the most widely em- ployed bioassay for detection and measure- ment of pituitary growth hormone (GH). FIGURE 9. munodiffusion with the antiserum to turtle GH. The amounts in each well (50 ul volume) are as follows: porcine and turtle GH, 10 wg each; sturgeon PE, 1.0 mg; shark PE, 1.6 mg; coelacanth and lungfish PEs, 2.0 mg each. Photographed after 48 hr of immunodiffusion. (From Hayashida 1977, Gen. Comp. Endocrinol. 32:221, by permission of the editors.) Ouchterlony plate showing the results of im- 2. In contrast to these findings, PEs of several species of existing primitive fishes were ca- pable of showing significant stimulation in the rat tibia assay. These include PEs of the TABLE 6. TIBIAL PLATE STIMULATING EFFECT OF PITUITARY EXTRACT FROM THE RATFISH.**” Daily Animal Tibial plate width Group dosage (no.) (um + SE) Significance‘ 1. Hypophysectomized controls — 11 Sie =e _ 2. Bovine GH 5.0 ug 8 197.8 + 4.9 I vs. 2 p < 0.001 3. Bovine GH 10.0 ng 4 220.0 + 5.1 2 vs.3 p < 0.01 4. Bovine GH 20.0 ug 8 249.0 + 3.8 3 vs.4p < 0.01 5. Ratfish 0.6 mg 4 Wop) IL se 720) 6. Ratfish 1.2 mg 8 17O6)-=31 7. Ratfish 2.5 mg 8 Li Nee 220 I vs. 7 p < 0.001 8. Ratfish 3.7 mg 8 177.4 + 4.1 Il vs. 8p < 0.01 4 Ratfish (Hydrolagus colliei). ® This assay was performed in female Long-Evans rats, hypophysectomized at 27 days of age, and utilized beginning 8 days postoperatively. The assay was conducted on two separate occasions. Since the results obtained in each of the corresponding groups were quite similar, the data were combined in the above table. © Significance of the difference between mean values in terms of p values of Fisher. From Hayashida (1977), Gen. Comp. Endocrinol. 32:221, with permission of the editors. HAYASHIDA: GROWTH HORMONE STUDIES MICROGRAMS PITUITARY TISSUE 06395-25) O40 116) 64) 250 a T T T T =| o = g oO 8 “N& ~ 40 < oa B COELACANTH PE a w = a s = 30+ a = = ; Ne N oO Ze BOVINE GH OQ 20- a B LUNGFISH PE eZ eS a = SHARK PE Zz STURGEON PE si SN & ob Is © BULLFROG GH w © RAT GH STANDARD (ae 4 = — sill : am) OSI 24 16 64 250 1000 4000 NANOGRAMS OF PURIFIED GROWTH HORMONE (somatotropin) 139 MICROGRAMS PITUITARY TISSUE 063.25 10 40 16 64 250 1000 f T T Ts T Th T 1 509 © “ol 8 Pe RATFISH PES e \ .. < SS [oa x a 8 COELACANTH PE = a 30 i < = N © 20+ Zé a Ze faa) — rq Oe cS) NG ira] ALLIGATOR PE a ©. Ne RATGH STANDARD Jt} i's 1 4E L — a= =| 225) Selina 4 16 64 250 1000 4000 16000 NANOGRAMS OF PURIFIED RAT GH (somatotropin) FiGure 10. The immunochemical relatedness of a substance in pituitary extracts of the coelacanth and other existing primitive fishes with respect to GH of tetrapods based on radioimmunoassay. The antiserum consisted of a monkey anti-turtle GH serum, employed at a final dilution of 1:90,000. (From Hayashida 1977, Gen. Comp. Endocrinol. 32:221, by permission of the editors.) chondrosteans, elasmobranchs, a dipnoan, holosteans and the only surviving member of the Crossopterygii, the coelacanth. All of these PEs gave dose-response slopes that were less steep than that due to mam- malian GH standards, with the exception of lungfish PE which gave a slope similar to mammalian GH. No slope was established for the coelacanth due to insufficient mate- rial. 3. Immunochemical studies utilizing immuno- diffusion in agar gel and RIA employing mon- key antisera to rat GH or turtle GH have demonstrated significant immunochemical relatedness of GH or a GH-like substance in the PEs of the primitive fishes, with respect to GH of tetrapods. Such a substance was demonstrated in the PE of a single coelacanth by RIA employing the antiserum to turtle GH. The results sug- gested that it is more closely related to GH or a GH-like substance in the pituitaries of the African lungfish and the shark, than to that of the sturgeon. 4. By both bioassay and immunochemical test- ing ratfish PE (GH) appeared to show signif- icant differences from shark PE (GH). DISCUSSION In the series of studies reported here both im- munochemical and biological techniques have been employed to examine the relatedness of growth hormones (GHs) and pituitary extracts (PEs) of existing primitive fishes and modern bony fishes, with respect to highly purified pi- tuitary GHs of tetrapods. The ability of GH or a GH-like substance in the PEs of primitive fish- es to stimulate growth in the mammal was clear- ly demonstrated on the basis of the tibia assay in hypophysectomized rats. These included PEs of elasmobranchs, a dipnoan (African lungfish), chondrosteans, as well as holosteans. A purified preparation of shark GH has been recently shown to be capable of significant stimulation in this assay (Hayashida and Lewis, manuscript in preparation). On the other hand PEs of several modern bony fishes were all found to be incap- 140 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, No. 134 able of showing such stimulation. However, a highly purified GH from this group of fishes (Tilapia) did show a small but significant degree of stimulation in this assay, but only when a very large dose was employed (Farmer et al. 1976a). The demonstration that antisera to tetrapod GHs (rat or turtle) are capable of completely neutral- izing the rat tibial plate stimulating activity of elasmobranch and sturgeon PEs strongly sug- gests that such activity of the PEs from these primitive fishes is due to a GH or GH-like sub- stance that is immunochemically related to tet- rapod GHs. The immunochemical studies in- cluding the use of immunodiffusion and radioimmunoassay, have indeed shown that there is a substance in the PEs from primitive fishes which is immunochemically related to purified tetrapod GHs. This has proven to be true for an elasmobranch in that PEs of the shark, Mustelus canis, reacted in a manner identical to that of a purified shark GH (Prionace glauca) by both immunodiffusion and radioim- munoassay (Hayashida et al. 1975). The findings obtained thus far with PEs of primitive fishes and modern bony fishes, have suggested that there must be significant differ- ences in the structure of GH from these two groups of fishes. Although recent studies have emphasized the notable similarities in molecular weights, amino acid composition, etc., among the GHs of several species of tetrapods (Farmer et al. 1976b) and the purified GHs of a modern bony fish (Tilapia) (Farmer et al. 1976a) as well as that of the shark (Lewis et al. 1972), closer examination of the amino acid composition data reveal a number of significant differences as well. For example, Tilapia GH contains signif- icantly higher amounts of serine and significant- ly lower amounts of methionine and phenylala- nine than tetrapod GHs; and shark GH shows that the percent composition of at least five out of eighteen amino acid residues is more like that of tetrapods than shown by Tilapia GH, while in another five of the eighteen residues the ami- no acid compositions are similar for shark, Ti- lapia and tetrapod GHs. These findings are very suggestive that structural sequence studies will most likely reveal the existence of significant differences between GHs of modern bony fishes and those of existing primitive fishes and that the GHs from the latter group of fishes will re- semble more closely the GH of tetrapods. The confirmation of these speculations must await the results of amino acid sequence studies which are now only in the preliminary stages. Only when more information becomes available in this area of research can we be in a position to learn more about the determinants responsible for immunochemical and biological relatedness. There are some interesting observations con- cerning relatedness among the PEs of the prim- itive fishes. Although shark and ratfish (Hydro- lagus colliei) are classified as members of the Chondrichthyes, the results of the present stud- ies suggest that these are two quite separate groups. For example, GH in shark PEs could be readily demonstrated by the immunodiffusion technique, employing either antiserum to rat GH or turtle GH, while precipitin lines were never detected with ratfish PEs even at higher concen- trations. Ratfish PEs did show a slight cross-re- action by the more sensitive radioimmunoassay procedure, while shark PEs as well as purifed shark GH, showed a partial but substantial de- gree of cross-reaction (Fig. 8A, B). Moreover, the results of the rat tibia bioassay revealed that elasmobranch PEs including two species of shark and the clear-nosed skate (Raja eglanter- ia) are capable of substantial stimulation al- though the dose-response slopes were very low, while ratfish PE at similar dose levels showed only a slight degree of stimulation and no dose- response slope. As far as the coelacanth PE is concerned the immunochemical findings suggest that the GH or GH-like substance in this PE is more closely related to GH in lungfish and shark PEs than it is to sturgeon PE (GH), and that it is more closely related to sturgeon PE (GH) than it is to the GHs of tetrapods. The results of the bioas- say indicate that coelacanth PE (GH) was defi- nitely capable of stimulating the growth of the rat tibial plate, similar to the results obtained with PEs of the lungfish, shark and sturgeon. Thus immunochemically and biologically one can state that the GH in PEs of existing primitive fishes, including the coelacanth, lungfish, shark and sturgeon, and extending to the holosteans as well, are more closely related to GH of tet- rapods than are the GHs of modern bony fishes. These findings provide additional evidence that the former group of fishes are less divergent from the main line of vertebrate evolution lead- ing to the tetrapods, than are the modern bony fishes which represent by far the most numerous of living vertebrates today. HAYASHIDA: GROWTH HORMONE STUDIES LITERATURE CITED FARMER, S. W., H. PAPKOFF, AND T. HAYASHIDA. 1976a. Purification and properties of teleost growth hormone. Gen. Comp. Endocrinol. 30(1):91—190. F , AND 1976b. Purification and prop- erties of reptilian and amphibian growth hormones. Endo- crinology 99(3):692—700. GESCHWIND, I. I. 1967. Molecular variation and possible lines of vertebrate evolution of peptide and protein hor- mones. Am. Zool. 7:89—108. GREENSPAN, F. S., C. H. Li, M. E. SIMpson, AND H. M. Evans. 1949. Bioassay of hypophyseal growth hormones: the tibia test. Endocrinology 45(5):455—463. GREENWooD, F. C., W. M. HUNTER, AND J. S. GLOVER. 1963. The preparation of '*!-labeled human growth hormone of high specific radioactivity. Biochem. J. 89(1):114—123. HAYASHIDA, T. 1969. Relatedness of pituitary growth hor- mone from various vertebrate classes. Nature (London) 222(5190):294-295. 1970. Immunological studies with rat pituitary growth hormone (RGH). II. Comparative immunochemical investigation of GH from representatives of various verte- brate classes with monkey antiserum to RGH. Gen. Comp. Endocrinol. 15(3):432-452. 1971. Biological and immunochemical studies with growth hormone in pituitary extracts of holostean and chon- drostean fishes. Gen. Comp. Endocrinol. 17(2):275—280. 1973. Biological and immunochemical studies with growth hormone in pituitary extracts of elasmobranchs. Gen. Comp. Endocrinol. 20(2):377-385. 1977. Immunochemical and biological studies with growth hormone in a pituitary extract of the coelacanth, 141 Latimeria chalumnae Smith. Gen. Comp. Endocrinol. 32(2):22 1-229. , AND A. N. CoNntopouLos. 1967. Immunological studies with rat pituitary growth hormone. Gen. Comp. En- docrinol. 9(2):217—226. , S. W. FARMER, AND H. Papkorr. 1975. Pituitary growth hormone: Further evidence for evolutionary con- servatism based on immunochemical studies. Proc. Nat. Acad. Sci. 72(11):4322-4326. , AND M. LaGios. 1969. Fish growth hormone: A bi- ological, immunochemical, and ultrastructural study of stur- geon and paddlefish pituitaries. Gen. Comp. Endocrinol. 13(3):403-411. , AND C. H. Li. 1958a. Immunological investiga- tions on bovine pituitary growth hormone. Endocrinology 63(4):487-497. , AND 1958b. An immunological investigation of human pituitary growth hormone. Science 128:1276— 1277. , AND 1959. A comparative immunological study of pituitary growth hormone from various species. Endocrinology 65(6):944—956. KNoBIL, E., AND J. HOTCHKISS. Annu. Rey. Physiol. 26:47-74. Lacios, M. D. 1975. The pituitary gland of the coelacanth Latimeria chalumnae Smith. Gen. Comp. Endocrinol. 25(2): 126-146. Lewis, U. J., R. N. P. SINGH, B. K. SEAvey, R. LASKER, AND G. PICKFORD. 1972. Growth hormone- and prolactin- like proteins of the blue shark (Prionace glauca). Fish. Bull. 70(3):933-939. OUCHTERLONY, O. 1953. Antigen-antibody reactions in gels: types of reactions in coordinated systems of diffusion. Acta Pathol. Microbiol. Scand. 32:231-240. 1964. Growth hormone. OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES The Biology and Physiology of the Living Coelacanth No. 134, 18 pages December 22, 1979 EVOLUTION OF THE CREATINE KINASE ISOZYME SYSTEM IN THE PRIMITIVE VERTEBRATES By Suzanne E. Fisher and Gregory S. Whitt! Department of Genetics and Development, University of Illinois, Urbana, Illinois 61801 ASTRACT The number of creatine kinase (CK, EC 2.7.3.2) isozyme loci and their differential tissue expressions were determined for species in 72 families in the groups Urochordata, Cephalochordata, Petromyzontiformes, Elasmobranchiomorphi, Crossopterygii, Dip- noi, Chondrostei, Holostei, Teleostei, Amphibia, Reptilia, and Mammalia. Urochor- dates have a single creatine kinase locus. The appearance of the second CK locus in the cephalochordates correlated with a reported threefold higher DNA content in these species compared to certain urochordate species (Atkins and Ohno 1967). All non- teleostean fishes examined have two creatine kinase loci (A and C). The differential patterns of expression of these loci among specific tissues vary greatly among these fishes. The Amphibia and Reptilia also have two homologous CK loci which are also expressed in a variety of tissues. In the successful and diverse teleost fishes a very different strategy of CK gene evolution and isozyme expression has been followed. As many as four functional CK loci are present and the differential expression of these loci has been increasingly restricted to specific cell types. INTRODUCTION There are many approaches to the study of the relationships of the primitive vertebrates. Investigations based on morphological elements can encompass both fossil and extant forms. In- vestigations of embryology may reveal relation- ' Please address all correspondence concerning this manu- script and reprint requests to G.S.W. 142 ships that are not obvious from a study of adult morphology; one example is the many similari- ties of cephalochordates and the ammocoete lar- va of cyclostomes (Romer 1966). Although a comparative biochemical approach is limited to studying extant species, it can be valuable in resolving questions about the systematics of higher taxa (Zuckerkandl and Pauling 1965). Within the primitive vertebrates there are many examples of once species-rich groups being rep- FISHER & WHITT: CREATINE KINASE ISOZYME EVOLUTION 143 resented today by one or a few species. In many cases, and certainly the coelacanth is a prime example, the extant species are morphologically very similar to their ancestors. Much of the work reported in this paper is based on the premise that where there is considerable struc- tural similarity between ancient and modern groups, it is reasonable to expect some conser- vation of the ancient genome structure and reg- ulation as well. Specifically, we are assuming that the number of structural genes, their ho- mologies, and the patterns of regulation of these genes reflect the ancestral condition in much the same way extant species reflect the ancestral morphologies and life histories. Of course mo- lecular divergence has occurred through time because of the inevitable incremental changes in nucleotide sequences resulting in amino acid substitutions in the corresponding proteins (Langely and Fitch 1974; Wilson et al. 1977). While it is not possible to directly determine the biochemical attributes of such extinct ancestral forms as the placoderm or acanthodian fishes, the existence of relic species such as the bowfin, lungfish, and coelacanth provide a unique op- portunity to make phylogenetic inferences about the origins of once numerous taxa. Although the functional macromolecules in the earliest period of molecular evolution were presumably formed by fortuitous and random combinations of smaller units, subsequent mo- lecular and cellular evolution is thought to have been more constrained and to have proceeded by modifications of pre-existing, functionally specific genes. The evolution of a given gene to a new function might be constrained if the older but also essential functions must be retained. This dilemma can be avoided by the formation of duplicate copies of the gene. The concept of evolution by gene duplication has been devel- oped by Susumu Ohno (1967, 1970, 1974) and others (Watts and Watts 1968a, 1968b; Watts, R. L. 1971). Most simply stated, Ohno has pro- posed that the major advance to the vertebrate grade must have required a large increase in ge- nome size. Additional copies of structural genes and regulatory elements were then free to accept mutations and to evolve to new functions. Ohno (1974) has argued that tandem duplications fre- quently produce additional copies of structural genes that are under the control of the same reg- ulatory element, which makes divergence in these duplicate loci more difficult. [However, there is evidence that two closely linked paral- ogous loci such as the Ldh B and C genes in birds (Zinkham et al. 1969) can be under quite separate control.| For this reason as well as evi- dence derived from measurements of DNA con- tent and examination of karyotypes, Ohno has proposed that the amplification of the genome that preceded the radiation of the vertebrates was a polyploidization event. The fact that cephalochordates have haploid DNA contents (Amphioxus lanceolatus 0.60 pg) three times higher than certain tunicate species (Ciona intestinalis 0.21 pg) provides consider- able evidence for at least one major genome am- plification event early in the evolution of the chordate line (Atkin and Ohno 1967). In addi- tion, in several cases among the vertebrates the loci encoding homologous polypeptides are lo- cated on separate chromosomes (Kucherlaptai et al. 1974; Wheat et al. 1972, 1973; Whitt et al. 1976). These observations are consistent with the postulate that a polyploidization event oc- curred early in chordate evolution. Obviously most extant species subsequently have under- gone extensive diploidization. The alternative interpretation, that these homologous loci were originally formed by tandem duplications and have since been placed on separate chromo- somes through chromosomal rearrangements, 1s unlikely in view of the karyotypic conservation observed in many lower vertebrates (Ohno 1974; Morizot et al. 1977; Bush et al. 1977). Isozymes [multiple molecular forms of the same enzyme (Markert and Moller, 1959)], have proved to be powerful tools for studying genome evolution. Those isozyme systems which are the products of homologous gene loci are well suited for studying the evolution of gene structure, function, and regulation (Markert 1975; Markert et al. 1975; Whitt et al. 1975; Fisher et al. 1976). A phylogenetic survey of a multilocus isozyme system allows a partial reconstruction of the probable times of the gene duplication events and the inferred course of their regulatory di- vergence. The significance of evolutionary changes in gene regulation is only starting to be realized. Most vertebrates presumably possess similar arrays of structural genes; the differ- ences between species of vertebrates therefore appears to be primarily due to changes in the timing and the level of expression of these genes during development. Subtle changes in gene reg- ulation can lead to dramatic changes in mor- 144 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, No. 134 phology (Wilson et al. 1974a, 1974b, 1977; King and Wilson 1975). Monitoring the tissue speci- ficity of expression of homologous isozyme loci through phylogeny is a simple and direct method of ascertaining the probable sequential changes in patterns of gene regulation during the course of evolution. The most extensively studied multilocus iso- zyme system is lactate dehydrogenase (LDH, EC 1.1.1.27). A survey of LDH gene expression in the fishes revealed an increase from one to three loci and considerable increases in the spa- tial and temporal specificity of gene expression (Horowitz and Whitt 1972; Shaklee et al. 1973; Whitt et al. 1973, 1975; Markert et al. 1975). A few families of teleosts (Salmonidae, Catostom- idae) have undergone a more recent genome am- plification (tetraploidization) which has resulted in the reduplication of the three LDH loci as well as all other loci (Allendorf et al. 1975; Engle et al. 1975; Ferris and Whitt 1977). The pattern of gene evolution that is inferred from the stud- ies of LDH is that the structure and expression of a duplicate locus is initially identical to the expression of the locus from which it arose. Through time, mutations are accepted at both the structural genes and the DNA regulatory se- quences controlling them, resulting in a differ- ential regulation of the duplicated loci. A more limited survey of glucosephosphate isomerase (GPI, EC 5.3.1.9) gene expression in the bony fishes showed that a gene duplication event probably occurred early in that line and that the higher teleosts have evolved a tissue specific pattern of expression of the two GPI genes (Avise and Kitto 1973). The creatine kinase (CK, EC 2.7.3.2) multi- locus isozyme system provides an excellent op- portunity to study the evolution of homologous structural genes and the evolution of gene reg- ulation. Creatine kinase is a dimeric enzyme with a molecular weight of 80,000 daltons. It cat- alyzes the readily reversible conversion of ATP and creatine to ADP and phosphocreatine (Daw- son et al. 1967; Masters and Holmes 1975). Con- sequently it is very important in maintaining en- ergy balance in a variety of cells. There is recent evidence that certain CK isozymes may fulfill roles in cell structure as well as catalytic roles (Turner et al. 1973; Eppenberger et al. 1975). Three CK loci are expressed in mammals; two loci (which in the past have been called M and B) encode isozymes localized in the cytosol and an additional locus which encodes an isozyme localized in the mitochondria (Masters and Holmes 1975). As many as four CK loci have been detected in the advanced teleost fishes (Eppenberger et al. 1971; Perriard et al. 1972; Scholl and Eppenberger 1972, Champion et al. 1975; Champion and Whitt 1976; Shaklee et al. 1975; Fisher and Whitt 1975, 1977). The CK system is valuable not only because of the proliferation of isozyme loci during the evolution of the fishes and the evolution of high- ly specific patterns of gene expression, but also because the origins of this enzyme can be traced back to a different but related enzyme in the invertebrates. There are clear homologies be- tween the creatine kinase isozymes of chordates and the phosphagen kinases of invertebrates. Watts and others (Robin 1974; Watts, D. C. 1968, 1971, 1975) have reported that inverte- brates utilize a variety of phosphagen kinases, the most common enzyme being arginine kinase (ARG K, EC 2.7.3.3). Coelenterata, Platyhel- minthes, and Arthropoda have monomeric ar- ginine kinases. Monomeric and dimeric arginine kinases are found in different species of mol- luscs. Dimeric creatine and arginine kinase iso- zymes are found in the echinoderms and uro- chordates (Virden and Watts 1966; Moreland et al. 1967). In the echinoderms there is an evolu- tionary progression from some species having only arginine kinase, to those having both argi- nine kinase and creatine kinase, to those having only creatine kinase. The active site peptides of creatine and arginine kinase are very similar (Watts, D. C. 1971). Immunological similarities have been reported between denatured forms of creatine kinase and a monomeric arginine kinase (Robin et al. 1976). Furthermore, it is possible through reversible in vitro denaturation to form a hybrid protein containing a functioning echi- noderm arginine kinase subunit and a function- ing mammalian creatine kinase subunit (Watts et.al. 1972): The present paper is a partial report on a sur- vey of creatine kinase gene expression in 72 families from the echinoderms through mam- mals concentrating, however, on the phylogeny of expression of CK genes in relatively **primi- tive’’ groups of fish—Petromyzontiformes, Elasmobranchiomorphi, Crossopterygii, Dipnoi, Chondrostei, and Holostei. Within these groups, many of which are represented today by a single or a few species, there is evidence of consider- FISHER & WHITT: CREATINE KINASE ISOZYME EVOLUTION 145 able diversity in the patterns of tissue expression of the two creatine kinase loci among the adult tissues. In the two main evolutionary lines de- rived from these primitive fishes—the teleost fishes and the land vertebrates—two very dif- ferent strategies of creatine kinase gene prolif- eration and expression have been observed. MATERIALS AND METHODS Nomenclature ithe icreatme kinase (CK, EC 2-7-3.2) iso- zymes and the corresponding CK loci have been named according to the rules of the IUPAC-IUB Commission on Biological Nomenclature (1973) and previous citations in the literature (Cham- pion and Whitt 1976). The alphabetical desig- nations A, B, C, and D reflect the order in which the loci have been described and not the evo- lutionary order of appearance of the four iso- zyme loci. Preparation of Tissue Extracts All fish used in this study were obtained as fresh or frozen specimens and stored at —20°C until use. Whenever possible ten tissues were examined: white skeletal muscle, heart, eye, brain, stomach muscle, gill, liver, spleen, gonad and kidney. The dissected tissue samples were homogenized in an equal volume of 0.1 M so- dium phosphate buffer, pH 7.0, using a motor- driven Potter Elvehjem homogenizer. The ho- mogenates were centrifuged at 12,000 ¢ for 30 minutes at 4°C and the resulting supernatants recentrifuged at 27,000 g for 30 minutes at 4°C. The final supernatants were used for electropho- resis. Electrophoresis Vertical starch gel electrophoresis (Buchler Instruments Inc., Fort Lee, New Jersey) was performed at 6°C with 15% starch gels (Elec- trostarch lots 371, 302, and 307, Electrostarch Co., Madison, Wisconsin). The buffer system used was the pH 8.6 EDTA-boric acid-Tris (EBT) buffer system de- scribed by Shaklee et al. (1973). 2,000 units of sodium heparin were added to each gel. Elec- trophoresis was performed at 250 volts for 16— 24 hours. Histochemical Staining After electrophoresis the gels were sliced lengthwise and stained according to Shaw and Prasad (1970). One slice was stained using all the reagents for creatine kinase and a second slice from the same gel was stained using all the staining reagents except the substrate phospho- creatine. The omission of phosphocreatine was employed as a control to detect the presence of adenylate kinase (AK, EC 2.7.4.3) activity which appears as an artifact of the creatine ki- nase staining procedure. All staining reagents were obtained from Sigma Chemical Company, St. Louis, Missouri. Purification The details of the purification of the individual CK isozyme will be published elsewhere (Fisher and Whitt, in preparation). Three homodimeric isozymes (A,, B, and C,) of the green sunfish (Lepomis cyanellus) and the A, isozyme of the spotted gar (Lepisosteus oculatus) were purified utilizing gel filtration on Sephacryl S-200, affin- ity chromatography on Blue Sepharose 6B (Easterday and Easterday 1973), and prepara- tive starch gel electrophoresis. Preparation and Use of Antibodies Antibodies were formed against the four pur- ified creatine kinase isozymes in adult male New Zealand white rabbits according to the following schedule: on day 0 purified CK in 0.5 ml of 0.1 M sodium phosphate pH 7.0 buffer was emul- sified in 1.0 ml Freund’s complete adjuvant and then injected into the footpads. Depending on the isozyme used as antigen individual rabbits were injected with as little as 20 wg up to as much as 1,000 wg. On day 10 the same amount of antigen as initially employed was emulsified in 1.0 ml Freund’s incomplete adjuvant and in- jected into the footpads. On day 17 the same amount of antigen dissolved in buffer was in- jected without adjuvant into the thigh and scap- ular muscles. Blood was collected from the mar- ginal ear vein twice a week on a continuous basis starting on day 20 and continuing for about 8 weeks. The antisera were titred according to the enzymatic equivalence point technique de- scribed by Markert and Holmes (1969). The as- say system for creatine kinase was that of Oliver (1955). All enzyme activity assays were per- formed in duplicate on a Beckman Kintrac VII spectrophotometer. Controls for adenylate ki- nase were carried out by omitting phosphocrea- tine. The antisera were concentrated by precip- itation with 33% ammonium sulfate. 146 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, No. 134 PURPLE SEA URCHIN (Arbacia punctulata ) ARGININE KINASE ® Arg K C Aad -e FiGureE |. Cross reaction of the single arginine kinase of the purple sea urchin, Arbacia punctulata with antiserum formed against the A, creatine kinase isozyme of green sun- fish, Lepomis cyanellus. This result is consistent with the pos- tulated origin of creatine kinase from a duplicated arginine kinase locus. The antibodies were used in the immunopre- cipitation plus electrophoresis technique devel- oped by Markert and Holmes (1969). Increasing amounts of antisera were added to aliquots of a given enzyme extract. The enzyme-antibody mixtures were incubated for 2 hr at 4°C, then centrifuged to remove precipitable antibody-en- zyme complexes. The supernatants were elec- trophoresed to remove any remaining enzyme- antibody complexes through the molecular siev- CREATINE KINASE LEATHERY SEA SQUIRT (Styela plicata) Origin Be ae Os oe owe C. x = 2% 7/0 359900 25 85%055%0 Of0K AK Se «KT Ko FiGurReE 2. Reaction of anti CK A,, and anti CK B,, and anti CK C, sera (of similar titres) with the single creatine ki- nase of Styela plicata. Only the anti A, serum cross-reacts, which suggests that the first creatine kinase of the chordates is homologous to the A, isozyme of teleosts. ing action of the starch gel. Electrophoresis and histochemical staining were performed in the usual manner. The reduction of staining inten- sity on the gel as a result of incubation with increasing antibody concentrations is an indi- cation of cross-reactivity. Controls for dilution of enzyme activity by antiserum addition were carried out with normal (pre-immunization) se- rum. The high level of sensitivity of this immu- noprecipitation and electrophoresis technique has been documented and discussed elsewhere (Holmes and Scopes 1974). RESULTS Patterns of differential creatine kinase gene expression were analyzed for over 100 species in 72 families of chordates. Evidence for the ho- mology of the arginine kinase of echinoderms and the creatine kinase of vertebrates is provid- ed in Figure |. Antiserum formed against the purified CK A, isozyme of green sunfish clearly cross-reacted with the single arginine kinase of FISHER & WHITT: CREATINE KINASE ISOZYME EVOLUTION 147 KINASE ISOZYMES OF THE SPOTTED GAR (Lepistosteus oculatus ) CREATINE KINASE . ® ee Origin —» ee ee CO Y Qo, oO “yy dg el og S.9,e8 Y S Wile >» ONCA6O.F 47 7? % an ek FIGURE 3. ADENYLATE KINASE e Sai ve eo a Creatine kinase isozymes of the tissues of the spotted gar (Lepisosteus oculatus). Any bands or smears appearing on both gels are adenylate kinase isozymes. Only the two bands unique to the CK zymogram are creatine kinase isozymes. the purple sea urchin (Arbacia punctulata). Equal volumes of sea urchin enzyme extract were mixed with two levels of antibodies. A di- lution control with normal serum was also per- formed for the highest amount of antiserum added. A comparison of the results in the third (SO ul anti CK A,) and fourth (50 yl normal se- rum) slots reveals that the anti CK A, serum precipitates the arginine kinase. A single argi- nine kinase was also detected in the striped sea cucumber (Thyonella gemnata) and the red- footed sea cucumber (Pentacta pygmaea). The arginine kinase of the sea cucumbers did not clearly cross-react with the antisera formed against any of the creatine kinase isozymes of the sunfish. The urochordates are considered to be the most primitive members of the chordate line. Hemichordates are currently considered a sep- arate phylum that has affinities with both the chordates and echinoderms (Kluge 1977). The urochordate Styela plicata (leathery sea squirt) expresses a single CK locus (Fig. 2). When anti- sera (of similar titres) formed against each of the A,. B,, and C, isozymes of the green sunfish were separately mixed with the tunicate creatine kinase, only the anti CK A, serum cross-reacted with this single creatine kinase. These results suggest that the A locus of higher fishes still re- tains considerable structural homology with the original creatine kinase locus of chordates. The tunicates and perciform fishes last shared a com- mon ancestor more than 500 million years ago and considerable caution is necessary when in- terpreting the specificity of immunochemical re- actions over such large evolutionary distances. The cephalochordates are the most primitive taxon in which the expression of two CK loci has been detected. After histochemical staining of the starch gels three bands of CK activity, 148 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, No. 134 KINASE /SOZYMES OF THE COEWACANTH (Latimeria chalumnae/ CREATINE KINASE FIGURE 4. ADENYLATE KINASE CREATINE KINASE Cs) AK- — Hy * Creatine kinase isozymes of the tissues of the coelacanth (Latimeria chalumnae). The two zymograms on the left show the creatine kinase isozymes (and the adenylate kinase control) in the tissues examined. The bands slightly anodal to the main CK band in skeletal muscle are sub-bands or conformational isozymes. The zymogram on the right shows that antiserum formed against the CK-A, isozyme of green sunfish clearly cross-reacts with the CK of coelacanth skeletal muscle. indicating the presence of two homodimers and a heterodimer of intermediate electrophoretic mobility, were detected for each individual of amphioxus (Branchiostoma floridae) analyzed. Immunological reactions indicate that one locus is more closely related to the A locus of higher fishes and the second locus is more closely re- lated to the C locus of higher fishes. Thus, it appears that the first two creatine kinase loci were those which have evolved to the A and C loci of perciform fishes. The alphabetical des- ignations of the CK isozymes and the corre- sponding loci were made before the evolutionary history of this isozyme system was known. Since the specimens of amphioxus we analyzed were quite small (less than 2 cm in length) they were homogenized whole and we cannot make a definitive statement on the tissue patterns of gene expression in this group. However, there was considerably more CK A, activity than CK CF Analyses of CK isozyme activity among the tissues of fish in the groups Petromyzonti- formes, Elasmobranchiomorphi, Crossoptery- gil, Dipnoi, Chondrostei, and Holostei indicate that two creatine kinase loci are exhibited in all species examined. However, there is consider- able variation among species in the patterns of expression of the A and C loci among the var- ious tissues. The creatine kinase isozymes of the spotted gar (Lepisosteus oculatus) are shown in Figure 3. The staining activity seen in both gels stained for creatine kinase and adenylate kinase represents the adenylate kinase isozymes. Ad- enylate kinase is not a phosphagen kinase and is not directly related to creatine kinase or ar- ginine kinase. Adenylate kinase catalyzes the conversion of 2 ADP molecules to ATP and FISHER & WHITT: CREATINE KINASE ISOZYME EVOLUTION 149 AMP and its presence results in extra stained bands on the gel stained for CK purely as an artifact of the staining procedure used. Only the bands unique to the creatine kinase zymogram are in fact creatine kinase isozymes. Of these two creatine kinases in the gar, the A, isozyme exhibits the greatest activity in skeletal muscle and to a lesser extent in kidney, liver, heart, eye, and gills. The C, isozyme activity was de- tected predominantly in stomach muscle and go- nad and to a lesser extent in eye, brain and kid- ney. A zymogram of the creatine kinase isozymes of the coelacanth tissues (Latimeria chalumnae #79) is shown in Figure 4. Although the stan- dard set of ten tissues we examined for other species was not available, it is clear that the coelacanth has at least two creatine kinase loci. The bands slightly anodal to the main band in skeletal muscle are sub-bands or conformational variants; they have previously been reported for the coelacanth (Hamoir et al. 1973) and are ob- served in many other species. It is difficult to interpret the lack of creatine kinase activity in gill, liver, testis, and kidney. It may be that the creatine kinase isozymes of the coelacanth are labile to heat and/or cold storage and that the CK activity was detected only in those tissues which initially had very high levels of the en- zyme. Alternatively, certain tissues of this fish may possess low or no levels of CK in vivo. However, in most other species we have found that most tissues have detectable levels of at least one of the CK isozymes. The zymogram to the right in Figure 4 shows that the antiserum formed against the CK A, isozyme of green sun- fish cross-reacted with the CK isozyme of coel- acanth skeletal muscle. Antisera formed against the CK C, and B, isozymes of green sunfish do not cross-react with the CK isozyme of coel- acanth muscle, which by inference is the A, CK. A summary of our data on the tissue specific patterns of expression of the CK loci in different taxa of fishes which express two creatine kinase loci is in Table 1. This table documents the di- versity in the patterns of gene expression which exists among species in the primitive fishes. There is not a consistent relationship of relative electrophoretic mobilities of the C, and A, iso- zymes. In the sea lamprey (Petromyzon mari- nus), pallid sturgeon (Scaphirhynchus albus) and paddlefish (Polyodon spathula) the C, iso- zyme is highly anodal. In the gar (Lepisosteus) and bowfin (Amia calva) the C, isozyme is slightly more anodal than the A, isozyme. In some sharks and rays, as well as coelacanth, lungfish (Lepidosiren paradoxa) and reedfish (Calamoichthys calabaricus) the A, isozyme is more anodal than the C, under our electropho- retic conditions. The tissue specificity of expres- sion of the two loci varies considerably among species. The only constant feature of CK gene expression which emerges for all species is that the only isozyme detected in skeletal muscle is the A,. In some chondrosteans (sturgeon and paddlefish) the A, isozyme is detected only in skeletal muscle and not in other tissues. In the reedfish, which is classified in a separate order in the same division, the A, isozyme is present in highest levels in the skeletal muscle but can also be detected in all other tissues examined. Similarly, the C, isozyme in some species can also be quite restricted in its expression among the adult tissues. For example, the C, isozyme is detected only in heart and brain in the lung- fish. In other species such as the sturgeon and gar, the C, is detected in many tissues. The tis- sues in which the C, isozyme activity is most commonly found are eye, brain, stomach and kidney. The conclusion we draw from the data presented in Table 1 is that there is no single characteristic pattern of expression of the two creatine kinase loci in the primitive fishes. Although there is some debate as to the exact ancestral taxa involved it is generally accepted that two separate lines—the amphibians and the teleost fishes—had their origins in ancestors of fish which we have shown currently to possess two creatine kinase loci. However, the number of CK loci and patterns of CK gene expression among adult tissues are very different between these two lines. We have examined the creatine kinase iso- zyme patterns of eight amphibian species; Am- bystoma tigrinum, Amphiuma means, Bufo americanus, Bufo marinus, Necturus maculo- sus, Rana pipiens, Triturus viridescens, and Xenopus laevis. A typical tissue specific cre- atine kinase pattern for amphibian tissues is shown in Figure 5. The CK isozyme pattern of the marine toad (Bufo marinus) indicates that two loci are present. It is probable that the single locus expressed in skeletal muscle is homolo- gous to the A locus of fishes. For the sake of simplicity we will refer to the second locus as the C locus, even though its orthology with the OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, No. 134 150 “Agupry ry ‘usajds dg ‘19al] v7] “[II3 1D ‘yoRwoys 1g ‘ureIg 1g ‘aka Aq ‘yWeay ay ‘ajosnwi np + 7 - — +++ ++ +4 +4 + +44 ++ ++ ++ + F+4+4 + + 7 + +44 ++ = ~ - + + + + + +++ + ++ + — +++ = - — ++ ++ + + + + ++ + + + ++ = + + +++ . - Se ee YI dg Wah tS) WS sndo] D Jo AWANOYy ++ ++ ++ ++ +++ Aq Cis *POUIWIeXd JON quasqy — “AYWANIV MOT + “AWANIE Q]BIPIWAd}Uy + + “AWANIB YBIH + + + AWIGOW [epouy aaneeay WV + ste = P of Staats ot 46 tf ats ar oar apeteeta \/ 0) DA]DO DIU SOULIOJIWY Brees ate = ae J + ae site ak at dete +++ YW<<9O snuojsojvjd snajsosida] eer te = Je it au de + Sets = dels Jit ae V<)O SNASSO SNIISOSIdAT aL = 43 aL ee ae ae a Ete A ++ +++ YW<9 = > = = = a E 2 WY = 2 z om zs m XS s a n* | i Zz he S Z m 2 HIINS SZINVYUGIT_ LIBRARIES SMITHSONIAN INSTITUTION NOILMLILSNI_ NVINOSHLIW! w Zz wn za . Ww ma 4 = = = < ae < a Bs) 5 7 = Bs) SS =) =e 7 =) i oe = ae - Ee a0 ». = fee a cad ee: \ 2 ae z ri z es a JNIAN INSTITUTION NOILALILSNI NVINOSHLINS S3I1YVYGIT_ LIBRARIES SMITHSONIAI Fs Se w = wn z pad: 7) < NE = = = = Ws = = > = af wip ae SS =| : : ; i gp EOS 3 : S - bGdg 2 XX 8 2 zZ = 2 “my = \ 2 sn s S > = > = = > ” S Fa ” z ” he Be HLINS S31YVYGIT LIBRARIES SMITHSONIAN _ INSTITUTION NOILMLILSNI_NVINOSHLIWN w = wn = = N WwW = Ww re WwW = wy So = a fe ac a SB = ivf Ll = 2 = ie S « Wey S c c a 5 a “fe 3 a - he Zz a 2 4 = * ONIAN INSTITUTION NOILALILSNI_NVINOSHLINS S3IYVUSIT LIBRARIES SMITHSONIA 2 z ee z = = a py, wo — . = co SB YY : ly, > 5 Ng 5 w = ly v = ee Oo = HLIINS SAIUYVYSIT LIBRARIES SMITHSONIAN INSTITUTION NOILALILSNI NVINOSHLIM ‘ NVINOSHLINS S31u¥Vvudl N¥INOSHLIWS SMITHSONIAN NVINOSHLIWS wr SMITHSONIAN LIBRARIES SMITHSONIAN ONIAN INSTITUTION NOILALILSNI NVINOSHLIWS S3S1Y¥VYdIt LIBRARIES Yy Ma Z een LIBRARIES NOILNLILSNI LIB RAR PES NOILNLILSNI HLINS S3IYVYUSIT LIBRARIES SMITHSONIAN INSTITUTION NOILNLILSNI NVINOSHLIW HK | p saiuvual INSTITUTION saiuvugiy INSTITUTION Sal1uYvual INSTITUTION SONIAN INSTITUTION NOILNLILSNI NVINOSHLINS S3IYWYSIT LIBRARIES SMITHSONIA ; ) INSTITUTIC Saluvua INSTITUTIC saiuvua INSTITUTIC Saluvua INSTITUTIC g!71 LIBRARIES SMITHSONIAN INSTITUTION NOILALILSNI NVINOSHLINS S3IYVYSIT =z y 2) = ” z w = . N = =] Y Ze > Ne: 2 Ng : a: = GG. = Ue. ON \_ o na ~® NY n 3) ee) n lp Gp D fe a ~~ O pe RAN AW Oo = S My / ale SN \E E AS Z E a Se alee = = eee. > = NN >* = > = ” FS =< ” : z w” Zz a) 1ON NOILNLILSNI_NVINOSHLINS SJ 1YVYdtil LIBRARIES SMITHSONIAN INSTITUTION mi w zs 5 ” i z uJ tid LJ , us ~ as = nh ~ Yin, 2 = oc — ac Le ee a < = y= = < GY yr a < a Ys: 7. S a = 4 [og v Soe = m e fo} = o aa rs) a a4 z= —T me air rs a 8!1 LIBRARIES SMITHSONIAN INSTITUTION NOILNLILSNI NVINOSHLINS S3IYVYSII Saruvugit INSTITUTION sa1uvugiy INSTITUTION INSTITUTION ION NOILNLILSNI S3IYVUGIT_LIBRARIES Soa SMITHSONIAN INSTITUTION s NVINOSHLINS S3IYVUEt NYINOSHIIWS S3aINWNAaIt NVINOSHLIWS Ko SMITHSONIAN NVINOSHLIWS SMITHSONIAN GL) SMITHSONIAN a. i >. ~ git LIBRARIES SMITHSONIAN _INSTITUTION NOILNLILSNI NVINOSHLINS S3IYVYSIT = = 5 ep) > é : a, ul a Coe @ = a = x = - = o z : = : : O = Oo = 5 a 3S z y Fa a Fa wl = TON _ NOMLALILSNI CNVINGSHIINS SSS tave att = ix = Oo - * Oo = 2) = o — w : = oo — ow Te F > We 5 S E ne : = Boo WN i- > ros > = = 2 WAY E es) = es = 2 m \S 2 in ” = a) <= w as = w = wn = G@i1_ LIBRARIES SMITHSONIAN INSTITUTION NOILNLILSNI NVINOSHLINS S31YVYEIT : Ww wn : Fai ate < z z < = WM: ye = 4 = =i = = hify, = 3 NN GE 3 = i 9 & ef 2 hE NY? : 2 - 27k = a > = >” = > = ”) 2 ” Fr “” z ” ION NOILNLILSNI_ NVINOSHLIWNS LIBRARIES SMITHSONIAN INSTITUTION if Za i Ss % Yy, S ay Yn , = a = & ~ €, 5 — < 3 < = & LG fia = = cS > < Se < Y , oe ‘¢ < oe a fa = e P ake = o~ ca fe) a o el rs) pec J z =) =z J re ad di1 LIBRARIES SMITHSONIAN INSTITUTION NOILNLILSNI NVINOSHLINS S31YVYdIT i z ic 2 ee : = a w ‘2 “w S w = ow a = o > a = ra > : > - > - Cat Se) = 6 = ms) rom a = 2) = ” uA ”. 5 iB z o Zz e jue ae TION NOILNLILSNI ~S3IYVYE!IT LIBRARIES SMITHSONIAN INSTITUTION tl | 3 9088 01302 6372 SMITHSONIAN INSTITUTION LIBRARIES il re eee aren babii theta ener wre Mey m cegeerertce eens Nite arn See ae mere eh arene ue AEE EARLE oe Ee SNe Ee ae An Re Sg RN Oe WEA Ry eet eh aerer URS ENON ae 7 ; :