INFORMATION TO USERS
This material was produced from a microfilm copy of the original document. While the most advanced technological means to photograph and reproduce this document have been used, the quality is heavily dependent upon the quality of the original submitted.
The following explanation of techniques is provided to help you understand markings or patterns which may appear on this reproduction.
1. The sign or "target" for pages apparently lacking from the document photographed is "Missing Page(s)''. If it was possible to obtain the missing page(s) or section, they are spliced into the film along with adjacent pages. This may have necessitated cutting thru an image and duplicating adjacent pages to insure you complete continuity.
2. When an image on the film is obliterated with a large round black mark, it is an indication that the photographer suspected that the copy may have moved during exposure and thus cause a blurred image. You will find a good image of the page in the adjacent frame.
3. When a map, drawing or chart, etc., was part of the material being photographed the photographer followed a definite method in "sectioning" the material. It is customary to begin photoing at the upper left hand corner of a large sheet and to continue photoing from left to right in equal sections with a small overlap. If necessary, sectioning is continued again — beginning below the first raw and continuing on until complete.
4. The majority of users indicate that the textual content is of greatest value, however, a somewhat higher quality reproduction could be made from "photographs" if essential to the understanding of the dissertation. Silver prints of "photographs" may be ordered at additional charge by writing the Order Department, giving the catalog number, title, author and specific pages you wish reproduced.
5. PLEASE NOTE: Some pages may have indistinct print. Filmed as received.
University Microfilms International 300 North Zeeb Road Ann Arbor, Michigan 48106 USA St John s Road. Tyler s Green High Wycombe. Bucks, England HP10 8HR I I 78-12,358
LeGRANDE, William Hunt, 1950- CYTOTAXONOMY AND CHROMOSOMAL EVOLUTION IN NORTH AMERICAN CATFISHES (SILURIFORMES, ICTALURIDAE) WITH EMPHASIS ON NOTURUS.
The Ohio State University, Ph.D., 1978 Zoology
University Microfilms International f Ann Arbor, Michigan 48106 CYTOTAXONOMY AND CHROMOSOMAL EVOLUTION IN NORTH AMERICAN
CATFISHES (SILURIFORMES, ICTALURIDAE) WITH
EMPHASIS ON NOTURUS
DISSERTATION
Presented in P a rtial F u lfillm e n t of the Requirements fo r
the Degree Doctor of Philosophy in the Graduate
School of The Ohio State University
By
William Hunt LeGrande, B .S ., M.S.
The Ohio State University
1978
Reading Committee: Approved By
Dr. Ted M. Cavender, Adviser
Dr. Tim M. Berra
Dr. Roy A. Tassava Advi ser Department of Zoology ACKNOWLEDGMENTS
I am grateful to my adviser. Dr. Ted M. Cavender, for his constant
encouragement and assistance from Inception to completion of this study.
The members o f my reading committee, Drs. Tim Berra, Elton Paddock and
Roy Tassava, read the rough draft of this thesis and greatly improved It
with their valuable suggestions. I am Indebted to many of my friends and
colleagues for assistance 1n collecting specimens for this study. I am
particularly grateful to my fellow students at OSU, Larry McGeehan and
Miles Cobum, for accompanying me on several extended field trips. Drs. L.
M. Hardy, R. K. Spealrs and the Biology Club at Louisiana State Univer
sity in Shreveport kindly allowed me to accompany them on a field trip
to the Ouachita Biological Station and assisted me 1n the collection of
Noturus eleutherus and ta y lo r i. Mr. Mike Corcoran o f Duke University
kindly supplied specimens of Ictalurus serracanthus. Dr. R. R. MUler,
University of Michigan, allowed me the use of his laboratory to prepare karyotypes of two Rhawdia spp. supplied by him. Drs, L. M. Hardy
(Louisiana State University in Shreveport) and Mike Fltzslmons (LSU 1n
Baton Rouge) both granted me free use of their laboratories during a portion of this study. Drs. Glen Clemmer, Neil Douglas, David E tn ier,
Robert Jenkins and William PfHeger all provided Information on specific collecting lo c a litie s fo r Hoturus in th e ir respective areas. The fin a l
11 draft of this thesis was prepared during my appointment as Visiting
Instructor in the Department of Biology at the University of Wisconsin at Stevens Point. The support and encouragement offered by colleagues at that institution are deeply appreciated. Support during a portion of
this research was provided through a Dissertation Year Fellowship and
Mary Osburn Memorial Fund Summer Fellowship from the Graduate School of
The Ohio State University. I am grateful for the unlimited use of the collections and facilities of the Division of Fishes at the OSU Museum of Zoology. Lastly* I wish to thank my parents for their constant moral and financial support. Without their manifold sacrifices, this study may never have been completed. VITA
October 11, 1950...... Bom-Shreveport, Louisiana
1956-1968 ...... Stonewall High School, Stonewall Louisiana
1968-1970 ...... Louisiana State University, Shreveport, Louisiana
1970-197 1 ...... Louisiana State University, Baton Rouge, Louisiana
1971 ...... B. S., Zoology, Louisiana State University, Baton Rouge, Louisiana
1971-197 4 ...... Graduate Teaching Assistant,Depart ment o f Zoology and Physiology, Louisiana State University, Baton Rouge, Louisiana
1973 ...... Summer Research Fellowship, Graduate School, Louisiana State University, Baton Rouge, Louisiana
197 4 ...... M. S., Zoology, Louisiana State University, Baton Rouge, Louisiana
1974-1976 ...... Graduate Teaching Associate, Depart ment of Zoology, Ohio State Univer s ity , Columbus, Ohio
1976 ...... Mary Osburn Memorial Fund Sumner Fellowship, Graduate School, Ohio State University, Columbus, Ohio
1976-197 7 ...... Dissertation Year Fellowship, Graduate School, Ohio State University, Columbus, Ohio
1977-presen t ...... V1 sitin g Instru cto r, Department of Biology, University of Wisconsin, Stevens Point, Wisconsin
1 v PUBLICATIONS
LeGrande, W. H. 1975. Karyology o f six species o f Louisiana fla tfis h e s (Ostelchthyes, PIeuronectlformes). Copela 1975:516-522.
LeGrande, W. H. and J. M. Fltzslmons. 1976. Karyology o f the mullets Hugil curem and M. cephalus (Perclformes:Muq1lidae) from Louisiana.
r5peTa197^88-37l . ------
FIELDS OF STUDY
Major Field: Zoology
Studies 1n Systematic Ichthyology: Dr. Ted M. Cavender
Studies 1n Cytotaxonotny: Dr. Ted M. Cavender
v TABLE OF CONTENTS
Page ACKNOWLEDGMENTS...... 11
VITA 1 v
LIST OF TABLES...... v111
LIST OF FIGURES...... u
INTRODUCTION...... 1
REVIEW OF THE FAMILY ICTALURIDAE...... 3
MATERIALS AND METHODS...... 9
Hypoton1c*C1trate Technique...... 10 Acetlc-Orceln Squash Technique ...... 13 Chromosome Observation and Analysis ...... 14
RESULTS...... 18
Genus Ictalu ru s...... 20 Genus Pyjodlctfs...... 22 Genus Noturus...... 23
MECHANISMS OF CHROMOSOME CHANGE...... 59
Changes 1n Diploid Number ...... 59 Changes Not Affecting Diploid Number ...... 63
ICTALURID RELATIONSHIPS: HISTORICAL REVIEW...... 66
ICTALURID CHROMOSOME COMPLEMENTS...... 74
ANCESTRAL DIPLOID NUMBERS...... 94
EVOLUTION OF ICTALURID KARYOTYPES...... 99
vl RELATIONSHIPS IN ICT ALU RIDS...... 106
Intergeneric Relationships ...... 106 Relationships Within the Subgenera Noturus and SchUbeodes...... 108 Relatlo'n'shTps In Rablda...... 116 Diploid Vs. Fundamental Numbers ...... 121 Noturus Relationships: Morphological E v id e n c e ...... 123-* Chromosomes Vs. Morphology ...... 126
SUMMARY...... 130
LITERATURE CITED...... 134
v11 LIST OF TABLES
Table Page 1. Previously reported karyotyplc data on Ictalurld catflshes...... 8
2. Stannary of karyotype data for the 26 species of Ictalurld catflshes 1n this study. Abbreviations: diploid number (211), fundamental nurtfcer (FN), number o f large chromosomes (LC)« number of large msm's (LM), number of cells counted (NC), percent o f hypomodal counts (HoMX), percent o f modal counts (MX), percent o f hypermodal counts (HrMX) ...... 19
3. Distribution of Diploid Numbers for cells counted 1n specimens o f Noturus alb ater. N« number o f ce lls counted fo r each specimen ...... 30
4. Karyotype data for non-1ctalur1d sllurfform fishes 83
v111 LIST OF FIGURES
Figure Page 1. Karyotypes of Ictalurus punctatus (OSUM 35520, not sexed), 2N-58. Multiple karyotypes 1n this and following figures are to Illustrate the variability between spreads from the same species due to d iffe re n t states of chromatid contraction ...... 36
2. Karyotypes of Ictalurus natal1s (OSUM 35535, male), 2N- 62. Note the two large, distinct sm pairs in the msm series ...... 37
3. Karyotype of Ictalurus me!as (OSUM 35534, male), 2N-60. Note the two largest msm's...... 38
4. Karyotypes o f Ictalurus nebulosus (OSUM 35502, male), 2N*60. Note the two largest msm's...... 39
5. Karyotypes of Ictalurus serracanthus (OSUM 35538, fem ale), 2N»52. Note the large number o f msm elem ents... 40
6. Karyotypes of Pylodictls ollvarls (OSUM 37001-S, not sexed), 2N-56. Note the large distinct metacentrlc pair In the msm series ...... 41
7. Karyotypes of Noturus g1lbert1 (OSUM 35499, male), 2N“54. Note the large number o f msm elements ...... 42
8. Karyotvpes o f Noturus Inslgnls (above. OSUM 34172, female) and Noturus exlTls (below, OSUM 37066, female), both 2N"54. Note the larger number of msm elements In inslgnls (above) than 1n e x llIs (below) ...... 43
9. Karyotypes o f Noturus nocturnus (above, OSUM 34172, fem ale), 2N»48 and Noturus leptacanthus (below, OSUM 34759, not sexed), 2N*Tff. Note the 1arge, d is tin c t s t element 1n the s tt series ...... - ...... 44
1 0 ..Karyotype of Noturus funebrls (OSUM 37064-S, female), 2N-44. Note the larg e , d is tin c t st element In the s tt series ...... 45
1x 11. Karyotypes o f Moturus phaeus (OSUM 34778, not sexed), 2NM2. Note the large st element 1n the stt series... 46
12. Karyotypes o f Noturus gyrlnus from Ohio (above, OSUM 35529-S, not sexed) ana Llvlnoston Parish, Louisiana (below, OSUM 34735, not sexed), both 2N-42. Note the large st element 1n the stt series ...... 47
13. Karyotypes of Noturus gyrlnus from Bossier Parish, Louisiana (above, not catalogued, not sexed) and Noturus lachnerl (below. OSUM 34830, fem ale), both 2N*4Z. Note the large st element In the stt series of both. These two karyotypes cannot be distinguished consistently ...... 48
14. Karyotypes of Noturus flavus from northern Ohio River Basin (above, OSUM 35510, female), 2N"48 and from Copper Creek, Clinch River Drainage (below, OSUM 35526, female), 2N*50. Note the large metacentrlc pair In Ohio flavus (above) and its absence 1n the Copper Creek specimen (below) ...... 49
15. Karyotypes of Noturus fla y 1pinn1s (OSUM 34655-S, female), 2N»52. Note the large number of msm elements 1n the karyotype of this species ...... 50
16. Karyotypes o f Noturus mlurus (above, OSUM 35515, not sexed), 2N*50 and Noturus a1!bater (below, OSUM 34045, male), 2N“66. Note the large, distinct metacentrlc pair 1n the msm series o f mlurus (above) and the large number of stt elements In albater (below) ...... 51
17. Karyotypes of Noturus a lb ate r, 2N*70 (above, OSUM 34066, not sexed ) and 2N"7^ (below, OSUM 34066, female). Note the large number of stt elements 1n both and the polymorphism o f 2N ...... 52
18. Karyotypes of Noturus eleoans (OSUM 35516, not sexed), 2N-46. Note the three large distinct pairs of msm elements and the large st pair 1n the stt series ...... 53
19. Karyotypes of Noturus h i1debrand1 latus (above, OSUM 34837, not sexed ) and Noturus Hildebrand! hlldebrandl (below, OSUM 34775, m ale), l)oth 2N*46. Note the "large number of msm*s and the large, distinct st pair (second largest o f the msm's) In both karyotypes ...... 54
20. Karyotype o f Noturus fla v a te r (OSUM 34065, male), 2N-44...... 55
x 21. Karyotypes o f Upturns eleutherus wale (above) and female (below;, ZN-4Z (osUH 134825). Note the large, distinct metacentrlc pair and the large, d is tin c t st pair 1n both karyotypes ...... 56
22. Karyotypes of Noturus stlgmosus (above. OSUM 34196, female) and Noturus munitus 1below, OSUM 34765, not sexed), both
23. Karyotypes of Noturus ta y lo rl male (above) and female (below), both 2M*4tf (OSUH 348^7). Note the heteromorphlc pair of chromosomes In the male (above) but not the female (below).
24. Relationships o f Noturus spp. as proposed by Taylor (1969). Numbers In parentheses refer to diploid and fundamental numbers r e s p e c t iv e ly ...... 72
25. Distribution of diploid numbers in the Ictalurldae 75
26. Distribution of diploid n inters 1n the Cyprlnlformes.... 76
27. Distribution of diploid numbers In the Perclformes 77
28. Distribution of diploid numbers In the Atherinlformes... 78
29. Distribution of diploid numbers 1n the Sllurfformes 82
30. Comparison of probable sympatry among the Noturus species studied and those which share a corrmon diploid number ...... 88
31. D istribution o f fundamental numbers In 1ctalur1ds ...... 100
32. Relationship between fundamental and diploid numbers 1n Ictalurlds (polymorphic N. albater om1tted). Upper and lower dashed lines represent Totally saturated and completely unsaturated karyotypes respectively ...... 101
33. Distribution of diploid numbers 1n the Ictalurldae by taxa studied. Numbers In parentheses Indicate 2N range 1n the polymorphic N. a lb a te r...... 104
34. Intergeneric relationships In the Ictalurldae. Based on Lundberg (1970,1975) and Taylor (1969) ...... 107
x1 35. Proposed hypothesis of phylogenetic relationships between the 20 species of Noturus studied. Based prim arily on chromosomal data. Numbers In parentheses are diploid and fundamental numbers respectively. Presumed synapomorphles, Indicated by dark, numbered bars, are as follows: (1) adnate adipose fin ; relatively short pectoral spines; subcutaneous eye; lengthened, generally toothed, sublateral process of the premaxi1 la with no lamina between the process and premaxi11a; 9 additional characters Indicated by Lundberg (1970) as listed on pp. 68-69; (2) 2N^54; body without distinct mottled pattern or darkly"pig mented saddles and bands; pectoral spines nearly straight, posterior serrae, when present, poorly developed and not recurved toward spine base; anterior dentations, when developed, are Irregularly spaced; (3) FN < 76; (4) 2N < 54; (5) 2N always < 50; large d is tin ct st element present in the s tt series; (6) 2N < 48; (7) 2N<46; (8) lateral line canal system on body does not pass posterior to the anterior margin of the adipose fin ; pectoral mode typ ica lly 7 or 8; (9) anal ray count elevated to 17-21; dark marginal bands 1n the median fins; (10) typically 20- 24 anal rays; (11) pectoral spines scimitar shaped; posterior serrae well developed and consistently recurved toward spine base; anterior serrae present and usually regularly spaced; 2 N < 5 4 : (12) 2N ^ 52; 03) 2N<48; (14) 2N<46; FN ^ 68; (15) 2N^ 44; (16) no large, distinct metacentrlc pair that 1s separable from the remaining msm's; FN*62; (17) 2N» 40; late ral lin e canal system on body does not pass posterior to the anterior margin of the adipose fin ; sex chromosomes Id e n tifia b le ; (18) elongate body, pec toral spines relatively shorter than other Rablda; relatively short head; reduced anterior serrae on pec toral spine; short humeral process of the postclelthrum; (19) four largest msm pairs distinctly separable from the remainder of that series. Note: Noturus albater 1s ten tatively presented as having diverged early from the elegans-h11debrandl lin e although chromo somal ly 1t snould be placed as a sis ter group to a ll other Rablda since 1t 1s the only member of that sub genus to Wave Increased 2N above the presumed ancestral condition and retains a pleslomorphlc FN ...... 109
36. Regression o f the number o f large msm's (LH) 1n Ictalurld karyotypes on 2N. (Noturus albater omitted)... 115
xi 1 INTRODUCTION
Since the publication of the first relatively simple procedure
for the display o f fish chromosomes by McPhail and Jones (1966),
the ichthyological literature has produced an ever increasing number
of papers dealing with karyotypes and karyotyping of fishes (see
Denton 1973, C h ia re lli and Capanna 1973, N iko l'skiy and Vasil'yev
1973 for reviews of karyotypic studies on fishes). Even so, little
direction or synthesis has been apparent in the burgeoning literature
on fish cytotaxonomy u n til recently. Much o f the e a rlie r work dealt
with purely descriptive karyology of one or a few representatives of
large taxa (families or orders) with only cursory efforts to extract
any significant implications about relationships from the chromosome
data. Comparison to previous concepts of relationships based on more
classical approaches to phylogenetic problems has often been lacking.
This has not always been the fault of the investigator, but an
inevitable consequence of the relative infancy of the field and the
great number, diversity and wide geographic distribution of many fish
groups, making representative sampling an almost monumental task.
Coupled with these problems has been the reluctance of cytogeneticists and ichthyologists to work with fis h chromosomes because of th e ir
small size and, often, high diploid numbers.
1 2
This study was initiated with four principal objectives in
mind:
1. Survey the karyotypes of as many extant species
of a selected family of fishes as possible so
as to ascertain what karyotypic diversity is
present within that family.
2. Infer what major mechanisms of chromosomal
change had been operative in the evolution of
these karyotypes.
3. Assess the intrafamilial relationships on
the basis o f chromosomal data.
4. Compare and integrate chromosomally derived
concepts of relationships with those
previously postulated on the basis of morpho
logical criteria in hopes of developing a
more solidly based view of relationships
within that family.
The family selected for this study was the Ictaluridae, which
includes all native North American catfishes. Ictalurids are an
ideal group for such a study since they represent a close-knit assemblage of fishes, moderately speciose, with several distinct
lineages (genera) and a distribution restricted principally to eastern North America. REVIEW OF THE FAMILY ICTALURIDAE
Native, North American freshwater catfishes belong to the family Ictaluridae. Ictalurids are the fourth largest family of freshwater fishes in the United States. Most extant species are restricted to eastern North America, with one (Ictalurus meridionalis) extending south on the Gulf Slope of Central America to B ritis h
Hondouras (M ille r 1966). Of the 43 Recent species described (Lundberg
1975, Taylor 1969 and Douglas 1972), 36 are found in the freshwaters of the United States, with representatives in most major drainages east of the Rocky Mountains. Additionally, two undescribed madtoms
(Noturus spp.) are currently under study (Jenkins 1976).
The fam ily is composed of six extant genera. The following classification of modern species currently in use is based on those by Lundberg (1975) and Taylor (1969).
Genus Ictalurus Rafinesque
Subgenus Ictalurus Rafinesque
furcatus group
K balsanus (Jordan and Snyder) K furcatus (LeSueur)
punctatus group
I_. australis (Meek) 1* dugesi (Bean)
3 X- lupus (Girard) JL* mexicanus (Meek) X- price^T (Gutter) X- punctatus Rafinesque
Subgenus Amiurus G ill
catus group
X* brunneus (Jordan) X. catus (Linnaeus) X- plat.ycephalus (Girard) X* serracanthus Yerger and Re1yea
natalis group
X- melas (Rafinesque) I. natal is (LeSueur) T. nebulosus (LeSueur)
Genus Noturus Rafinesque
Subgenus Noturus Rafinesque
N_. flavus Rafinesque
Subgenus Schilbeodes Bleeker
N. e x ilis Nelson N.- gyrinus (Mitchi 11) Hi' msignis (Richardson) N_. 1 achneri Taylor N. leptacanthus Jordan N.* nocturnus Jordan and G ilbert g ilb e rti~ Jordan and Evermann
funebris group
IX funebris G ilbert and Swain N. phaeus Taylor
Subgenus Rabida Jordan and Evermann
N.* albater Taylor PL eleutherus Jordan N_. tayTori Douglas
hiIdebrandi group
N_. bai leyi Taylor N_. hi Idebrandi (Bailey and Taylor) 5
elegans group
N. elegans Taylor N_. trautmani Taylor
furiosus group
furiosus Jordan and Meek N_. munitus Suttkus and Taylor N_. placidus Taylor N^. stigmosus Taylor
miurus group
N_. flavater Taylor PT flavipinnis Taylor N_. miurus Jordan
Genus P rie te lla Carranza
P_. phreatophila Carranza
Genus P ylodictis Rafinesque
P_. olivaris (Rafinesque)
Genus Satan Hubbs and Bailey
S^. eurystomus Hubbs and Bailey
Genus Trogloqlanis Eigenmann
T. pattersoni Eigenmann
Of these six genera, three (Prietella, Satan and Trogloqlanis) are highly specialized, monotypic genera adapted to subterranean existence and known only from a few preserved specimens taken in the southwestern United States and Mexico; they w ill be considered no further in this study.
Of the 15 species in Ictalurus, 9 are found in the fresh waters of the United States and six primarily on the Gulf slope of eastern
Mexico and Central America. A ll s tr ic tly Mexican forms belong to the 6
subgenus Ictalurus. While the species of I_. (Ictalurus) are widely
distributed in the United States, three of the species of the subgenus
Amiurus, brunneus, platycephalus and serracanthus, are restricted
to the Atlantic and Gulf Slopes of the southeastern U. S. (Yerger and
Relyea 1968). The monotypic flathead c a tfish , P ylodictis o liv a r is ,
is principally a larger river species, widely distributed in the
Mississippi Basin (Trautman 1957). Ictalurus and P ylodictis are
commercially important fishes.
With the description o f Noturus ta y lo ri by Douglas (1972) and
the discovery of two as yet undescribed madtoms in Tennessee and
North Carolina (Jenkins 1976) the total number of recognized species
in Noturus is 26. These are almost e n tire ly re stricte d to the
eastern United States, with three species (N_. flavus, gyrinus and
miurus) penetrating into southern Canada (Scott and Crossman 7973).
Representatives of the genus are found in most major drainages of the
eastern United States, although many of its species are characterized
by a relatively restricted distribution, being found in only one or
a few drainages. L ittle is known about the biology of most madtoms
due to their secretive nocturnal habits (Taylor 1969).
With the study of the flathead bullheads of the southeastern
U. S. (Yerger and Relyea 1968) and the monographic revision of the
madtoms (Taylor 1969), most problems of taxonomy have been worked out
in the fam ily, enhancing its value as a model fo r the study of chromosomal evolution.
In spite of the wide distribution and economic importance of many ictalurids, little information has been available on their chromosomal constitution until recently (Table 1). Prior to the initiation of my study, karyotypes had been published for only K punctatus (Muramoto, Ohno and Atkin 1968), I_. nebulosus (Roberts
1973) and Noturus gyrinus (Levin 1973). A recent study by Hudson
(1976) included karyotype data on these and eight additional ictalurid species. Table 1. Previously reported karyotypic data on ictalurid catfishes. Abbreviations: diploid number (2N), fundamental number (FN).
Species ______2N FN_____ Source ______
Ictalurus punctatus 56 Muramoto, Ohno and Atkin 1968 Ictalurus- punctatus 56--58 84--99 Hudson 1976 Ictalurus furcatus 58 84 Hudson 1976 Ictalurus melas 60 80-86 Hudson 1976 Ictalurus nebulosus 60 -- Roberts 1973 Ictalurus nebulosus 60 76-32 Hudson 1976 Ictalurus natal Ts 62 84 Hudson 1976 Ictalurus brunneus 62 96-106 Hudson 1976 Ictalurus plat.ycephalus 54 92 Hudson 1976 Ictalurus" catus 48 64-68 Hudson 1976 Pylodictis olivan's 56 75-80 Hudson 1976 Noturus gyrfnus 42 Levin 1973 Noturus gyrmus 42 64-66 Hudson 1976 Noturus insignis 54 91-96 Hudson 1976 MATERIALS AND METHODS
The origin of the material used fo r chromosome analysis of each
species is given under individual accounts in the RESULTS section.
With the exception of Pylodictis olivaris and Ictalurus punctatus,
all material has been collected from native populations by seining.
Because of their nocturnal habits, the most effective method for
collecting madtoms (Noturus) is night seining, as noted by Taylor
(1969). Specimens were transported from the field to the laboratory where they were maintained in well aerated aquaria until they could be processed.
Chromosome preparations were made employing either the hypotonic- c itra te method (LeGrande 1975, LeGrande and Fitzsimons 1976), an acetic-orcein squash procedure (modified from McPhail and Jones 1966) or both. Regardless of the technique employed, each specimen received an intraperitoneal injection of Velban (0.5 mg/ml; Eli Lilly Co.)
2 1/2 - 3 1/2 hr before sacrifice. No weight specific dosage was employed, but fishes received from 0.1 to 1.0 cc of Velban, the specific dosage depending on the size of the specimen. Longer periods of Velban pretreatment w ill often result in inaccurate counts and descriptions of chromosome form due to excessive chromatid contraction and/or premature chromatid separation. Following injection, specimens were held in well aerated holding tanks until sacrificed.
9 10
A detailed schedule for processing the appropriate tissues in each of the two techniques employed is given below.
Hypotonic-Citrate Technique
This procedure usually gave the highest number of countable spreads and greater resolution of chromosome morphology. Primary hemopoetic tissues (kidney and spleen) were used as a source of dividing cells. The procedure was as follows:
1. The excised tissue was placed in a drop of 0.6%
sodium citrate solution (the amount of tissue
needed was not large; 10-20 mm3 was generally
sufficient). Excessive amounts of tissue
usually resulted in poorer slides due to crowding
of the ce lls when the suspension was la te r con
centrated and placed on slides.
2. The tissue was minced with fine scissors and sus
pended in 1-3 ml of c itra te solution (the volume
depending on the amount o f tissue used).
3. The suspension was allowed to stand with occasional
aspirations with a pasteur pipette. The aspirations
were in itia lly vigorous, but as the cells swelled
in the hypotonic solution, the aspirations were
gentler. A fter approximately 20-25 minutes in the c itra te , the tissue suspension was filte re d through cheesecloth to remove large pieces.
A fter 30 minutes in c itra te , the cell suspension was centrifuged for 5 min at approximately 600-
1200 rpm.
The supernatant was quickly poured off the cell button.
While continuing to hold the centrifuge tube upside down, the remaining citrate was washed from the lip of the tube with fresh fixative (3 parts absolute methanol : 1 part glacial acetic acid).
It was imperative that all citrate solution be removed as it tended to interfere with proper fixation of the cells.
The tube was turned upright and 1-2 ml of fresh fixative were slowly added without disrupting the button of c e lls. The entire process of draining the supernatant and adding fixative were accomplished as rapidly as possible.
The button was allowed to stand 7-10 min in f i x ative then resuspended with a clean pasteur pipette.
Care was taken to scrape the walls of the tube with the pipette tip to remove any cells that might have adhered. The suspension was centrifuged.
The supernatant was decanted, fresh fixative added, the cells resuspended and centrifuged again.
Step 11 was generally repeated twice, fo r a total of three washes in fixative after the
in itia l fix and centrifugation. These washes were necessary to remove small particulates and provide cleaner preparations. One or two of these washes were eliminated if the cell button was small, as some cells were lost with each successive wash.
After the final centrifugation, all but 0.3-
0.5 ml of supernatant (depending on the size of the remaining cell button) were aspirated o ff.
The cells were resuspended in the remaining fixative and dropped (3 or 4 drops) onto clean, labeled glass microslides.
The fixative was immediately ignited and allowed to burn completely. When the flame had died, the slides were vigorously shaken to remove any remaining flu id and allowed to dry completely before staining. 13
The dry slides were usually stained as soon as possible; older slides did not stain as well as fresh material. Slides were stained in Giemsa (10 ml stock solution : 40 ml distilled water) for approx imately 40-50 min (time and concentration of stain may vary for best results). Slides were removed from the staining solution, rinsed in tap water, then dehydrated by the following procedure: dipped several times in each of two baths of nure acetone, dipped several times in one bath of 1:1 acetone-xylene and allowed to stand 2-5 min in each of two baths of pure xylene. The slides were then removed from the xylene, a miscible mounting medium added and covered with a 22 X50 mm glass covers lip .
Acetic-Orcein Squash Technique
This technique was usually employed with individuals of less than 40 mm in standard length because of d iffic u lty in obtaining enough tissue from the kidney and spleen of such small specimens to process by the hypotonic-citrate method. G ill epithelium was used as the source of dividing cells. The procedure was basically that of
McPhail and Jones (1966) as modified by Uyeno (personal communication).
A detailed treatment schedule was as follows:
1. The anteriormost two g ill arches were removed
from one or both sides of the sacrificed specimen.
2. The intact g ill arches were placed Into d is tille d
water in small petri dishes and allowed to stand
fo r 30 min. 14
3. After hypotonic treatment the distilled water was
removed from the dish with a clean pasteur pipette,
disturbing the arches as l i t t l e as possible.
4. The arches were covered with fresh, filtered,
acetic-orcein stain {4% orcein in 60% acetic acid).
5. The arches were allowed to stain for 15 minutes
before slide preparation began.
6. The arches were removed from the stain with micro
forceps and swabbed lightly on a slide in a small
drop of clean 60% acetic acid.
7. The arch was discarded and any clumps o f tissue
removed from the slide with micro-forceps.
8. A 22 X 22 mm coverslip was placed over the suspension,
covered with a piece of f i l t e r paper and squashed
with gentle thumb pressure.
9. The slide was then sealed with Kroenig cement to
retard drying and stored in a refregerator until
observations were made.
Chromosome Observation and Analysis
Chromsome preparations were scanned fo r countable spreads under
200-400X magnification with an Olympus phase-contrast microscope.
Coordinates of suitable spreads were recorded and these were later counted under 1500X m agnification. As many c e lls as possible were counted or u n til a clear mode was evident. The modal number fo r 15 each species (N_. albater excepted) made up between 50% and 75% of the counts. Counts were always negatively skewed about the mode (both fo r individual specimens and species). Hypomodal counts ( i. e . , any number of chromsomes in a spread that was less than the modal count) were always less than 50% of the counts. Hypermodal counts ( i. e . , any count greater than the mode) were always less than 5%, except in
I_. serracanthus where only 15 cells were counted. This trend in the d istrib u tio n of counts resembles that seen in most studies on fish chromosomes and is accepted as a rtifa c tu a l. Hypomodal counts are lik e ly to be due to the loss of one or more chromosomes from the spread during slide preparation, counting errors or obstruction of one or more chromosomes due to overlap of the elements. Hypermodal counts could result from the less lik e ly addition of one or more stray chromosomes during preparation, overlap of adjacent spreads or premature chromatid separation in one or more of the elements. As usually expected of diploid numbers, modal counts for a ll species
(except some of the polymorphic N_. albater specimens) fe ll on even numbers, while many of the hypo- and hypermodal counts were odd values.
Thus, in keeping with accepted procedure, the modal count for each species is taken as the correct diploid number.
High quality spreads with the modal number of chromosomes were photographed (fo r both sexes when material was available) for analysis of chromosome morphology and preparation of a typical karyotype for the species. Since many countable spreads do not provide sufficient resolution of the morphology of individual elements, emphasis has been placed on the analysis of a few high quality spreads from each species 16
rather than a larger number of poorly conformed, distorted spreads
in which not a ll elements were clearly visib le . Poorer spreads were often compared to the prepared karyotypes to check fo r any gross deviations. Observations of photographic enlargements were sometimes supplemented by reexamination of the actual spread with the microscope
to ensure proper interpretation of chromosome morphology. The number of karyotypes prepared fo r each species ranged from one to ten. I have placed l i t t l e emphasis on quantitative analysis of chromosome complements in icta lu rid s because of variation in the size and, to a lesser extent, morphology of the chromosomes due to d iffe re n t states of chromatid contraction from cell to c e ll. For this reason, the measurements of chromosomes were made from only one or a few cells fo r each species and these values were used q u a lita tive ly to describe the number of elements of d iffe re n t morphologies. The following abbreviations are used throughout the text to refer to various aspects of chromosome form and size:
%TCL: Percent of total complement length; calculated by
dividing the total length of a chromosome pair by
the total length of a ll chromosomes in the complement
and multiplying by 100. This is essentially an
index of the size of a particular chromosome pair
in relation to the size of all other pairs. 17
L/S: Arm ratio; calculated by dividing the length of
tlie long arm of a chromosome by the length of
its short arm. This is an index of centomeric
position.
LC: The number of large chromosomes in the complement.
A large chromosome was a rb itra rily defined as any
chromosome pair with a value o f %TCL greater than
5.0%.
LM: The number of large biarmed chromosomes; th is is
the number of the large chromosomes (LC) that f a ll
within the metacentric-submetacentric series.
LC+2N: The number of large chromosomes (LC) plus the
diploid number (2N).
FN: Fundamental number; calculated by assigning each
metacentric or submetacentric chromosome a value
of two and each subtelocentric or acrocentric
chromosome a value of one, then summing these
values for the entire complement. This may « otherwise be considered the number of major
chromosome arms.
Nomenclature fo r centromeric position follows the c rite ria established by Levan, Fredga and Sandberg (1964): L/S <1.7, metacentric
(m); 1.7^-L/S< 3.0, submetacentric (sm); 3.0 (st); L/S>7.0, acrocentric (t). RESULTS The results of this study are presented below as individual species accounts (annotated as necessary) and summarized in Table 2. Each species account includes the following information: origin of specimens used for karyotype analysis, disposition of material saved (OSUM = Ohio State University Museum of Zoology), to ta l number of cells counted, modal 2N (hypomodal, modal and hypermodal count per centages in parentheses), Fundamental Number (FN, best approximation), number o f large chromosomes plus the 2N (LC+2N, best approximation), and the Karyotype Formula stating the number of metacentric-sub- metacentric (msm) and subtelocentric-acrocentric (stt) elements. Annotations (where needed) describe any distinctive pairs in the complements or other aspects of the karyotype that may be of particular interest. Because of the large number o f elements in ic ta lu rid karyotypes that possess small to large second arms, as well as the large number of chromosomes that are on the borderline between morphological categories (i.e ., metacentric-submetacentric, submetacentric- subtelocentric, subtelocentric-acrocentric) the complements have been divided into only two morphological groupings, msm elements and stt elements. For the same reasons, FN is d iffic u lt to assess accurately 18 Table 2. Summary of karyotype data for the 26 species of ictalurid catfishes in this study. Abbreviations: number of specimens (N), diploid number (2N), fundamental number (FN), number of large chromosomes (LC), number of large msm's (LM), number of cells counted (NC), percent of hypomodal counts (HoM%), percent of modal counts (M%), percent of hypermodal counts (HrM%). Species N 2N FN LC LM LC+2N Formula NC Horn M* HrMS Ictalurus punctatus 4 58 92 ------58 34msm,24stt 62 25.8 74.2 0.0 Ictalurus natalis 2 62 84 2 2 64 22msm,40stt 59 30.5 67.8 1.7 Ictalurus melas 3 60 76 -- 60 16msm,44stt 36 38.9 58.3 2.8 Ictalurus nebulosus 9 60 76 -- 60 16msm,44stt 124 31.5 64.5 4.0 Ictalurus serracanthus 1 52 90 8 6 60 38msm,14stt 15 33.4 53.3 13.3 Fylodictis olivaris 3 56 82 4 2 60 26msm,30stt 67 31.1 67.2 1.5 Noturus qilberti 2 54 82 4 2 58 28msm,26stt 46 34.7 65.3 0.0 Noturus insignis 6 54 74 4 58 20msm,34stt 186 27.4 71.0 1.6 Noturus exilis 2 54 68 6 60 14msm,40stt 57 43.0 57.0 0.0 Noturus nocturnus 10 48 72 10 8 58 24msm,24stt 129 24.1 75.1 0.8 Noturus leptacanthus 10 46 74 16 12 62 26msm,20stt 53 24.5 75.5 0.0 Noturus funebris 2 44 68 14 12 58 24msm,20stt 35 48.7 51.3 0.0 Noturus phaeus 3 42 68 14 12 56 26msm,16stt 30 23.3 73.4 3.3 Noturus gyrinus 11 42 72 14 10 56 30msm,12stt 32 26.6 71.2 2.2 Noturus lachneri 9 42 72 12 10 54 30msm,12stt 138 34.8 63.0 2.2 Noturus flavus (Copper Cr.) 2 50 70 6 56 20msm,30stt 38 44.8 52.6 2.6 Noturus flavus 8 48 70 8 2 56 22msm,26stt 208 27.8 71.2 1.0 Noturus flavipinnis 2 52 82 10 4 62 30msm,22stt 57 38.7 59.6 1.7 Noturus miurus 11 50 74 12 8 62 24msm,26stt 111 39.6 58.6 1.8 Noturus albater 13 66-72 82 4 -- 70 230 - - <____ * - - Noturus elegans 3 46 82 8 8 54 36msm,10stt 30 46.7 53.3 0.0 Noturus h. hildebrandi 15 46 80 12 10 58 36msm,10stt 126 35.7 61.1 3.2 Noturus h. latus 6 46 80 12 10 58 36msm,10stt 53 33.8 64.2 2.0 Noturus flavater 1 44 62 14 10 58 20msm,24stt 21 33.3 66.7 0.0 Noturus eleutherus 7 42 66 16 10 58 24msm,18stt 107 32.7 63.6 3.7 Noturus stigmosus 1 42 62 12 8 54 20msm,22stt 45 26.7 73.3 0.0 Noturus munltus 8 42 62 16 10 58 20msm,22stt 66 42.4 57.6 0.0 Noturus taylori 9 40 61-62 16 12 56 24msm,16stt 138 40.1 59.4 1.4 20 in ictalurids as cells in different states of chromosome contraction may vary slightly in FN due to more rapid contraction of the long arm of a chromosome than its short arm (Bogart 1969 1n Thompson 1976). Thus a st element with a f a ir ly large second arm may measure as a sm in more contracted material. The FN values presented here are my best approximation from the least contracted cells in the material avail able to me for study. The karyotypes of the species studied are presented in Figures 1 to 23, with the msm elements arranged in a group above s tt elements. Genus Ictalurus Rafinesque Ictalurus (Ictalurus) punctatus Rafinesque- Material: OSUM 34420 (3 not sexed), Louisiana, East Baton Rouge Parish, Drainage ditch near Louisiana State University Campus, OSUM 35533 (1 not sexed), Ohio, Allen Co., Miami-Erie Canal; Cells counted: 62, 2N: 58 (25.8%, 74.2%, 0.0%); FN: 92; LC+2N: 58; Karyotype Formula: 34 msm, 24 s t t; Figure 1. The msm and s tt groups were both composed of elements that gradu ally changed in both size and form, precluding matching of homologous chromosomes. Second arms were id e n tifia b le on nearly a ll s tt elements. Ictalurus (Amiurus) natal is (LeSueur)- Material: OSUM 35535 (1 male), Ohio, Allen Co., Miami-Erie Canal, OSUM 34063 (1 not sexed), Missouri, Ozark Co., North Fork o f the White River; Cells counted: 59; 2N: 62 (30.5%, 67.8%, 1.7%); FN: 84; LC+2N: 62; Karyotype Formula: 22 msm, 40 s t t; Figure 2. 21 The largest msm pair was easily recognized as a sm pair. Second arms on many of the s tt elements were very small and d iffic u lt to Identify. There was a distinct size gap between the two largest msm pairs and the remainder of that series. Ictalurus (Amiurus) me!as (Rafinesque)- Material: OSUM 35534 (1 not sexed), Ohio, Allen Co., Miami-Erie Canal, OSUM 35501 (2 males), Ohio, Sandusky Co., East Harbor State Park, Lake Erie; Cells counted: 36; 2N: 60 (38.9%, 58.3%, 2.8%); FN: 76; LC+2N: 60; Karyotype Formula: 16 msm, 44 s tt; Figure 3. The two largest pairs of msm elements were sm, d is tin c tly larger than the remainder of the msm's. Many of the s tt elements had tiny or nonidentifiable second arms. Ictalurus (Amiurus) nebulosus (LeSueur)- Material: OSUM 35537-S (1 female), Ohio, Franklin Co., Mirror Lake, Ohio State University Campus, OSUM 35505 and 35507 (1 female, 1 not sexed), Ohio, Lucas Co., Maumee Bay, Lake Erie, OSUM 35502 (4 males, 2 females), Ohio, Sandusky Co., East Harbor State Park, Lake Erie; Cells counted: 124; 2N: 60 (31.5%, 64.5%, 4.0%); FN: 76; LC+2N: 60; Karyotype Formula: 16 msm, 44 s tt; Figure 4. The largest two pairs of msm elements were separable from the remainder of that series by a size gap. The largest pair was sm while the second largest pair was on the borderline between m and sm, depending on the state of chromatid contraction in the cells. This karyotype could not be consistently distinguished from that of I_* me!as within the limits of resolution possible in the materialavailable. Many of the stt 22 elements had, as In K me!as, no id e n tifia b le second arms or very tin y ones. Ictalurus (Amiurus) serracanthus Yerger and Relyea- Material: OSUM 35538 (1 female), Florida, Columbia-Alachua Co. Line, Santa Fe River; Cells counted: 15; 2N: 52 (33.4*, 53.3%, 13.3%); FN:90; LC+2N: 60; Karyotype Formula: 38 msm, 14 s tt; Figure 5. A single specimen was available fo r study and only a few cells were found to be o f adequate q u a lity fo r counts. Karyotypes were prepared and are te n ta tive ly presented fo r th is species pending analysis of larger samples. The most distinctive elements of the msm series were the two largest pairs. The largest pair of that series was sm. The remaining msm and s tt elements were graded both in size and form, making accurate matching of homologs impossible. Genus Pylodictis Rafinesque Pylodictis olivaris (Rafinesque)- Material: OSUM 37001-S (3 not sexed), Ohio, Guernsey Co., Senecaville National Fish Hatchery; Cells counted: 67; 2N: 56 (31.3%, 67.2%, 1.5%); FN: 82; LC+2N: 60; Karyotype Formula: 26 msm, 30 s tt; Figure 6. The most d is tin c tiv e pair o f chromosomes (which was m) was the largest (about 7.0%TCL) in the complement. The second largest pair of the msm series was usually recognizable by its size and centromere placement. The remainder of both the msm and stt series could not be accurately matched. Most o f the second arms (where present) of the s tt elements were small. 23 Genus Noturus Rafinesque Noturus (Schilbeodes) g i1berti Jordan and Evermann- M aterial: OSUM 35499 (2 males), V irg in ia , Roanoke Co., Roanoke River; Cells counted: 46; 2N: 54 (34.7%, 65.3%, 0.0%); FN: 82; LC+2N: 58; Karyotype Formula: 28 msm, 26 s t t; Figure 7. Both series were composed o f elements not separable in to homologous pairs. Noturus (SchiIbeodes) insignis (Richardson)- M aterial: OSUM 35527 (1 not sexed), Pennsylvania, Cameron Co., Sinnemahoning River, OSUM 37067 (1 male, 1 female), Pennsylvania, Cameron Co., Sinnemahoning River, OSUM 35498 (1 male), V irg in ia , Roanoke Co., Roanoke River, OSUM 35504 (1 male, 1 female), North Carolina, Burke Co., Paddy Creek; Cells counted: 186; 2N: 54 (27.4%, 71.0%, 1.6%); FN: 74; LC+2N: 58; Karyotype Formula: 20 msm, 34 s tt; Figure 8 (top). The karyotype consisted of elements graded in both size and shape, precluding consistent matching of homologs in both msm and stt categories. Specimens from the various drainages sampled could not be chromosomally distinguished. Noturus (Schilbeodes) e x ilis Nelson- Material: OSUM 34064 and 37066-S (2 females), Missouri, Ozark Co., North Fork o f the White River; Cells counted: 57; 2N: 54 (43.0%, 57.0%, 0.0%); FN: 68; LC+2N: 60; Karyotype Formula: 14 msm, 40 s tt; Figure 8 (bottom). Several elements in the msm series were d is tin c tiv e . Four of the msm's, including the largest pair, appeared to be m, with the other 24 three pairs in that series strongly sm. The stt series was graded In size and form making id e n tifica tio n of homologs impossible. Most of the s tt elements, however, were observed to possess small to large second arms. Specimens of the small, disjunct population in the lower stretches of the Tennessee and Cumberland River Basins were not available for study. Noturus (Schilbeodes) nocturnus Jordan and G ilbert- Material: OSUM 34777 (1 not sexed), M ississippi, Amite Co., Brush Creek, OSUM 34172 (1 male, 5 females), Missouri, Stoddard Co., Unnamed drainage ditch, OSUM 34842 (1 male, 1 female, 1 not sexed), Louisiana, Livingston Parish, Hog Branch; Cells counted: 129; 2N: 48 (24.1%, 75.1%, 0.8%); FN: 72; LC+2N: 58; Karyotype Formula: 24 msm, 24 s tt; Figure 9 (top). The two largest pairs in the msm series were characteristically sm and m respectively, while the remainder of that series could not d e fin ite ly be matched into homologs. Of the s tt's , only the largest pair was readily recognized by the size gap between i t and the remaining members of that series. All but the smallest stt consistently had identifiable second arms. Karyotypes from Missouri and Louisiana specimens could not be distinguished. No karyotypes were prepared from the Mississippi specimens. Noturus (Schilbeodes) leptacanthus Jordan- Material: OSUM 34736 (5 females, 3 males), Louisiana, Livingston Parish, Hog Branch, OSUM 35530 (1 not sexed), Louisiana, Washington Parish, M ille r Creek, OSUM 34758 (1 female), Louisiana, Washington Parish, Hays Creek; Cells 25 counted: 53; 2N: 46 (24.5%, 75.5%, 0.0%); FN: 74; LC+2N: 62; Karyotype Formula: 26 msm, 20 s t t; Figure 9 (bottom). The fiv e largest pairs of elements were usually composed of two pairs of m and three pairs of sm chromosomes. The largest chromosome pair in the complement was generally a st pair (about 7.0 %TCL). The remainder o f both the msm and s tt were not accurately matchable. No apparent sexual dimorphism was noted. Noturus (Schilbeodes) funebris G ilbert and Swain- M aterial: OSUM 34759 (1 not sexed), Louisiana, Washington Parish, Hays Creek, OSUM 37064-S (1 female), Louisiana, Washington Parish, M iller Creek; Cells counted: 35; 2N: 44 (48.7%, 51.3%, 0.0%); FN: 68; LC+2N: 58; Karyotype Formula: 24 msm, 20 s t t; Figure 10. The msm series wascomposed primarily of m elements, with only two or three pairs approaching a sm condition. There was a size gap apparent between the six largest msm pairs and the six smallest. A d ditio na lly, the largest s tt element was separable from the remainder of that series on the basis of size. A ll of the s tt elements possessed identifiable second arms of various sizes. If additional material o f higher q u a lity becomes available, the placement of some elements with large second arms in the stt series may need reevaluation. Noturus (Schilbeodes) phaeus Taylor- Material; OSUM 34778 (1 not sexed), M ississippi, Amite Co., Brushy Creek, OSUM 34838 (2 not sexed), Tennessee, Henry Co., North Fork of the Obion River; Cells counted: 30; 2N: 42 (23.3%, 73.4%, 3.3%); FN: 68; LC+2N: 56; Karyotype Formula: 26 msm, 16 stt; Figure 11. 26 Most of the msm elements were close to a m condition with only two or three pairs approaching a sm condition. The largest st was separable from the remaining stt chromosomes by its size. Some stt's partic ularly the second largest pair, might become placed in the msm series, with analysis of additional, higher quality material. Noturus (Schilbeodes) gyrlnus (M itchill)- Material: OSUM 35529-S (1 not sexed), Ohio, exact locality not known, OSUM 35506 (1 not sexed), Ohio, Lucas Co., Maumee Bay, Lake Erie, OSUM 34171 (1 male, 1 female, 1 not sexed), Missouri, Stoddard Co., Unnamed drainage d itch , OSUM 34735 (6 not sexed), Louisiana, Livingston Parish, Hog Branch; Cells counted: 132; 2N: 42 (26.6%, 71.2%, 2.2%); FN: 72; LC+2N: 56; Karyotype Formula: 30 msm, 12 s tt; Figures 12 and 13 (top). L ittle could be accomplished toward id e n tific a tio n of homologous pairs due to the graded size and shape of the elements in both series. Almost a ll elements had id e n tifia b le second arms in the s tt series. About the third largest pair in the msm series was a distinctive sm (almost st) pair. The largest stt pair was noticeably larger than any other pair in that group. Specimens from Ohio, Missouri and Louisiana could not be separated confidently on the basis of karyotype. Noturus (Schilbeodes) lachneri Taylor- Material: OSUM 34830 (6 females, 3 males), Arkansas, Garland Co., M erriott Branch of the Middle Fork o f the Saline River; Cells counted: 138; 2N: 42 (34.8%, 63.0%, 2.2%); FN: 72; LC+2N: 54; Karyotype Formula: 30 msm, 12 s tt; Figure 13. 27 As in N_. gyrinus, l i t t l e matching of homologs was possible. However, there was the same distinct pair of sm chromosomes, about the third largest msm pair, as seen in N. gyrinus. Also as in N_. gyrinus, there was a large st pair separable from other pairs in the stt series. All but the smallest of the stt had consistently identifiable second arms. This karyotype was not consistently distinguishable from that of N_. gyrinus. Noturus (Noturus) flavus Rafinesque- Material: OSUM 35518 (1 male, 1 not sexed), Ohio, Pickaway Co., Big Darby Creek, OSUM 35508 (2 females, 2 males), Pennsylvania, Erie Co., Conneaut Creek, OSUM 35510 (2 females), Pennsylvania, Crawford Co., French Creek; Cells counted: 208; 2N: 48 (27.8%, 71.2%, 1.0%); FN: 70; LC+2N: 56; Karyotype Formula: 22 msm, 26 s tt; Figure 14 (top). Material: OSUM 35517 and 35526 (2 females), V irginia, Scott Co., Copper Creek; Cells counted: 38; 2N: 50 (44.8%, 52.6%, 216%); FN: 70; LC+2N: 56; Karyotype Formula: 20 msm, 30 s tt; Figure 14 (bottom). The Virginia material was found to differ from other populations of N_. flavus sampled. While FT flavus from Ohio and Pennsylvania were found to be indistinguishable in th e ir karyotype, both having a diploid number of 48, the Virginia specimens from the Clinch River headwater trib u ta ry, Copper Creek, had 2N=50. The most d istin ctive pair in the Ohio and Pennsylvania material was a large m chromosome pair, separable from all other msm's in the complement. This pair was conspicuously absent in the Copper Creek specimens, which had, instead, two more pairs of stt elements than other Nk flavus karyotyped. In material 28 that was not overly contracted, second arms were usually Identifiable on most of the subtelocentric-acrocentric elements. Noturus ( Rablda) fla v lp ln n ls Taylor- M aterial: OSUM 35531 and 34655-S (2 females), Virginia, Scott Co., Copper Creek; Cells counted: 57; 2N: 52 (38.7%, 59.6%, 1.7%); FN: 82; LC+2N: 62; Karyotype Formula: 30 msm, 22 s tt; Figure 15. A ll chromosomes in the complement possessed id e n tifia b le second arms. Matching of homologs was not possible. Noturus (Rabida) mlurus Jordan- M aterial: OSUM 35515 (1 not sexed), Ohio, Union Co., Big Darby Creek, OSUM 35512 (2 not sexed), Missouri, Stoddard Co., Unnamed drainage ditch, OSUM 34737 (1 not sexed), Louisiana, Livingston Parish, Hog Branch, OSUM 35521 (5 not sexed), Ohio, Ross Co., Salt Creek, OSUM 35513 (2 not sexed), Ohio, Ross Co., Salt Creek; Cells counted: 111; 2N: 50 (39.6%, 58.6%, 1.8%); FN: 74; LC+2N: 62; Karyotype Formula: 24 msm, 26 s t t; Figure 16 (top). The largest chromosome pair in the msm series was a large (about 8.0% TCL) m pair that was id e n tifia b le in a ll spreads examined. This largest pair was followed by two pair of m's of about the same size as each other, with the remainder of the msm complement composed mostly o f sm chromosomes that were not matchable because of the gradual change in both size and form of the elements. The smallest two or three pairs of msm’s were generally m in form. Some o f the sm elements approached a st condition. The s tt formed a graded series; some 29 (especially the larger elements of the group) possessed noticeable second arms, but most of the stt had small to nonidentlflable second arms. No consistent geographic variation 1n the karyotypes was apparent. Noturus (Rabida) albater Taylor- Material: OSUM 34066 and 34045 (3 males, 10 females), Missouri, Ozark Co., North Fork o f the White River; Cells counted: 230; 2N: 66-72; FN: 82; Figures 16 (bottom) and 17. The karyotype o f th is species was observed to vary from specimen to specimen. Table 3 is a summary of the d is trib u tio n o f chromosome counts from the 13 specimens examined. There was a d is tin c t modal count in most o f the specimens, but the modes fo r d iffe re n t specimens varied from 66-72, with modes of 66, 67, 70, 71 and 72 represented in one or more specimens. Most o f the chromosomes were very small and d iffic u lt to resolve clearly even in enlarged photomicrographs. How ever, photographs of ce lls from specimens with modal counts of 66, 70 and 72 were prepared for tentative analysis of the karyotype. From these preliminary karyotypes, it appears that the difference in diploid number in these fishes results from a change in the number of msm elements, because the number of chromosome arms (FN) seemed to be the same for all three diploid numbers analyzed. The karyotype formuluae fo r these three specimens were as follows: 2N=66, 16 msm, 50 s t t; 2N=70, 12 msm, 58 s t t; 2N=72, 10 msm, 62 s tt. L it t le was accomplished toward matching of homologs because of the small size of many of the elements and their gradual change in size and form. The largest stt Table 3. Distribution of Diploid Numbers for cells counted in specimens of Noturus albater. N=number of cells counted for each specimen. Specimen No. Sex N Mode 61 62 63 64 65 66 67 68 69 70 71 72 73 74 532 female 25 72 3 1 2 -- 1 3 12 2 - 533 female 20 70 1 — 3 16 538 male 17 66 -- 1 1 1 3 10 I 539 female 22 70 1 3 17 1 540 female 37 71 -- -- 1 — 1 1 — 1 5 24 4 — — 542 female 30 71 1 -- 1 1 2 1 2 1 5 15 1 - - - - 543 male 18 67 1 1 — 1 9 6 549 female 18 71 2 1 — 3 7 5 — — 550 female 6 70 1 1 4 551 female 7 — 1 1 1 2 2 554 male 5 70 1 — — 4 556 female 17 71 2 2 3 3 7 557 female 8 70 1 1 2 4 — — — — TOTALS 230 5 4 5 3 7 19 15 14 14 62 57 22 2 - pair could usually be separated from the remainder of that series. Second arms were d if f ic u lt to distinguish on many of the s tt. Much larger samples from other populations must be analyzed before the karyotype or nature o f this polymorphism can be established fo r certain. Noturus ( Rabida) elegans Taylor- M aterial: OSUM 35514 (1 not sexed), Kentucky, Casey Co., Green River, OSUM 35516 (1 male, 1 female). Kentucky, Allen Co., Little Trammel Fork; Cells counted: 30; 2N: 46 (46.7%, 53.3%, 0.0%); FN: 82; LC+2N: 54; Karyotype Formula: 36 msm, 10 s tt; Figure 18. In the msm series, the largest three pairs were distinguishable from the remainder of the series. Of those three, the largest was a sm pair, the second largest distinctly m and the third largest on the borderline between m and sm. The remainder o f the msm elements could not be matched as homologs. The most easily id e n tifie d pair o f s tt series was the largest, a st pair; it was separable from the remaining stt by its size. Noturus (Rabida) hi!debrandi hi!debrandi (Bailey and Taylor)- M aterial: OSUM 34775 (7 males, 8 females), M ississippi, Amite Co., Brushy Creek; Cells counted: 126; 2N: 46 (35.7%, 61.1%, 3.2%); FN: 80; LC+2N: 58; Karyotype Formula: 36 msm, 10 s t t ; Figure 19 (bottom). Noturus (Rabida) hildebrandi latus Taylor- Material: OSUM 34837 (1 male, 3 females, 2 not sexed), Tennessee, Henry Co., North 32 Fork of the Obion River; Cells counted: 53; 2N: 46 (33.8%, 64.2%, 2.0%); FN: 80; LC+2N: 58; Karyotype Formula: 36 msm, 10 s tt; Figure 19 (top). The material represents collections from the type locality of each of the two described subspecies. No consistent differences could be detected between the two forms in e ith e r diploid number or chromosome form, although the largest pair of msm in h_. latus seemed relatively larger than its counterpart in N. h^ hildebrandi in some spreads. Because chromosomes in d iffe re n t ce lls may vary s lig h tly in form due to differing states of chromosomal contraction, it is possible that this difference may simply reflect an artifactual condition. The karyotype of each subspecies possessed distinctive m pairs that were the first and third largest pairs in that series. The second largest msm pair was noticeably sm, approaching a st condition. No outstanding elements were seen in the s t t series. No apparent sexual dimorphism was noted. Noturus ( Rabida) fla va te r Taylor- M aterial: OSUM 34065 (1 male), Missouri, Ozark Co., North Fork o f the White River; Cells counted: 21; 2N: 44 (33.3%, 66.7%, 0.0%); FN: 62; LC+2N: 58; Karyotype Formula: 20 msm, 24 s t t; Figure 20. The material from this single specimen was limited, but left little doubt as to the diploid number of 44 for the species. The karyotype is presented tentatively because minor modification may be necessary with additional material of higher quality. The five largest msm pairs were separable from the five smallest, but matching within each of 33 these groups was not possible. Similarly, no distinctive stt pairs were noted. Noturus ( Rabida) eleutherus Jordan- M aterial: OSUM 34825 (3 males, 2 females), Arkansas, Clarke Co., Caddo River, OSUM 35525 (2 females), Virginia, Scott Co., Copper Creek; Cells counted: 107; 2N: 42 (32.7%, 63.6%, 3.7%), FN: 66; LC+2N: 58; Karyotype Formula: 24 msm, 18 s t t; Figure 21. The exact number of msm elements was d if f ic u lt to diagnose due to the small size and overcontraction of many spreads, but the best karyotypes had 24 msm elements, with the rest stt. The two collections represent material from the two disjunct populations of this species. No striking differences were noted between the karyotypes of these two populations. There was a very large (about 10.5%TCL), characteristic m pair present in a ll of the spreads examined. This pair was easily distinguished from all other msm elements. The second largest msm pair was usually about 7.0%TCL. The remainder of the msm chromosomes could not be matched accurately nor could the s tt chromosomes. The largest stt pair was obviously larger than any other pair in that series. No sexual dimorphism noted. Noturus (Rabida) munitus Suttkus and Taylor- M ate rial: OSUM 34696 and 34765 (1 female, 7 not sexed), Louisiana, Washington Parish, Bogue Chitto River; Cells counted: 66; 2N: 42 (42.4%, 57.6%, 0.0%); FN: 62; LC+2N: 58; Karyotype Formula: 20 msm, 22 s t t; Figure 22 (bottom). The four largest pairs o f msm elements were separable from the remainder 34 o f that series on the basis of size. Most o f the elements in that series were more nearly m than sm. Noturus ( Rabida) stigmosus Taylor- M aterial: OSUM 34196 (1 female), Ohio, Pickaway Co., Big Darby Creek; Cells counted: 45; 2N: 42 (26.7%, 73,3%, 0.0%); FN: 62; LC+2N: 54; Karyotype Formula: 20 msm, 22 stt; Figure 22 (top). As In N. munitus, there was a distinct size gap between the four largest and six smallest pairs of msm elements. No other matching of elements in eith e r the msm or s tt series was possible. Noturus (Rabida) ta y lo ri Douglas- M ate rial: OSUM 34827 (4 males, 3 females, 2 not sexed), Arkansas, Clarke Co., Caddo River; Cells counted: 138; 2N: 40 (40.1%, 59.4%, 1.4%); FN: 61 (male)- 62 (female); LC+2N: 56; Karyotype Formula: 23 msm, 17 s tt (male) and 24 msm, 16 s tt (female); Figure 23. A case of apparent sexual dimorphism was noted in the karyotypes o f th is species. A pair o f heteromorphic chromosome elements appeared in the males as unmatched m and t chromosomes. The chromosomes singled out in Figure 23 as the sex chromosomes have been a r b itr a rily selected from the size and form categories where the heteromorphic pair consistently occurred. It is not possible to identify accurately the specific elements o f th is apparent sex chromosome pair because of the graded size and form of the chromosomes in both series. Accurate id e n tific a tio n and confirmation of th is heteromorphism as a "sex chromosome pair" must await analysis o f much larger samples o f both males and females of th is Caddo River endemic, as well as analysis of meiotic material. I tentatively refer to this as an XX-XY sex chromosome system In which the males are heteromorphic. Aside from th is , l i t t l e could be accomplished toward matching o f homologs 1n each series. The largest six pairs o f msm chromosomes were usually separable from the smallest five pairs and the largest stt pair (a st) was distinguishable from the remaining s tt elements 1n both sexes. 36 U R ftftX IflfH ft jf An k x * Aft*ha v * » A a x a a r k i * * A < A ft A ft Aft A f o f t & ft ft A*'Sft&4 ft A A A ft /» a . ^ ftltftXftAAA ft JIA A A ft** K x kNRA*>R« in H f t A f A A A M U II ft Af t A 0 A ftA /> 6 * f t « Figure 1. Karyotypes of Ictalurus punctatus (OSUM 35520, not sexed), 2N=58. Multiple karyotypes In this and folTowing figures are to illustrate the v a ria b ility between spreads from the same species due to different states of chromatid contraction. 37 X £ M a * a * k HAKMXAkJLX* m * (\AMAAAAAA A A A A Ail Aft/* a a a X I A A A A * * * * ^ M K V A H A A A A • • A A ft 0 A A A ^ A a A A A O itA AAft/S A A A A A A A A ^ > A a n Figure 2. Karyotypes of Ictalurus natal is (OSUM 35535, male), 2N=62. Note the two large, distinct sm pairs In the msm series. 38 D/it******* A A * A •» « A ft 0 A It A IV 6 A A • A ll A A A A ■*■ M l < l #i 0 a A A A A Figure 3. Karyotype of Ictalurus melas (OSUM 35534, male), 2N*60. Note the two la rg e s t msm’ s . 8 f t ft A i c i r H H r t * « f * * A It A ft A ItO fttO AAAftft A A* A A I A M n A*-AA-* A QltAft * A <1 Aft A A A ^ fl K * * * * * a x ir hv xx f i M & M 0 < m ft A f t i ' t A pAft <\ 0 A A a /v A l l | I S A A A Figure 4. Karyotypes of Ictalurus nebulosus (OSUM 35502, male), 2N=60 Note the two largest msm1"^ KXX£ftftA*«» K * AAA «•«***» « ft A M A A • • A * * * * * * A a * * IlXlf )*«»•« IMlll Ills 4 I I ft A I I * A » i ft I ft R ■ • V B S 9 S 9 4 A I AA ft ft 4 1 NAt^th!*?1*88 °f If ta1urJS serracanthus {OSUM 35538, female). Note the large number o f msmTTements™ 41 H 1M K K X > 1 X X ft 4 R ft x x * fllUflftftftft** ft 1 0 «A6 n X A U X A O 8 /C )| A ft R ft ft A A \ O f l A A > « ^ a < \ » a — o * a « a A ft A IfcA A .**'" * -« Figure 6. Karyotypes of Pylodlctls ollyarls (OSUM 37001-S, not sexed), 2N=56. Note the large, distinct metacentrlc pair 1n the msm series. 42 AK A A M «« AM H I « fT * * 4 ft l « A • f l * r t f t / » • Aft A A A l t d a * DAAK a * RfiAAS«-AAft» A*AAXDAA« a * * « * XX « * a A A a Figure 7. Karyotypes of NoturuS gllberti (OSUH 35499, male), 2N=54. Note the la rg e number o f msm eTements. YlO fXJLA /tftftft V A X * v f i A R A -**A A A a ® A M * V A AAAAA A A A » M i i t t i m A I A» AAAMA A A A t rt— «**•> Figure 8. Karyotypes of Noturus insiqnis (above. OSUM 35523. fprnxlAl and Noturus exllls (below, OSUM 3?0fil, female). both 2N-54. Note the larger number of msm elements in Insiqnis (above) than in exilic (below). ------44 m n 9 M | | 0 0 «*BI*I I • » • # K U X K U / U M M « < » • t> 0 a & * * ft A IV « A A A A A A Fiqure 9. Karyotypes of floturus nocturnus (above, OSUM 34172, female), 2N=48 and Noturus leptacanthus (below, OSUM 34759, not sexed), 2N=46. Note the large, cfistinet st element In the stt series. 45 | U x x i x x * «» X » » « t l f i h fifttXkAAAAA Figure 10. Karyotype o f Noturus funebris (OSUM 37064-S, female), 2N=44. Note the large, distinct st element in the stt series. X V tf Jt tf« ft ft ft K ft K m m m * J\ 0 ft ^ * a * * * X* # X HfMliffc** V X - * *» x v a a *. « J* * « « « A ^ A /» A A A. A NFo?e7ha\-,^y°t^U ^t!™ I f F ^ e T 3W 8’ "0t S“ ed>- 2NM2' Illl II III! I I U I X | « M A i m i m i • • A A X x v n «xs ^ & A A A A 5 A K 4 i A A fl A - A ft A A A A s.5^»ss?^2% nas« A & X K-*. Y 5 4 ^ ^ A ( < I 4 4 a a a a * A A fffiffff*«x«K ****•«**,» 4 I A « j i f t a t v V fi * M 4A • Xft 3483o!Ve™NeKab o ? r ? N .S f^ d^ ^ I ^ Bla c h ^ Louisiana 0 both-These tHo ka— s°e- o o t deX t^S s„siK res IK K JUS (U K H U M M t j &R9fln#a ( M i l i H I J • II o i # » XX/I A» A A K IS, X X M K A I (* | I A ft A . » — A X A A A A AAA. AA. Figure 14. Karyotypes o f Noturus flavus from northern Ohio River Basin (above, OSUM 35510, female), 2N=48 and from Copper Creek, Clinch River Drainage (below, OSUM 35526, female), 2N=50. Note the large metacentric p a ir in Ohio flavus (above) and it s absence In the Copper Creek specimen (below). f t J} X ft X * a R I t A A #tr a 4 f| j< A * « 4 4 a a i r 4 4A AAyv#^ a a • a A ^ K A H A IKJ»» A, tf » A * J t A At A * »***•«. A * * A a a a a a a Figure 15. Karyotypes of Noturus flavlplnnls (OSUM 34655-S, female), 2N*52. Note the large number of msm elements In the karyotype of this species. 51 U x n i i t a t Aft A A ft *« BtK M A m. m 4 4 I U lllllH 4 I M | M H 6 4 A I * » * A ft A Jr y * k »» k A A A » m « | • * M d AAlt ^ ^ ^ ^ A A H • n «* Figure 16. Karyotypes of Noturus mlurus (above. OSUM 35515 n o t Itote the W a ^ d lT ? ! ^ aifeaJSf (below, OSUM 34045, malef. 2N=66. ,15 ?h’ ? Ct "A centric pair 1n the msm series of mlurus (above) and the large number of stt elements In albater (below) - 52 K K f t f i k n * x X * A (S A ft-A A -N A A A a a A A A A a a A A * A X t t A A H A ^ A A 4 I t A A X X a •» » A H X X V *. * * * « « ft ft ft I) ft Ml R Aft D M A M Afi ft A A li A A A A ftflA A AAAAAAAOAA A A £ Ai ^ A *"• H ^ ^ ^ A ^ A A» ^ A A I Figure 17. Karyotypes of Noturus albater, 2N=70 (above, OSUM 34066, not sexed) and 2H=72 (below, OSUM 34066, female). Note the large number o f s tt elements in both and the polymorphism o f 2N. 53 IfKXttAt»** ft* ft»* * * * * * t < * * * • |fKXXAi»*fc ft ft ft ft ft K A ft A A I t A s A I Figure 18. Karyotypes of Noturus elegans (OSUM 35516, not sexed), 2N®46. Note the three 1arge,d1stlnct palrs of msm elements and the large st pair in the stt series. 54 )i X ft ^ * X * A A *< x X X X X * a * a A A/1 A A SSrt-^^A A C J M 4 w x c u f t a MAI»«•*»»* * * » • ft « A Jt * * * « * * «AIMA»*»* A A Figure 19. Karyotypes o f Noturus hlldebrandl latus (above. OSUM 34837, not sexed) and Noturus hlldebrandl hlldebrandl (below, OSUM 34775, male), both 2N=46. Note the large number o f msm's and the large, distinct st pair (second largest of the msm's) 1n both karyotypes. ' H K H i H ) \ H H V i ft # 1 * g A ft » *» * * A — '' / w f t * * Figure 20. Karyotype o f Noturus fTavater (OSUM 34065, male), 2N=44. 56 1 lift I H I M i $ i h f J t l » | A 14 • M • XIfftlftftA A 4 A * 3 D H W f t * tf rf H/l M I** » Flaure 21. Karyotypes of Noturus eleutherus male (above) and female (below), 2N*42 (OSUM 34825). Note the Targe,distinct metacentrlc pair and large* distinct st pair In both karyotypes. i s * * k v x k k ** • • ff * • « H « 9 t A A A A A4S A A tfl ^ (n» M i A Jl II 4 * ^ *K V M If K ft a ■ A • m m » * • « » A fl II ft A f t A A n H A A A a 4 * 0+ Figure 22. Karyotypes of Noturus stlgmosus (above, OSUM 34196* female) and Noturus munltus (below, ^SsUM 34765, not sexed), both 2N*42. Note that the four TaYgest pairs of msm's are separable from the remainder of that series 1n both karyotypes. The karyotypes of these two species cannot be distinguished consistently. 58 KXXX X X XX JU K X X X H 4 A A X s * * X A J&flftA'tAaiff ' Y ft ft A A * 4 x ? U K ? m j U K XKAA5^Xl»« * x t x HA (V-AA A A A ' A A ^ ^ Figure 23. Karyotypes o f Noturus ta v lo ri male (above) and female (below), both 2N=40 (OSUM 34827). hote the heterotnorphic pair of chromosomes in the male (above) but not the female (below). MECHANISMS OF CHROMOSOME CHANGE The process of speciation is synonymous with change in the genic constitution of the diverging populations. It is, then, not surprizing to find karyotypic differences between species since the chromosomes bear the genes. Although gross chromosomal change is not necessary fo r divergence, i t often accompanies i t (Gold 1977). White (1973) and John and Lewis (1969) have discussed in some de ta il the mechanisms o f chromosomal change that function or could function in the evolution of karyotypes, while Gold (1977), Denton (1973) and Kirpichnikov (1973) have dealt s p e c ific a lly with those mechanisms that seem to have been important in the evolution of fish karyotypes. Changes in karyotype that often accompany transpecific evolution may be divided into two categories, those in which 2N changes and those in which chromosome form changes without a concomitant change in the 2N (Gold 1977). Changes in Diploid Number As indicated by Gold (1977) there are three primary mechanisms whereby changes in chromosome number may occur during evolution: 59 60 polyploidy, aneuploidy and Robertsonian rearrangement. Polyploidization involves an increase in the chromosome number in an integer multiple of a base number. This results in triploid (3X), tetraploid (4X), etc. chromosome numbers. Operation of poly ploidy in the evolution of a group's karyotypes is generally indicated by a bimodal (or multimodal) distribution of chromosome numbers within the group, with the higher mode(s) approximately equal to some integer multiple of the lowest mode. Several examples of probable polyploidy have been discovered in fishes. Some families for which polyploidy has been suggested include the Catostomidae (Uyeno and Smith 1972), Salmonidae (Gold 1977 and references cited th e re in ), Cobitidae (Kobayasi 1976), Callichthyidae (Hinegardner and Rosen 1972) and the Cyprinidae (Muramoto et a l. 1968, Klose et a l . 1969, Aghasadeh and R itte r 1971, Kobayasi 1971 and Gold and Avise 1976). The loss or gain of chromosomes in other than multiples of the base chromosome complement is known as aneuploidy. Aneuploidy has rarely been demonstrated in fishes; the only known case is a trisomic individual of Salmo fontinails (Davisson, Wright and Atherton 1972). As an e vo lu tio n a rily important mechanism of chromosome change in fishes, aneuploidy seems to have had little impact. Since aneuploidy would probably result in gross genetic imbalance by the loss or acquisition of large blocks of genes (a chromosome) it is understand able that it has rarely occurred. It seems much more likely that e v o lu tio n a rily important mechanisms of chromosome change would involve duplications of whole sets of genes (as in polyploidy) or a rearrangement of genes already present. Such changes would minimize 61 the upset to balanced sets of coadapted genes present and probably be more easily tolerated in evolution. The fin a l, and probably most important, mechanism a lte rin g 2N 1n fishes is Robertsonian rearrangement. As currently conceived and used in lite ra tu re on fis h cytotaxonomy, Robertsonian rearrangements refer to the formation of a biarmed (m or sm) chromosome from two nonhomol- ogous uniarmed (st or t) chromosomes; the converse would be the formation of two uniarmed elements from a biarmed chromosome by some mechanism of centric dissociation (^fission?). Such rearrangements have often been implicated as a major mechanism in the evolution of fis h karyotypes {Manna and Prasad 1971, Gold 1977, LeGrande and Fitzsimons 1976, Denton 1973 and others). Most cytogeneticists agree on the probable mechanism involved in centric fusion, as reviewed by White (1973). Centric fusions are essentially whole arm translocations (see White 1973, p. 224, Figure 7.12). Simultaneous breaks are thought to occur in two nonhomologous uniarmed chromosomes in or around the centromeric region of each. The major chromosome arms of each then rejoin (translocation) forming a large biarmed chromosome and the small centric or acentric fragments fuse forming a chromosome fragment that is lo st in succeeding genera tions. Centromeres usually lie in regions of genetically inert hetero chromatin (White 1973) so the small acentric or centric fragment lost is probably of little genetic consequence. While l i t t l e dissention seems to surround the proposed mechanism of centric fusion, there is a great deal of controversy over the facil ity with which centric dissociation (=fission?) may occur (White 1973, John and Lewis 1969). This stems mainly from the acceptance or rejection of the telomere hypothesis. The telomere hypothesis suggests that there are natural chromosome "ends" which may not become in te r stitia l and that fresh interstitial breaks are unstable in a terminal position (White 1973). A byproduct of this hypothesis is that truly terminal (=Telocentric of Levan, Fredga and Sandburg 1964) centromeres are non-existent (White 1973). Although centric fusions are easily reconciled with the telomere hypothesis (White 1973), simple centric fissions are not since nontelomeric terminal centromeres would be created by simple fission of the centromere. White (1973, p. 227, Figure 7,14) postulates that centric dissociations occur only where there are telomere donor chromosomes present. Dissociations, then, are thought to be less likely to occur than centric fusions because they are a more complicated process and evolutionarily less parsimo nious (White 1973). Thompson (1976) has rightly pointed out that these telomere donors are lost from the complement in White's model of centric dissociation so that the actual 2N is not changed. Thompson theorized increases in chromosome number w ith in the bounds of the telomere hypothesis by concurrent centric fissioning and pericen tric inversion. Although an intriguing possibility, such a rearrange ment is s till incompatible with the telomere hypothesis. The pericen tric inversion following centric fissioning in Thompson's model would leave exposed, freshly broken, nontelomeric ends on the second arms of the new chromosomes. John and Lewis (1969) suggest that simple centric fissions are a tenable mechanism for increasing the diploid number by the formation of chromosomes with truly terminal centromeres. 63 The exact mechanism of centric dissociation is unexplained at the present time. As w ill be seen in the following discussion of the distribution and evolution of chromosome numbers in the ictalurids, non-polyploid increases in number occur and some mechanism allowing centric dissociation must be operative in certain fish groups. Thompson (1976) has also shown similar increases in 2N in certain cich lid s. In most groups o f fishes with a reasonable large number of species studied chromosomally, most deviation from the presumed ancestral 2N is toward decrease {see below). This suggests that increase in the 2N is a rare ly exploited mechanism of chromosomal evolution in fishes. The mechanism for centric dissociation may be more complex, thus less parsimonious, than that for centric fusion. Centric dissociations are probably less likely to occur and become fixed during evolution in fishes than are centric fusions. Changes Not A ffecting Diploid Number An array o f chromosome changes that would not a ffe c t the 2N is possible. Such rearrangements could include non-Robertsonian trans locations, deletions, para- and pericentric inversions, duplications, etc. With current cytogenetic techniques being employed with fishes, the vast majority of such changes would probably go undetected. Two species with apparently identical karyotypes may have dramatically different gene arrangements. Considering the well documented influence of gene arrangement on fitness (Position Effect) in Drosophila (Dobzhansky 1970) i t is reasonable to assume that the same phenomenon is important in other organisms as w ell. The importance o f structural rearrangements of chromosomes in evolution has been the object of a series o f a rtic le s by Wilson and his coworkers (Wilson 1975, Wilson, Bush, Case and King 1975 and references cited in those works). Wilson points out that the effect of a gene may be altered with a change in its position relative to its regulator genes. Structural rearrange ments may, then, be as important ( if not more so) as point mutations in providing genetic novelties upon which selection may act. This would, fo r once, give some credence to the idea that differences in the karyotypes of related species have selective value and are not simply the re su lt o f fix a tio n of random chromosome changes that have had no evolutionary consequences. Cytotaxonomically, the most important rearrangements falling into the category of those not affecting 2N are ones which can be identified cytologically by their effect on centromeric position, including principally unequal pericentric inversions and translocations. Translocations of unequal segments between nonhomologous chromosomes could produce id e n tifia b le changes in the arm ra tio of both chromosomes. This may or may not a ffe ct the FN depending on the size of the translocated segments. Unfortunately, in fishes, interspecific differences due to unequal translocations are d iffic u lt to assess and track due to the small size and high number o f chromosomes in most species. Probably the most easily tracked changes in fish karyotypes are unequal pericentric inversions that a lte r FN. Such rearrangements are detectable as differences in the number of chromosomes bearing major 65 second arms (metacentrics and submetacentrics), with no change 1n the relative size of the elements (%TCL) 1n the complement. As Manna and Prasad (1971) have pointed out, pericentric Inversions, 1n addition to Robertsonian fusions, have probably played a more important role 1n the evolution of the karyotype in fishes than many previous workers had thought. ICTALURID RELATIONSHIPS: HISTORICAL REVIEW The most comprehensive study of ictalurid relationships, to date is that of Lundberg (1970). He has reviewed c ritic a lly the fossil record of the Ictaluridae and compared i t extensively with the osteol ogy of many extant species. Taylor's (1969) systematic revision of Noturus provides the most encompassing study of relationships within that genus. Lundberg (1970) concluded that ictalurids appear to form a monophyletic assemblage of species, based on the following synapo- morphic osteological character states: 1. Extensive invasion of jaw musculature on the skull roof. 2. Articulation of the anterior process of the supracleithrum with the ventral surface of the pterotic wing. 3. Presence of a sublateral process of the premaxilla. 4. Absence of vomerine teeth. In addition to their shared, derived features, the ictalurids have retained several plesiomorphic features, indicating that the 66 67 family is representative of relatively primitive siluriforms. According to Lundberg (1970)* some characters indicative of the family's relative primitiveness are: 1. Broad sutural contact between the metapterygoid and anteroventral margin of the hyomandibular. 2. The endopterygoid is often a large plate-like element that is tightly bound to the metapterygoid. 3. The membranous sheet of bone between the anterior and posterior limbs of the transverse process of the fourth vertebra is not extensive. 4. Possession of a moderately large "neural" complex. 5. Presence of a small rib associated with the f if t h vertebra in young ictalurids. 6. Many species with six separate hypurals and a discrete parhypural. 7. The secondary hypurapophysis is always restricted to the first hypural. 8. Presence of a maxillary tooth patch in Hypsidorus farsonensis, an extinct ictalurid (the only extant siluriform not having lost such maxillary teeth is the prim itive South American Piplomystes) Ictalurids most closely resemble members of the Afro-Asian catfish family Bagridae, and may be "derived" from them (Lundberg 1970, 1975). The members of the subgenus Ictalurus have retained 68 more characteristics of primitive silurlforms than any other ictalurids. Lundberg (1970, p. 135) states: "Ictalurus furcatus ...in overall anatomy... has probably retained more primitive features than any other liv in g species of ic ta lu rid ." This does not imply, however, that I_. furcatus is without specializations of its own. Some derived features of I_. furcatus that have been pointed out by Lundberg include: 1. Supraoccipital process longer than in any other ictalurid and most other catfishes. 2. Depth of the Weberian complex is greatly increased by the elongation of the neural spine on the fourth vertebra. 3. I t has more caudal vertebrae than other ictalurids. 4. The anal fin is elongated. 5. The swimbladder is completely separated into an anterior and posterior chamber by a partition. Lundberg (1970) presents evidence that Noturus is a monophyletic group on the basis of shared, derived character states that include: 1. A subcutaneous eye. 2. Enlarged foramen for the superficial opthalmic branches of the f if t h and seventh nerves. 3. Lateral foramen placed fa r anteriorly in the lateral ethmoid. 69 4. Larger lower jaw than Its sister genus P rietella. 5. Lengthened, generally toothed, sublateral process of the premaxilla with no lamina between the process and premaxilla. 6. Elongated palatine. 7. Extremely narrow opercle with an angle between the dorsal and anterior edge of about 128°. 8. Generally, a reduced number of hypurals by the loss of the uppermost element and irregular patterns of fusion between the remaining bones. 9. Reduced hypurapophyses. 10. Increased number of branched caudal rays. 11. Facets on the second dorsal basal for the spine are markedly concave. He concurs with Taylor (1969) that the subgenus Noturus is probably more closely related to the subgenus Schi1beodes than to Rabida. Since Lundberg's study included only a single representative of each subgenus of Noturus, he was unable to add any new insight into intrasubgeneric relationships. In that respect, Taylor's study remains the most comprehensive. Lundberg (1970, 1975) also agrees with Taylor (1969) that Pylodictis and Noturus are probably more closely related to each other than either is to other ictalurids, although Pylodictis has retained some primitive features (Lundberg 1970). 70 The bullheads present a problem in terms of intrafamilial relationships. Taylor (1954) synonymized them with the genus Ictalurus, degrading Amiurus to subgeneric status. This was based on Taylor's belief that X- (Amiurus) catus was morphologically intermedi ate between Ictalurus (sensu s tric to ) and Amiurus. Lundberg (1970) states, however, that X- (A.) catus is a "perfectly good bullhead" osteologically and that there are few "certainly derived characters" that are shared between Ictalurus (£.) and Amiurus that must be derived independently i f the bullheads are to share a more recent common ancestry with the Noturus- Pylodictis lineage than with Ictalurus (s_.s_,). Within the subgenus Rabida, Taylor erected four species groups of supposedly closely related forms: 1. hiIdebrandi group (N^ hiIdebrandi and baileyi) 2. elegans group (N. elegans and trautmani) 3. furiosus group (N. furiosus, placidus, munitus and stigmosus) 4. miurus group (N_. miurus, flavipinnis and flavater) Taylor did not, as Lundberg (1975) indicated, place fi. albater in the hi Idebrandi group, N. eleutherus in the elegans group or N_. ai 1 berti in the funebris group. Each of those groups was viewed as a closely related assemblage of species by Taylor. He indicated that the hiIdebrandi and elegans species groups were probably more closely related to each other than 71 to other species or groups within Rabida. S im ilarly, the miurus and furiosus groups were depicted as sharing most recent common ancestry. Noturus albater was not viewed as an intimate relative of any other madtom species, but Taylor noted that i t shared some features with both the hildebrandi and elegans groups and placed i t closer to N. elegans in his phylogram (Figure 24). Similarly, the affinities of eleutherus were not intimate, but i t was placed closest to the furiosus and miurus groups, being somewhat intermediate between those two and the hildebrandi and elegans groups. Within this scheme, N_. hildebrandi was viewed as a relatively prim itive Rabida. The species in Schilbeodes were, according to Taylor (1969), more divergent than those in Rabida and i t was suggested that they might represent a "heterogenous" (=polyphyletic?) assemblage. The presence of a terminal mouth and ten preopercular-mandlbular pores in N. gyrinus, lachneri and exi1is was thought indicative of a closer relationship between these three species than with other madtoms, although exi1 is had diverged significantly in other characters {e.g., possession of nine pelvic rays, presence of serrations on the posterior margin of the pectoral spine, more depressed head). N_. gyrinus was proposed as the most primitive of the three. Noturus insignis, prior to Taylor's monograph, had been considered an intimate relative of exi1 is due to their sharing striking morpho logical similarities, including an elongate body, long anal fin and dark marginal bands in the median fins. Taylor, however, considered these characters to be convergent rather than indicative of PLEASE NOTE: Print on some pages is small and indistinct. Filmed in th best poss ible way. UNIVERSITY MICROFILMS. 00 ft © CO cc CO CO © NOTlftUS F1SUre d^?oNr,^L^,Sr„XrsaSS ^ . Tayl0r (1%9)' N“* erS f -J ro 73 relationship between the two. He considered ins ignis to be more closely related to nocturnus and that these two were probably moderately close to the funebris group. The funebris group was composed of the similar, but allopatric species funebris and phaeus. The relationships of leptacanthus did not seem to be close to any other Schilbeodes, indicating possible early divergence. Taylor considered giIberti to represent the most highly specialized member of the subgenus, but placed i t near the base o f the Schilbeodes lin e in his phylogeny. He pointed out that i t shared some morphological characters with the supposedly primitive N^. (Noturus) flavus, including cream-colored margins on the caudal fin and a relatively short, weak dorsal spine. Taylor's primary criterion for retaining flavus in a separate subgenus was tha t a ll specimens examined by him retained the pectoral radials in the plesiomorphic, unfused condition. Lundberg (1970) noted a d d itio n a lly that the retention o f an unreduced body la te ra l lin e canal system was another plesiomorphic character state found in fla vu s. Noturus (N_.) flavus has, however, apomorphous char acters such as an elongate, posterior process of the premaxillary tooth patch and 10 pelvic rays. ICTALURID CHROMOSOME COMPLEMENTS The most striking feature of the ictalurid karyotype is the d iv e rs ity of 2N seen in such a compact assemblage of species. Diploid numbers were found to range between 40 in ta y lo ri and 72 in N. albater (Figure 25, Tables 1 and 2). There was l i t t l e overlap in 2N (Figure 25) between Noturus and other ictalurids (Pylodictis and Ictalurus). Madtoms (Noturus) were responsible for a large part of the diversity of ictalurid 2N's. Noturus 2N*s ranged from 40-54 and 66-72 (N_. a lb a te r) , while other ic ta lu rid s had 2N's between 48-62, but mostly between 56-62 (Figure 25, Tables 1 and 2). There was no strong modal 2N as seen in most other fis h groups that have had a reasonably representative sample of their constituent species examined cytologically. The general trend, at least in fishes, is toward a strong modal 2N with the variation from this mode skewed toward hypomodal values. Illustrated in Figures 26, 27 and 28 are the distributions of reported 2N's in three of the most studied orders of fishes, the Cypriniformes, Perciformes and Atheriniformes. These distributions were generated by me primarily from the checklists o f chromosome numbers in fishes by Denton (1973), C h ia re lli and Capanna (1973) and N iko l'skiy and Vasil'yev (1973) as well as some of the more recent primary literature. Duplicate 74 SPECIES 10 iue 5 Dsrbto o dpod ubr i te Ictaluridae. numbers the in diploid of Distribution 25.Figure Vm iV & & & rrr#* NOTURUS YOITS ICTALURUS & PYLODICTIS ILI NUMBER DIPLOID UV *si 76 no CYPRINIFORMES 171 s d . in 5 Families 2N= 50+2 (82%) 100 M SO OIPLOID NUMBER Figure 26. D is trib u tio n o f d ip lo id numbers in the C ypriniform es. 77 110 PERCIFORMES 172 sp. in 29 Families ioo 2N= 48+2 (82%) •o SO TO so so 40 30 20 10 <40 SO >eo d i p l o i d m o m o c h ibution of diploid numbers in the Perciformes. 78 ATHERINI FORMES 217 sp. 1n 6 Families 2N= 48+2 (69%) 10 20 30 40 SO 60 DIPLOID NUMMn Figure 28. Distribution of diploid numbers 1n the Atherlnlformes. 79 reports o f the same chromosome number fo r a p a rtic u la r species were recorded only once. Where two or more d is tin c t 2N's had been re ported for a single species, generally no attempt was made to decide which was probably correct, and both were recorded in the d is trib u tio n . The only exceptions were cases where there was some reason to doubt the v a lid ity of one or more o f the reported numbers (e.g., old reports based on poorer, pre-1960 techniques are often suspect). Obviously polymorphic species were not included. Since some species have duplicate numbers recorded, the number of species indicated for each distribution may be a slight overestimate of the true number in each taxon fo r which 2N's are known. However, general trends in the distribution of 2N's in major fish groups are apparent. Consider the distribution of 2N's in the order Cypriniformes (Figure 26). This is an interesting comparison since cypriniforms are a sister group to the Siluriformes, the order containing the Ictaluridae. Together, these two orders form a large, closely related assemblage of primary freshwater fishes known as the Ostarlophysi. Within the Cypriniformes, 171 species in 5 families have been examined cytologically (Figure 26). There is a strong mode at 2N=50, with most species deviating from this mode having 2N's of 48. Of the 171 species, some have been shown to represent examples of polyploidy (i.e ., most of those species indicated as >60 in Figure 26), such as some cyprinids, some cobitids and catostomids; if these are removed from this distribution, the mode would be even more impressive, 89% of 157 species with 2N=48-52. In other words, cypriniforms have apparently undergone little change in 2N throughout their evolutionary history. The trend toward strong modal 2N's is equally impressive in the largest, and probably most diverse, order of extant fishes, the Perciformes. Some 172 species in 29 families have been studied chromosomally {Figure 27). There is a strong mode at 2N=48, with less than 3% having a hypermodal number. About 82% of all perciform fishes for which a chromosome number is known have 2N's o f 48+2. The strong mode with almost no hypermodal values indicates conservative number change in perciforms, with increases playing little role in the evolution of their karyotypes. Within the order Atheriniformes (Figure 28) about 217 species in 6 families have been studied. The majority of these are representative of the large cyprinodontoid families Cyprinodontidae and Poeciliidae. Once again, we see a strong mode (at 2N=48) w ith 69% of the species analyzed having 2N-48+2; less than 3% have hypermodal numbers (>48). This d i s t r i bution is quite similar to that seen in perciform fishes. Consider ing s t r ic t ly North American representatives o f fis h fam ilies that have been karyotyped, those families having received the most attention include the Cyprinidae (Hubbs, M iller and Hubbs 1974), the Centrarchidae (Roberts 1964), the Cyprinodontidae (Chen and Ruddle 1970, Chen 1971 and Setzer 1970) and the Percidae (Ross 1973, Cavender, Ross and LeGrande unpublished data), each with at least 20 species in two or more genera having been studied. A ll o f these fam ilies are as speciose or more so than ictalurids in North America. In each case, there always appears a strong modal 2N: 50 in North American cyprinids, 48 in centrarchids, percids and North American 81 cyprinodontids. In s ilu rifo rm fishes (Figure 29, Tables 1, 2 and 4) there is a bimodal d is trib u tio n of 2N's (54 and 58) with only 34% of the species studied with 2N=56+2. Since 30 of the 67 species studied are ic ta lu rid s , i t cannot be concluded that extreme d ive rsity w ill be characteristic of the whole order. Removing ictalurids from this distribution (see Figure 29) leaves a weak mode of 2N=58 with about 30% of the 36 remaining species with 2N+58+2 and 44% with 2N=56+2. Extensive sampling of additional siluriform species in different families w ill be needed to reveal how extensive chromosomal rearrangements affecting diploid number have been in the evolution of silurifo rm s. For the present, however, i t appears that 2N changes have occurred to a much greater extent in siluriforms than in cypriniforms, perciforms or atheriniforms (which together probably represent about 3/4 of known species of living fishes). Of the major fish groups having been studied in some detail, the Salmoniformes also have diverse karyotypes. However, i t is generally agreed (Gold 1977) that polyploidy has been involved, at least to some extent, in the evolution of these karyotypes and may explain some of their diversity. There is no strong evidence to suggest extensive polyploidization as important in siluriforms, although Hinegardner and Rosen (1972) have shown increases in DNA values correlated with increased 2N's in some siluriform species; this indicates that some isolated cases of polyploidy are probably present in catfishes. Hudson (1976) has recently presented DNA SPECIES 10 ^ iue 9 Dsrbto o dpod ubr i te Siluriformes. the numbers in diploid of Distribution 29. Figure ILI NUMBER DIPLOID ER ™ ° v AV> v S > ICTALURIDAE SILURIFORMS 6 p i 9 Families 9 in sp.36 Families 10 in sp.67 Without Ictalurids Ictalurids Without 2N= 58+2 (30*) 2N= 56+2 (34*) 2N= 58+2 (24*) 2N= 56+2 (44*) S I L U R I F O R M E S <-‘*V 131 rs> CD Table 4. Karyotype data for non-ictalurid siluriform fishes. Species ______Family______2N______FN_____ Source Mystus gu]io Bagridae 58 98-10; Natarajan and Subrahmanyam 1974 Mystus guTio Bagridae 58 — Khuda-Bukhsh and Manna 1974 Mystus mystus Bagridae 54 — Muramoto, Ohno and Atkin 1968 Mystus seenghala Bagridae 50 80 Srivastava and Das 1969a Mystus tengara Bagridae 54 102 Rishi 1973; Nayyar 1966 Mystus vittatus {form A) Bagridae 54 98 Manna and Prasad 1974 Mystus vittatus (form B) Bagridae 58 84 Manna and Prasad 1974 Mystus vittatus Bagridae 50 — Srivastava and Das 1969b Pelteobagrus nudiceps Bagridae 56 -- Nogusa 1960 Rita rita Bagridae 54 Nayyar 1966 Clarias batrachus Clariidae 50 Prasad 1971 Heteropneustes fossi l is Heteropneustidae 56 96 Srivastava and Das 1968 Heteropneustes fossil is Heteropneustidae 58 -- Nayyar 1966 Hypostomus plecostomus Loricariidae 54 82 Muramoto, Ohno and Atkin 1968 Loricaria parva Loricariidae 48 — Post 1965 Clupiosoma garua Schilbeidae 66 — Nayyar 1966 Kryptopterus bichirris Siluridae 60 Hinegardner and Rosen 1972 flmpok bimaculatus Siluridae 40 Nayyar 1966 Parasilurus asotus Siluridae 58 -- Nogusa 1960; Muramoto, Ohno and Atkin 1968 Mallago attu Siluridae 86 _ _ Nayyar 1966 Rhamboia Taticauda Pimelodidae 58 100+4 Present study Rhambdia sp. Pimelodidae 58 100+4 Present study Table 4. Continued. Species ______Family______2N FN Source Arius fe lis Ariidae 54 80 Present study Corydoras aqassizi Callichthyidae 98 180 Schee et al 1972 Corydoras aneus Callichthyidae 132+ 222+ Schee I et a! 1972 Corydoras arcuatus Callichthyidae 46 92 Schee et al 1972 Corydoras axel rodi Callichthyidae 46 92 Schee et al 1972 Corydoras bondi Callichthyidae 46 92 Schee * et al 1972 Corydoras elegans Callichthyidae 50 100 Schee et al 1972 Corydoras julii Callichthyidae 92 92 Schee 9 et al 1972 Corydoras melanistus Callichthyidae 46 92 Schee et al 1972 Corydoras metae Callichthyidae 92 180 Schee * et al 1972 Corydoras aff. osteocardus Callichthyidae 76 126 Schee et al 1972 Corydoras paTeatus Callichthyidae 44 88 Schee 9 et al 1972 Corydoras rabauti Callichthyidae 58 112 Schee 9 et al 1972 Corydoras schultzei Callichthyidae 58 106 Schee 9 et al 1972 Corydoras schwartzi Callichthyidae 46 92 Schee f et al 1972 00 85 values for erythrocytic nuclei in 11 species of ictalurids, showing elevated DNA contents in Noturus insignis and gyrinus. There is no correlation between the elevated DNA content and chromosome number that might be construed as evidence of polyploidy. In fact, these species have 2Nls significantly lower than in other icta lu rid species in either Hudson's (1976) study or mine. I f these high DNA values are confirmed, they may represent cases of reduplication, polynemy, addition of heterochromatin or some other non-polyploid increase in DNA. That certain fish groups have conservative karyotypes, while others seem to have extensive chromosomal rearrangements affecting 2N remains largely unexplainable. Inter- and/or intraindividual (intraspecific) variation in chromosome number does not seem to be widespread in the family. I f the diversity of 2N represented in ictalurids is simply a pro life ra tion of a propensity for random chromosome change, we should see a reflection of this in the form of extensive chromosomal polymorphism. Only two cases of polymorphism of 2N have been seen in this study: N^ albater and flavus. Hudson (1976) indicates a great deal of variation in chromosome form within species of ictalurids, but con sidering the large number of elements possessing second arms in the icta lu rid karyotype and the numerous chromosomes that are on the borderline between subtelocentric and submetacentric, these variations are probably more apparent than real. Bogart (1969 in Thompson 1976) has shown that long and short chromosome arms may contract at d iffe r ent rates, thus a subtelocentric element may measure as a 86 submetacentric in overly contracted material. This makes numerical analysis of a few karyotypes from a small sample of the population (such as that carried out by Hudson) of doubtful value since it is d ifficu lt to place accurately all elements of a karyotype into one morphological category or another. Strict numerical analysis of chromosome measurements and arm ratios in fishes should be viewed cautiously at this time. Another possibility that could explain variation in 2N's between members of a family of fishes is that the differences are postmating isolating mechanisms, ensuring failure (or at least sterility) of hybrid offspring that might be produced. Taylor (1969) has noted that, at least in Noturus, natural hybridization is extremely rare. The hybrids gyrinus X miurus (Trautman 1948, Taylor 1969 and Menzel and Raney 1973) and N_. e x ilis X miurus (Taylor 1969) are the only reported natural hybrid combinations among madtoms. Of the former cross, eight specimens are known and only one specimen in the latter. In neither case is there any sound evidence that the hybrids are fertile. There is no documentation of natural introgression 1n ictalurids. This scarcity of hybrids among Noturus points to the presence of very strong iso la tin g mechanisms. Among other ic ta lu rid s , the only case of extensive hybridization (possibly introgression?) is between I_. nebulosus and K me!as (Trautman 1957), two closely related and broadly sympatric species sharing a common 2N (60) and identical karyotypes. The extremely low frequency of hybrids in Noturus is interesting since many madtom species are sympatric. At many localities, several 87 species of Noturus may be taken within a lim ited stretch of stream. Illustrated in Figure 30 is a comparison between probable sympatric Noturus species 1n this study and those which share a common 2N. The data on sympatric species combinations were derived prim arily from the distribu tion maps presented by Taylor (1969). Species indicated as occupying the same drainages in at least a portion of th e ir range were accepted as sympatric whether or not they were noted to have been taken together in the same collections. There are 190 possible interspecific pairs between the 20 species studied. Of that, 63 (33%) of the pairs are sympatric while 127 (67%) represent species pairs with nonoverlapping distributions. Twenty-three (12%) of the possible pairs share the same 2N; 11 of that 23 are pairs of species in the same subgenus while the other 12 are species pairs in diffe ren t subgenera. Nine (14%) of the sympatric pairs share a common 2N (4 pairs in the same subgenus, 5 in d iffe re n t subgenera). Only 14 (11%) of the a llo p a tric pairs share a common 2N (7 in the same subgenus, 7 in d iffe re n t subgenera). Thus, there seems to be no marked difference in the percent of a llo p a tric vs. sympatric pairs (11% vs. 14% respec tively) that share a 2N. In fact, a slightly higher percent of sympatric pairs have the same 2N. Even when two sympatric species of Noturus share a diploid number (i.e ., gilberti-insignis, nocturnus-flavus, gyrinus-phaeus, gyrinus-eleutherus, gyrinus-stigmosus, gyrinus-munitus, and eleutherus- stigmosus) each always possesses a karyotype distinguishable from the other member of that pair on the basis of chromosome morphology (i.e . centromeric positions). For example, eleutherus and stigmosus are 88 in»ignj» • — Probable Sympatry (127) • « l MS ■ - Common 2N (23) gtlbarti — Common 2N + Probable Sympatry (9) Mivui rtoctuffius laplacanthua lunabria pba*ws gyrlnua I ic h n tr I I la vlptnnla miurus hiidabrandi • lagans a lb a ta r (lavaisr alautharua stigmoaua munliu* lay lori Figure 30. Comparison of probable sympatry among the Noturus species studied and those which share a common diploid number. 89 sympatric over a large portion of their ranges (Taylor 1969). They have often been collected in close association with each other (Trautman 1957). They share a common 2N (42) but they can be d is tin guished karyotypically since eleutherus possesses a large distinct m pair not seen in stigmcsus. Similar differences could be elucidated for all of the species pairs just mentioned. In fact, there are only two pairs (both allopatric and closely related species pairs) among the twenty species studied that share indistinguishable karyotypes (gyrinus-lachneri and stigmosus-munitus) . From these observations one might infer that the differences in chromosome number or karyotypes may be important postmating isolating mechanisms that prevent production of F^ hybrids. This is an equivocal conclusion, however, considering the reported hybrid combinations among madtoms both involve species pairs belonging to different subgenera and d iffe rin g in 2N (2N=50 in miurus vs. 54 in e x ilis and 42 in gyrinus). The combination taken most frequently, albeit rarely, is the one with the greatest 2N difference (2N=42 gyrinus X 2N=50 m iurus). Also, viable F^ hybrids have been produced under laboratory conditions between a variety of species differing in 2N or karyotype in the genera Ictalurus and Pylodictis (Hudson 1976). If genetic or cytogenetic barriers to the production of F^ hybrids were present, we would expect the natural breakdown of these more frequently in closely related species (same subgenus) that share more similar karyotypes. For example, Noturus qiIberti and ins ignis are members of the same subgenus (Schilbeodes) and share a common 2N; th e ir karyotypes differ only in FN. No hybrids between these two species 90 are known, even though they live 1n such close proximity that they may be taken in the same seine. In summary, the evidence does not seem to support the idea that karyotype differences in Noturus are strong cytogenetic barriers to the formation of at least viable (possibly heterotic according to Menzel and Raney 1973) F^ hybrids. Whether these hybrids are f e r t ile is not known. Setzer (1970) has shown, however, that in a population of Fundulus notatus (2N=40) and Fundulus olivaceus (2N=48) that backcrosses as well as F^’s were present as indicated by various intermediate karyotypes. The karyotypic differences between these two closely related sibling species could be totally explained as simple Robertsonian fusions. The implications of Setzer's work are clear; Robertsonian changes alone appear in s u ffic ie n t to prevent F^ hybridization or even introgression (at least in these two Fundulus). Hybrids between these sibling species were obviously fe rtile even though they would have apparently been heterozygous fo r 4 centric fusions. That is, in the F^ hybrid (2N=44) there would be 4 m's (from notatus) that would each be homologous to d iffe re n t pairs of nonhomol- ogous t's (from olivaceus). Four trivalents would probably be formed (see below) at meiosis in the hybrids and proper segregation of the elements o f these triva le n ts (forming genetically equivalent gametes) most lik e ly occurs fo r the F-j's to be fu lly f e r t ile ; F^ f e r t i l i t y Is indicated by the backcross individuals present in the population. Established Robertsonian rearrangements are obviously in s u ffic ie n t to disrupt proper synapsis and segregation of the chromosomes at meiosis in the hybrids. 91 In an individual heterozygous for a simple Robertsonian rear rangement, the two arms of the biarmed element would each be homolo gous to one of two nonhomologous uni armed chromosomes. The two appropriate uniarmed chromosomes could each pair with th e ir homolo gous arm o f the biarmed element, forming a triv a le n t. Segregation at Anaphase I could occur such that the biarmed element goes to one pole while the two uni armed elements go to the opposite pole. This would result in equal distribution of the genes f.uni- on the hetero- morphic chromosomes between the two daughter c e lls , thus forming genetically "normal" gametes. If the trivalent did not segregate as just described (obviously a realistic possibility) then the hetero zygote would produce some gametes with extra or deleted sets of genes. This would probably result in reduced fe rtility (possibly sterility) of the heterozygote. The case of Fundulus hybridization ju s t mentioned as well as the occurrence of perpetual heterozygosity for a Robert sonian fusion in the males of two species of Mexican cyprinodontid fishes (Uyeno and M ille r 1971, 1972) suggests that there is not always a mechanical b a rrie r to synapsis and proper segregation in hetero zygotes for centric fusions (or dissociations, for that matter). In both cyprinodontids, the females are always X-|X^X2X2 (a ll t's ) while the males are always X^X^Y (the Y being a large m). The X-j is homol ogous to one arm of the Y, while X 2 is homologous to its other arm. The Y is thought to have been formed by centric fusion of an original uniarmed Y to an autosomal acrocentric (X2). The males, then, might be considered perpetually heterozygous for a centric fusion. At meiosis in the males, trivalent formation like that described above 92 could be predicted; each arm of the biarmed element would synapse with different X's. If a large proportion of those trivalents segregated such that one o f the t's accompanied the m to the same pole, then f e r t i l i t y would be greatly reduced since many o f the gametes would have extra or deleted sets of genes. The evolution of such a chromosomal system seems un likely under those circumstances. If the differences in 2N between two related species are the result of Robertsonian rearrangements alone, then hybrids between them could be viewed as heterozygotes for simple centric fusions (or dissociations). Since mechanical barriers to synapsis and proper segregation of the trivalents that would be formed at meiosis in the hybrids cannot be assumed apriori, then such a hybrid might (and, I believe, usually would) be fu lly fe r tile . I f other chromosomal rearrangements (e.g., non-Robertsonian translocations and/or pericentric inversions) become established in each of the two related species in addition to the Robertsonian changes, then i t is much more likely former linkage groups would have been disrupted sufficiently that hybrids between them would experience pairing and segregation difficulties; that would probably result in sterility or reduced f e r t i l i t y . Such may well be the case in ic ta lu rid s because (see below) the differences observed in th e ir karyotypes are not to ta lly explain able on the basis of Robertsonian change; other rearrangements must be invoked to explain the variation. Considering, then, the evidence that natural hybrids do occur between Noturus in different subgenera (and differing markedly in their karyotypes, also) as well as the production of laboratory hybrids (F-|) between Ictalurus and Pylodictis species, the genic constitution of ictalurids probably has not been sufficiently altered during their divergence to prevent hybridization. There are probably no cytogenetic barriers to the formation of F^'s, but apparently sub sequent non-Robertsonian rearrangements that have become established in each species would keep such hybrids from being f e r t ile ( i. e . , the karyotype differences could s till act as postmating isolating mechanisms). Finally, the extremely low frequency of natural hybrids in the absence of any apparent genetic or cytogenetic barriers to the production of F^'s suggests that strong premating iso la ting mechanisms (ecological or ethological?) are probably more important in the pre vention of hybridization. So little is known about the biology of most madtoms that it is impossible to single out what specific premating mechanisms are operative in the maintainance of the integ rity of sympatric Noturus. In the early stages of madtom divergence, interspecific chromosomal differences that arose in allopatric stocks may well have played (and continue to play) an important role in the prevention of introgression by inducing hybrid ste rility when karyotypically d is tin c t stocks returned to sympatry. With time, more e ffic ie n t premating mechanisms might have developed in sympatric stocks isolated be postmating mechanisms. ANCESTRAL DIPLOID NUMBERS The d iv e rs ity of ic ta lu rid chromosome complements exacerbates the problem o f deducing the ancestral state o f the ic ta lu rid karyo type and derived conditions in extant species. Before evolutionary interpretation of the karyotypes and interrelationships is undertaken, a hypothetical ancestral 2N should be proposed. Lundberg (1970, p. 10) has reiterated c rite ria fo r deducing (at least for osteological characters) ancestral states. These are: 1. Distribution of states in outside, but related groups. 2. Ontogenetic precedence, i. e . , a state is lik e ly to be ancestral if it appears earlier in ontogenetic development than another state. 3. Structural intermediacy in a serially homologous sequence. 4. Paleontology. Obviously, at least two of these criteria (ontogenetic precedence and paleontology) w ill not be applicable to the establish ment of ancestral karyotypes. In the case o f ontogenetic precedence, 94 95 the karyotype is viewed, for all practical purposes, as a fixed character state within any one individual (usually species) and generally does not undergo ontogenetic change. Secondly, since it is impossible to karyotype fossils, the knowledge of the chromosomal constitution of extinct species must forever remain a matter of speculation. This essentially leaves two tenable criteria for arriving at an estimate of the ancestral 2N: distribution of karyo- typic states in outside groups and structura l intermediacy. If one looks at the distribution of 2N's in other siluriforms as well as within the various lineages of the Ictaluridae, the 2N's that seem to occur consistently are those between 54 and 58 (Figure 29, Tables 1, 2, and 4). I should point out, though, that ictalurids are now the most studied siluriform family as far as chromosomes are concerned and that there is only meager information available on the chromosome complements o f n o n -icta lu rid catfishes. Of the information now available, though, the species in the families 8agridae and Pimelodidae that have been studied are of particular interest and importance in interpreting the ancestral condition of the ic ta lu rid chromosome number. These two fa m ilie s, along with the Ariidae and Doradidae, are thought to represent a series that is close to the basal s ilu rifo rm stock (Gosline 1975 a, b). Of these, the Bagridae and Pimelodidae, together with the Icta lu rida e form a group of closely related families (Gosline 1975b), with the Ictaluridae restricted to North America, the Pimelodidae to South and Central America and the Bagridae to A frica and southeast Asia. Lundberg (1970, 1975) suggests that in several features (e .g ., presence o f 96 nasal barbels, muscle crests on the hyomandibular, longitudinal dermal ridges on the supraoccipital, possession of 6 or more infra orbitals and the reduction of the sphenotic spine) the Ictaluridae most closely approach certain bagrid subfamilies (Bagrinae and Bagroidinae) and may be most closely related to or even derived from bagrids. Of the 11 bagrids and pimelodids karyotyped, 9 (82%) have 2N=56+2. Of all non-ictalurid siluriforms studied (38 species), 42% have 2N=56+2. The only a riid and pimelodids that have been karyotyped were in the present study; both pimelodids (Rhamdia spp.) had 2N=58, while Arius felis had 2N=54. Within the family Ictaluridae, 2N's of 56+2 occur in all major 1ineages: Genus Ictalurus: Ictalurus punctatus and furcatus 2N=58 Genus Pylodictis: Pylodictis olivaris, 2N=56 Genus Noturus: Noturus e x i1 i s , ins ig n is , g ilb e r t i, 2N=54 From extensive osteological study, including comparison with fossil forms, Lundberg (1970) has suggested that species in the subgenus Ictalurus retain more plesiomorphic character states than any other lineage within the Ictaluridae. Of the 2 species in that subgenus that have been k a ry o ty p e d punctatus and furcatus both have 2N=58. Additionally, Pylodictis (2N= 56) retains several primitive osteological features according to Lundberg (1970, p. 129). Lundberg (1970 p. 60) has marshalled osteological evidence on synapomorphic character states that indicates ic ta lu rid s represent 97 a monophyletic assemblage of species (see above). If correct, then the assumption that all ictalurid karyotypes havebeen derived from some common ancestral 2N should be valid. Thus, if the direction of karyotypic change has been toward both increase and decrease of the 2N about this ancestral state, then the distribution of numbers in extant species should represent a "serially homologous sequence" of 2N's in which the intermediate values should be near the ancestral value. Of the ictalurids that have thus far been karyotyped, 2N's range between 40 and 72, with most values in between represented in one or more species. Intermediate between these extremes would be 2N=56. As far as FN's are concerned, similar observations on the dis tribution of states in the above mentioned families indicate that high FN's (>80) are most cornnon in siluriforms that have been studied. In ictalurids, FTTs>76 are encountered only in the more specialized genus Noturus. Numbers o f 80-100+ are most common in Ictalurus and P ylodictis and are seen in some Noturus (Tables 1 and 3). The FN's known for non-ictalurid siluriforms (including the bagrids, pimelodids and ariids) are all >80. Considering the evidence outlined above, it seems logical to postulate that the primitive ictalurid karyotype consisted of 56+2 chromosomes and that deviations from this number represent derived states; those numbers nearer 54-58 are probably closer to the ancestral condition. In view of Lundberg's observations that the subgenus Ictalurus has probably retained more primitive features than other living Ictalurids, and the fact that 2N=58 is seen in this subgenus 98 as well as in closely related bagrids and pimelodids, ancestral ictalurids probably possessed a 2N of 58. If ictalurids are near the basal stock of the Siluriform es as suggested by Gosline (1975b) then 2N=58 may have been the ancestral condition for all catfishes. Likewise, the plesiomorphic condition for FN's is assumed to have been characterized by the presence of many large second arms on the chromosomes of the p rim itiv e 2N=58 karyotype. Fundamental numbers of greater than 80 are considered primitive. Thus, the 2N=58 karyo type with high FN's as seen in L punctatus and furcatus probably approaches the p rim itiv e state of the ic ta lu rid (silu rifo rm ? ) karyo type. A study of Diplomystes would be of particular interest in regard to th is problem. Diplomystes belongs to the monotypic South American catfish family Diplomystidae. Based primarily on its retention of m axillary teeth (Greenwood, Rosen, Weitzman and Meyers 1966) as well as its p rim itiv e caudal skeleton and Weberian structure (Lundberg and Baskin 1969), Diplomystes is thought to be the most primitive extant catfish species. Its karyotype would be of great interest in assess ing the ancestral state of catfish karyotypes and is currently under investigation by Dr. Hugo Campos (personal corrmunication 1976). I f Diplomystes has retained a relatively primitive karyotype as well as morphology, we might expect a 2N of around 58 with a relatively high FN (>80). EVOLUTION OF ICTALURID KARYOTYPES As discussed ea rlie r, the mechanism that has most often been invoked to explain changes of 2N in fishes is Robertsonian rearrange ment (i.e., centric fusions and/or dissociations). The most important feature of Robertsonian change is that alteration of the 2N is accomplished without a concomitant change in FN. Therefore, i f the changes in 2N within the family Ictaluridae could be explained solely as Robertsonian changes, then the FN's should have remained constant. The distribution of FN's (Figure 31) indicates this is not the case, but that diversity of FN's as well as 2N's is prevalent. A plot of FN vs. 2N (Figure 32) shows that FN, at least in some cases, has changed with 2N. Although Robertsonian rearrangements may, and probably have, occurred, additional rearrangements (such as unequal translocations, deletions, duplications or pericentric inversions) that affected centromeric position but not 2N have been superimposed on any such Robertsonian changes. Such a situation is consistent with the concept that these changes in karyotype between icta lu rid species may (at one time) have been important postmating isolating mechanisms. Because of the complexity of the icta lu rid karyotype and the diversity of both number and form of the chromosomes, i t is impossible to speculate at this time on what specific rearrangements 99 SPECIES 10 9 7 8 iue 1 Ditiuin f udmna nmes n cal i s rid lu Icta In numbers fundamental of istribution D 31. Figure UDMNA NUMBER FUNDAMENTAL tumtu; 100 101 100 y=32. 22 + .85x r2=.66 90 UJ 5 80 Q 70 u. 60 50 40 50 60 70 80 DIPLOID NUMBER Figure 32. Relationship between fundamental and diploid numbers in ictalurids (polymorphic N_. alba ter omitted). Upper and lower dashed lines represent totally saturated and completely unsaturated karyotypes respectively. 102 have taken place in the evolution of each species' karyotype. Perhaps with the development o f more advanced methods o f d iffe re n tia l chromosome band staining now being extensively employed in mammalian cytogenetic studies and their successful adaptation to fish material in the future, specific patterns of rearrangement might be reconstruct ed. However, even with advanced techniques of d iffe re n tia l band s ta in ing, speculation on the evolution of these karyotypes w ill s till be d iffic u lt due to the large number of changes that have apparently taken place. As already pointed out, I_. furcatus and punctatus have probably retained 2N's close to the ancestral condition. This is also true of P ylodictis which has decreased in number from the presumed ancestral condition of 58 to 2N=56. Besides the difference in 2N's, the karyo type of £. olivaris differs from that of punctatus and furcatus in the possession of relatively fewer msm's (i.e ., has a lower FN). Several rearrangements affecting centromeric position have apparently taken place in the divergence of Pylodictis and £. (Ictalurus) spp.; the decrease of the 2N from an ancestral 2N=58 could be explained easily as a Robertsonian fusion which produced the large, character istic m pair in Pylodictis from two pairs of stt's. Within Noturus, the third major phyletic lineage in the Icta lu rida e, the presumed ancestral number is most closely approached by three species in the subgenus Schilbeodes: insignis, exl1 is and g ilb e rti, all 2N=54, but distinguishable on the basis of FN. The distribution of 2N's in the taxa studied (Figure 33) indicates that parallel changes from the presumed plesiomorphic condition have 103 occurred in several lines within the family. This is most apparent in the subgenera Amiurus, Schilbeodes and Rabida. Within Amiurus, diploid numbers have both increased and decreased during th e ir evolution from the ancestral state. In Noturus, the major trend has been toward decrease in 2N, with the notable exception of (Rabida) albater, where the 2N has drastically increased to 66-72, higher than any other ic ta lu rid . A parallel decrease has occurred from 52 to 40 in Rabida and 54 to 42 in Schilbeodes. The most derived ictalurid karyotype (with the possible exception of the polymorphic albater) is found in a small madtom endemic to the Caddo River Drainage in the Ouachita region o f southwestern Arkansas, N^. (Rabida) ta y lo r i. Its low 2N (40) and the presence of many large msm's (probably fusion products) and low FN (62) attest to its apomorphous condition. Additionally, it is the only ictalurid (or siluriform) in which I have been able to distinguish probable sex chromosomes. The mechanism appears to be an XX (female) - XY (male) system, the male possessing a pair of heteromorphic chromosomes. An interesting morphological correlate of taylori1s derived state is the reduction o f the body la te ra l lin e canal system, which, in the material that I have seen during my chromosome study, rarely passes posterior to the anterior margin of the adipose fin. In cursory examination of other Noturus spp. available to me, this extreme reduction appears to be shared with only two other madtoms, N^. gyrinus and lachneri, both in the subgenus Schilbeodes and both possessing the lowest 2N's (42) 1n that subgenus. Since these two species apparently lie on a separate lineage (subgenus) w ithin Noturus from 104 ICTALURIDAE 2N-40-72 ICTALURUS 2N: 58 G. ICTALURUS AMIURUS 2N:48-62 G. PYLODICTIS 2N: 56 /'" noturus 2N:48 G. NOTURUS ■< SCHILBEODES 2N:42-54 \JJABIDA 2N:40-52 (66-72) Figure 33. Distribution of diploid numbers in the Ictaluridae by taxa studied. Numbers in parentheses indicate 2N range in the polymorphic albater. 105 taylori, karyotypic and morphological sim ilarities have apparently evolved independently in both subgenera. Thus, at least in th is one character, a derived morphological state supports my contention of lowered 2N‘s as indicative of a derived karyotype in ictalurids. RELATIONSHIPS IN ICTALURIDS Intergeneric Relationships Because of the diversity of karyotypes within the family, as well as the apparent parallel and/or convergent nature of certain chromosomal changes w ithin and between taxa studied, karyotypes have so far added little to our concept of intergeneric relationships in ictalurids. Perhaps with the study of remaining species of the sub genus Ictalurus more realistic interpretation of intergeneric rela tionships based on chromosomes w ill be possible. Figure 34 contains the basic phylogenetic relationships postulated between the major lineages o f ic ta lu rid s (adapted from Taylor 1969 and Lundberg 1970, 1975). The numbers in parentheses in that figure indicate the range of 2N's found w ithin each major group. Pylodictis and Noturus probably share a more recent common ancestry with each other than with any other ictua lu rids (Taylor 1969, Lundberg 1970). Lundberg (1970) presents evidence that Amiurus may have shared a more recent common ancestry with the P ylo dictis- Noturus lin e than with Ictalurus (s^.s^.), while Taylor (1969) places Amiurus closer to Ictalurus (s^.s^.). Because Ictalurus (s ^ .) species studied karyotypically have retained the piesiomorphic condition, there is no sound karyotypic evidence 106 (66-72 (58) (56) (48-50) (42-54) (40-54 I.(Ictalurus) Pylodictis N.(Hoturus) N. ( Schilbeodes) N.(Rabida) Figure 34. Intergeneric relationships in the Ictaluridae. Rased on Lundberg (1970*1*175) and Taylor (1969). 107 108 either to confirm or refute the relationship between Amiurus and the Pylodictis-Noturus line. If further study shows clearly that Amiurus is closer to the Pylodictis-Noturus line, and that those two together form a sister group of Ictalurus (,s..s. ), then resurrection of the genus Amiurus would be indicated. Relationships Within The Subgenera Noturus And Schilbeodes While the chromosome data on Ictalurus do not provide clear evidence on intrageneric relationships, karyotypes have, in some cases, provided new insight into the relationships and evolutionary trends within the genus Noturus. The data are consistent with some of the relationships postulated by Taylor (1969), while contradicting his phylogenetic scheme in other points. Taylor's proposed relationships are shown in Figure 24, while my concept (based prim arily on chromo some data) of relationships between the species studied are depicted in Figure 35. Both Taylor and Lundberg agree that the subgenus Noturus appears morphologically most similar to Schilbeodes. Its exact position, though, remains problematic. Noturus flavus has certa inly retained some primitive character states as pointed out by both Taylor and Lundberg. Among such plesiomorphic states that might be noted are its large body size, unreduced body lateral lin e canal system and retention of the pectoral radials in an unfused state. Noturus flavus shares a common 2N with hL (Schi 1 beodes) nocturnus. Although there are distinct differences in the karyotypes sufficient to separate the two, the karyotypes retain some sim ilarities. Those Schi1beodes 109 Figure 35. Proposed hypothesis of phylogenetic relationships between the 20 species of Noturus studied. Based primarily on chromosomal data. Number's 1n parentheses are diploid and fundamental numbers respectively. Presumed synapomorphles. Indicated by dark, numbered bars, are as follows: (1) adnate adipose fin; relatively short pectoral spines; subcutaneous eye; lengthened,;generally toothed, sublateral process of the premaxllla with no lamina between the process and premaxllla; 9 additional characters Indicated by Lundberg (1970) as listed on pp. 68-69; (2) 2N<54; body without distinct mottled pattern or darkly pigmented saddles and bands; pec toral spines nearly straight, posterior serrae, when present, poorly developed and not recurved toward spine base; anterior dentations, when developed, are Irregularly spaced; (3) FN< 76; (4) 2N<54; (5) 2N always <50; large distinct st element present In the s tt series; (6) 2 N < 4 8 ; (7) 2 N < 4 6 ; (8) lateral line canal system on body does not pass posterior to the anterior margin of the adipose fin; pectoral mode typically 7 or 8; (9) anal ray count elevated to 17-21; dark marginal bands In the median fins; (10) typically 20-24 anal rays; (11) pectoral spines*scimitar shaped; posterior serrae well developed and consistently recurved toward spine base; anterior serrae present and usually regularly spaced; 2N<54; (12) 2N<52; (13) 2N<48; (14) 2N <46; FN<68; (15) 2N <44; (16) no large, distinct metacentrlc pair that 1s separable from the remaining rasm's; FN-62; (17) 2N*40; lateral line canal system on body does not pass posterior to the anterior margin o f the adipose fin ; sex chromosomes identifiable; (18) elongate body, pectoral spines relatively shorter than other Rabida; relatively short head; reduced anterior serrae on pectoral spine; short humeral process of the postclelthrum; (19) four largest msm pairs distinctly separable from the remainder of that series. Note: Noturus albater 1s tentatively presented as having diverged early from the-eleqans-hlldebrandi lin e although chromosomally I t should be placed as a sister group to a ll other Rabida since I t 1s the only member of that subgenus to have Increased 2N above the presumed ancestral condition and retains a pleslomorphlc FN. lachnerl ( ohaeus (42,6E) funebrls (*4,66) leptacanthus (46,74) noeturnus (48.72) flavus (48-50,70) 1ns1gn1s (54,74) ro g1)bert1 (54.82) mo tO£ 3 3 ,§ U> i n flaviplnnls (52,82) miurus (50,74) IIO llbater (66-72,82) M 1 debrand 1 (46,82) elegans (46,80) flavater (44,62) eleutherus (42,66) stiamosus (42,62) OLl taylori (40.62) I l l which possess a 2N higher than flavus (i.e ., more primitive karyotypes) are also characterized by larger body sizes (although they do not reach the size extremes o f fla vu s) . A d d itio n a lly , my observations indicate that N_. nocturnus also retains an unreduced la te ra l lin e canal system as seen in fla v u s . Considering that N_. stigmosus, a rather derived Rabida in its karyotype, sometimes retains unfused pectoral radials (Taylor 1969, Lundberg 1970), th is condition in flavus could be secondarily derived. Also, a population of N_. flavus with a distinctive karyotype of 2N=50 in the headwaters of the Clinch River represents an intermediate condition between those SchiIbeodes with 2N=54 and 2N=48-42. The status o f th is "chromosomal race" is s t i l l in d e fin ite . Only two specimens from the Clinch headwaters were obtained and karyotyped; it is not known whether these represent a distinct species, distinct subspecies or simply a chromosomal poly morphism. The Copper Creek m aterial has not yet been studied s u f f i ciently to ascertain if morphological differences commensurate with the karyotypic differences are present. The origin of the 2N=50 race is also open to debate. Two main possibilities could explain the existence of this distinctive form. First, the 2N=50 karyotype may have been secondarily derived from the 2N=48 karyotype by dissociation of the large metacentric element found in northern Ohio River and Lake Erie Basin specimens. A more in teresting p o s s ib ility is that the Clinch River specimens represent a re lict population of ancestral flavus stock. Robertsonian fusion of two pairs of stt's may have resulted in the large metacentric pair and 2N of 48, a rearrangement that subsequently became fixed over most o f the range of flavus. I f 112 the latter case 1s true, it may indicate an origin of flavus possibly in the headwater areas of the Clinch River. Continued study of the geographical variation of N. flavus karyotypes and morphology w ill certainly be necessary before the status of this form and flavus as a whole may be c la rifie d . Observations such as those above make the retention of N_. flavus in a monotypic subgenus questionable. Although additional study of this problem is needed, Schilbeodes and Noturus are tentatively consolidated here, as indicated in Figure 35. Since Noturus is the oldest of those two names (Taylor 1969), i t is retained as the sub generic designation. Prior to Taylor's study, N_. exilis and insignis were considered close relatives on the basis of th e ir elongate body form, possession of a long (usually 17-19 rays in insignis, 19-21 in exi1is) anal fin and presence of dark marginal bands in the median fins. Taylor, how ever, felt that their striking morphological similarities were con vergent and that exi1 is and insignis were not closely related. Karyo- ty p ic a lly they d iffe r l i t t l e , with only a minor change in FN. Noturus e x i I ts certainly has no marked resemblance to 1achneri or gyrinus in its karyotype; both lachneri and gyrinus are species with which Taylor had aligned exi1 is. It is apparent that exi1is and insignis have been separate fo r quite some time since they have diverged s ig n ific a n tly in head morphology, p a rticu la rly in e x ilis 1 possession of a terminal mouth, more depressed skull and a single internasal pore. While Taylor viewed N_. gilberti as highly specialized, ithas retained a relatively primitive karyotype (2N=54) along with exilis 113 and insignis. Taylor’s indication that gilberti was probably the "most highly specialized member" of its subgenus is inconsistent with his indication of an early divergence for gilberti in his phylo- gram. Noturus g ilb e rti and the e x ilis - ins ignis assemblage have diverged significantly as evidenced by the much smaller number of msm's in both exi1is and insignis. Although Taylor pointed out that gilberti resembled flavus in general body form, possession of cream- colored margins on the truncate caudal fin and in th e ir re la tiv e ly short dorsal spines, there is little karyotypic similarity to suggest a closer relationship between those two species than with other Noturus. Noturus gilberti possibly represents one of the most primitive members of the subgenus Noturus (=Schi1beodes) , having diverged quite early from the remainder of that subgenus. Chromosomally, then, e x i1is , insignis and q i1be rti, with their 2N=54 karyotypes, are all probably close to the basal stock which gave rise to N_. (Noturus) . Beginning with N_. nocturnus, there is a smooth gradation of 2N from 48 to 42 in " Schi1beodes" . Noturus leptacanthus (2N=46) was thought by Taylor to represent one of the most primitive "SchiIbeodes” . It possesses karyotypic sim ilarities to other "SchiIbeodes" and its karyotype could have been derived easily from a nocturnus-1ike karyo type by Robertsonian fusion. Noturus nocturnus, leptacanthus, funebris, phaeus, gyrinus and lachneri all possess karyotypes that differ mainly in 2N, with only slight (possibly artifactual) variation in FN. Ictalurids (including these "Schi1beodes") with smaller 2N's 58) generally have more large msm's (LM). There is a s ig n ific a n t negative correlation (P<.01, r^=72.3) between the number of LM elements and 114 the 2N for all ictalurids studied {Figure 36). This suggests that Robertsonian fusion was probably an important mechanism in the decrease of diploid numbers within the family. The largest stt pair is distinctly separable from the remaining stt's in nocturnus, leptacanthus( funebris, phaeus, gyrinus and lachneri and may represent a synapomorphy fo r those species. Noturus funebris and phaeus are morphologically close species, resembling each other in body form, meristics and possession of an extremely long anal fin (Taylor 1969). Taylor placed them in a closely related species group (the funebris group) on the basis of these sim ilarities. Their karyotypes were found to differ in 2N, phaeus 2N=42 and funebris 2N=44. Since the data suggest a trend toward decrease in 2N in Noturus karyotypes, th is would indicate that funebris is probably the more primitive, representing more closely the stock from which the two originated. The karyotypes, although differing in 2N, are quite similar; phaeus could have been derived from a funebris-like karyotype by Robertsonian fusion, adding one more large msm pair. In spite of the slight difference in 2N, the s trik in g morphological, m eristic and karyotypic resemblances indicate an intim ate relatio nsh ip between phaeus and fun eb ris. Noturus gyrinus and lachneri share a common 2N (42) with phaeus, but the karyotypes are d is tin c tly d iffe re n t. The strongly submeta- centric element seen in both lachneri and gyrinus (about the third largest msm pair) is not present in phaeus' karyotype. There appear to be more msm pairs in the karyotypes of gyrinus and lachneri. I am not confidently able to separate gyrinus and 1 achneri on the basis 2 10 8 595x 6 4 2 r\ 4 0 60 70 8050 DIPLOID NUMBER Figure 36. Regression of the number of large msm's (LM) in ictalurid karyotypes on 2N. (Noturus albater om itted). 116 of th e ir karyotypes. This agrees with prominent morphological s im ila ritie s between the two (Taylor 1969). Both possess a terminal mouth and shortened la te ra l lin e canal system. I agree with Taylor's suggestion that gyrinus and lachneri are intimately related. Relationships In Rabida Noturus flavipinnis was thought to be extinct (Taylor 1969), but has subsequently been rediscovered by Taylor, Jenkins and Lachner (1970) in a Clinch River headwater trib u ta ry (Copper Creek) in south western Virginia. It retains a karyotype that is probably closer to the ancestral condition than any other Rabida. The high 2N and FN are indicative of plesiomorphic karyotypes in ictalurids. Taylor placed flavipinnis in a closely related species group with miurus and f 1avater (the miurus group). Although miurus (2N=50) is close to flavipinnis (2N=52) in 2N, it has diverged from it in FN and in the possession of a very characteristic large m pair not present in flavipinnis. Noturus flavater differs dramatically from flavipinnis and miurus in its karyotype. Noturus fla va te r has a karyotype (2N=44, FNi=62) that seems to more closely resemble that found in species of the furiosus species group, exemplified by stigmosus and munitus in this study. Both stigmosus and munitus, with 2N's of 42 and FN's of 62 are closely approached by the karyotype of flavater. Cursory examination of pectoral spine morphology in specimens of miurus, flavipinnis, flavater, munitus and stigmosus available to me in this study suggests greater sim ilarity of the spines of flavater with those 117 of stigmosus and muni tus than with miurus and flavipinnis. Spines of fla v a te r, munitus and stigmosus possess deep, lo n g itu d in a l, pa ra lle l grooves not seen in the spines of miurus and flavipinnis. Meristic data in Taylor's (1969) monograph indicate little difference between N_. flavater and furiosus. The major character states used by Taylor to align flavater with flavipinnis and miurus were sim ilarity of color pattern and possession of an elevated number of caudal fin rays, both highly variable characters within Noturus. I believe that N_. flavater is not closely related to either miurus or flavipinnis, but forms a karyotypically plesiomorphic sister group to the eleutherus-furiosus group-ta y lo ri assemblage w ith in Rabida, as evidenced by its low 2N (44) and FN (62) shared with that group o f species. Noturus eleutherus, thought by Taylor to be an intimate relative of no other Rabida, also has some karyotypic affinities to the furiosus group in its low 2N (42) and FN (66). The presence o f a d is tin c tiv e large metacentric pair of chromosomes in the karyotype of eleutherus does, however, distinguish i t from the furiosus group. This large pair superficially resembles the large m pair also characteristic ° f N. miurus. However, the two are thought not to be homologous since that would necessitate an independent reduction of 2N and FN from miurus to eleutherus ; this pair is not present in Rabida with 2N=46-44 or in other species with 2N=42. This large m pair indicates some karyotypic divergence from the furiosus group which parallels the morphological divergences pointed out by Taylor (1969). Noturus eleutherus seems to have a slightly higher FN, but this may be artifactual. Interestingly, though, eleutherus and stigmosus (a 118 member of the furiosus group) s u ffic ie n tly resemble each other super fic ia lly that they were, for many years, considered to represent the male and female o f the same species (Trautman 1957). Chromosomal evidence provides some support for Taylor's sug gestion of relationship between N_. elegans and N^. hildebrandi. Both share a common 2N (46) and have nearly the same FN (80 in hlldebrandi and 82 in elegans). The difference in FN could certainly be arti- factua l, due to the d iffic u lty of separating sm and st elements in ictalurid karyotypes. If real, however, it could easily be explained as a pericentric inversion difference. The karyotypes are similar, although differences sufficient to separate the two are present. The distinctive sm pair (about the second largest msm pair) in hi!debrandi is probably comparable to the d is tin c t large st pair in the karyotype of elegans; the difference is explainable as a pericentric inversion. Removing that pair from consideration, the msm series (particularly the larger pairs) differ little between the two species. Such chromo somal sim ilarities in conjunction with shared morphological charac te ris tic s (elongate body, short heads, reduced anterior serrae on the pectoral spines, short humeral processes of the postcleithrum) pointed out by Taylor (1969) indicate closer relationship between the elegans and hildebrandi groups than between either of them and other Rabida (with the possible exception of albater). The increased 2N (higher than any other ic ta lu rid ) suggests that centric dissociation has probably played a major role in the evolution of a lb a te r1s karyotype. The specimens with modal counts of 67 and 71 are not considered to represent true aneuploids as might 119 be expected of prime number chromosome counts. I suggest that those are specimens heterozygous fo r one or more centric dissociations. Since an individual heterozygous fo r a single dissociation would have two s tt's , each homologous to d iffe re n t arms of an msm, the genes of every linkage group would s till be represented the same number of times in the karyotype. Thus, i f an individual was heterozygous fo r 1, 3, 5 ,____ etc. centric dissociations (or fusions, for that matter) it could have a 2N indicative of aneuploidy while s till being geneti ca lly euploid. An analogous case would be the Mexican cyprinodont fishes (see above) with m ultiple sex chromosomes (Uyneo and M ille r 1971, 1972). Although the male is characterized by a prime number of chromosomes (47) i t would not be considered aneuploid since each arm of the single biarmed Y is homologous to different acrocentric X's. In spite of the polymorphic karyotype found in N_. a lb a te r, i t retains a FN (82) close to that of both hildebrandi and elegans. Thi's is consistent with Taylor's observations of morphological a ffin itie s between a lb a te r, elegans and hildebrandi. Although albater is probably most closely related to the hildebrandi and elegans groups its exact position is s till unclear. Possibly, the divergence of albater1s karyotype is relatively recent and, in fact, s till in progress as indicated by a failure to undergo fixation of a particular pattern of rearrangement. Since only one population of th is species was sampled, I do not know i f th is polymorphism occurs throughout the range of albater. Additional study of karytoypic variation over alba te r1s entire ^ange w ill be necessary before fin a l judgement may be passed on its phylogenetic position or the exact nature of its 120 polymorphic karyotype. The retention of a high FN cannot alone be taken as indicative of relationship to hildebrandi or elegans since this represents a plesiomorphic condition. In fact, on the basis of chromosomal change only, albater should be considered a s is te r group to all other Rabida. However, its retention of morphological characteristics of both the hildebrandi and elegans groups (e.g. , short humeral process, reduced anterior pectoral spine serrae, elongate body and short head) may suggest an early divergence from the ancestral stock which gave rise to those two species groups. Because of this, I have tentatively placed albater near the hildebrandi- elegans line in Figure 35 in spite of its poor karyotypic fit in that scheme. Although having diverged significantly in the development of a sex correlated chromosome heteromorphism, the karyotype of N_. taylori resembles that of stigmosus and muni tus o f the furiosus group in its low 2N (40) and FN (62). There is general s im ila rity o f chromosome form (especially in the larger msm's) of taylori, stigmosus and munitus which suggests close relationship to (if not derivation from) the furiosus group. Noturus furiosus and placidus karyotypes would greatly aid our understanding of the exact relationships of some of the Rabida apparently peripheral to the furiosus group (including taylori, eleutherus and flavater). 121 Diploid Vs. Fundamental Numbers I have placed more emphasis on changes in 2N's than FIM1 s as indicators of relationships in Noturus for two reasons. First, while it is relatively easy to ascertain the probable correct diploid number for a specimen (or species), it is much more d ifficu lt to accurately assess the proper FN in ictalurids and other siluriforms. The problem with establishing FN's is due to the large number of elements which possess small to large second arms in the karyotypes of siluriforms. This, in conjunction with the large number of chromosomes on the borderline between sm and s t categories makes assessment of FN's problematic. Small differences in FN (+2-4) may well represent artifact because of differential chromosome contraction or sampling error re su ltin g from the small number o f ce lls fo r which the FN's were calculated for each species. Secondly, it seems much more realistic to theorize secondary elevations or reductions of FN than 2N. A change in FN can be effected as the result of a simple pericentric inversion in a single chromosome; that rearrangement could then become fixed in the population with sub sequent generations. A change in the 2N of the same magnitude would involve the more complex, thus less parsimonious process of centric fusion; this requires simultaneous chromosome breaks and fusion of appropriate fragments in two non-homologous chromosomes forming a single element from two independent elements. Also, if the trend of FN decrease with increasing specialization were taken as the sole c rite rio n fo r arrangement o f the madtoms in a phylogenetic scheme, 122 this would necessitate the positioning of some species such that decrease in 2N would be followed by an increase then subsequent decrease. For example, Noturus miurus (2N=50) has a FN o f 74 while N_. hiIdebrandi and elegans {both 2N=46) have FN's of 80-82. A phylogeny based on decreasing FN's would necessitate the divergence ° f miurus follow ing that of hildebrandi and elegans in Rabida since i t has a FN intermediate between those two species and other Rabida with 2N's of less than 46 ( fla v a te r, eleutherus, stigmosus, munitus and ta y lo ri). This would imply that a fte r in it ia l decrease in 2N from an ancestral of about 52 (as seen in flavipinnis) to 46 (as in hiIdebrandi and elegans) some form of centric dissociation would have to be invoked to explain the higher 2N in miurus. A decrease in 2N would again be needed to explain the lower 2N's in the remaining Rabida. Since centric dissociations do not seem to be as parsimonious an evolutionary event as centric fusions in fishes, this would probably be less likely than independent reduction of the FN by pericentric inversions in miurus. Thus, I believe that the transformation of 2N's more nearly reflects the course of evolutionary change and divergence in madtom karyotypes; inconsistencies between 2N change and the transition of FN's are explainable as e ith e r a rtifa c t of technique or secondary pericentric inversions that increased or decreased FN's independently on several occasions. 123 Noturus Relationships: Morphological Evidence Pelvic Ray Counts--Ictalurids typically possess 8-10 pelvic rays. Lundberg (1970) and Taylor (1969) note that most silu rifo rm s possess modally 6 pelvic rays, a condition thought to be primitive for the order. This suggests a trend toward increase in pelvic ray counts within ictalurids. Most of the Noturus in this study have a mode of 9 pelvic rays. In Rabida, N_. hildebrandi and baileyi usually have 8. In Noturus (Noturus) (_s.J_.), a mode of 8 is seen in leptacanthus, gyrinus and lachneri. If the lower modes in hildebrandi, leptacanthus, lachneri and gyrinus are taken as indicative of a more primitive condition, a phylogeny based on changes in pelvic ray formu lae would indicate a more basal position fo r those species than one generated from chromosomal data. However, i f a modal count of 9 is taken as primitive for Noturus, then the chromosomal phylogeny would be inconsistent only in the placement of phaeus and funebris (both with modes of 9) between leptacanthus and g.yrinus- lachneri (with typically 8). The position of hiIdebrandi (and baileyi?) within Rabida would not be inconsistent with chromosomal data, but simply reflect a parallel decrease of the pelvic count as seen in N^ (Noturus) and indicate a synapomorphy for those two species. It is interesting to note here that Pylodictis usually has 9 (often 10) pelvics as seen in most Noturus species. Since Pylodictis and Noturus are thought to have shared most recent common ancestry with each other, parsimony would suggest a primitive mode of 9 pelvics in the Pylodictis-Noturus lineage. This would necessitate parallel reduction of the pelvic 124 counts in Rabida and N^. (Noturus), where a primitive mode of eight for that line would indicate parallel increase in Rabida, N. (Noturus) and Pylodictis. Pectoral Ray Counts--Lundberg (1970) notes that most catfishes typically possess 9 soft pectoral rays, a state thought to be plesiomorphic for siluriforms (including the ictalurids). Pectoral modes of less than or greater than 9 rays are thought to represent apomorphous states. All members of N_. (Noturus) possess modes of 9 except flavus (10, often 9), gyrinus (7, often 8) and lachneri (8) according to Taylor (1969). Rabida usually have 8 pectoral rays, except in hildebrandi (9, often 8) and albater (9). If a mode of 9 is taken as p rim itive fo r the genus, then the position of hildebrandi and albater in Figure 35 would imply secondary increase to a plesiomorphic mode of 9 from an apomorphous condition of 8 since both flavipinnis and miurus (olaced nearer the basal stock of the subgenus) have modes of 8. Changes in pectoral ray counts within N^. (Noturus) from a plesiomorphic state of 9 is consistent with the phylogeny deduced from chromosomes. The increase to a mode of 10 in f l avus would represent an autapomorphy fo r that species, while reduction to modes of 7 or 8 in gyrinus and lachneri could represent a synapomorphy for those two species. Precaudal Vertebrae--Lundberg's (1970) pcv counts do not agree with those of Taylor (1969) for the same species, therefore Lundberg's conclusions concerning trends of change in pcv within ictalurids can not be conclusively applied to Noturus data in Taylor. Also, counts 125 are not available for fL flavipinnis, taylori and bai leyi, while only small samples of several other species were counted by Taylor. If high pcv modes of 9-11 (Taylor's counts) are taken as primitive, then reduction to a mode of 8 in N. (Noturus) nocturnus, leptacanthus, funebris, phaeus, gyrinus and lachneri could represent a synapomorphy for that group of species; this is consistent with the unity o f this group o f species deduced from chromosomal data (Figure 35). Within Rabida, an assumption of the same p rim itive condition would necessitate independent reductions to modes of 7-8 pcv in eleutherus and the hiIdebrandi- elegans lin e . Such a trend of decrease is inconsistent with Lundberg's conclusion that pcv have tended to increase within ictalurids, but would seem possible in light of the elevated pcv's seen in Pylodictis, if Noturus and Pylodictis do indeed share most recent common ancestry among ic ta lu rid s . Caudal Vertebrae--As with precaudal vertebral counts, Lundberg's and Taylor's counts do not agree, therefore, Lundberg's primitive states cannot be u tiliz e d fo r comparison with Noturus. Also, no counts are available on for flavipinnis, taylori or baileyi. However, if cv modes of 26-30 are taken as prim itive (25-28 in Lundberg, but Taylor's values tend to be about 1-2 greater) then a ll N_. (Noturus) appear to have retained a plesiomorphic condition. All Rabida for which data are available have modes of 26-29 except eleutherus, stigmosus and munitus, which possess lower numbers (modes of 22-25). This agrees with the apomorphous position of those three species in a chromosomally based phylogeny. 126 Chromosomes Vs. Morphology I believe that a phylogeny generated primarily from trends of change in the karyotypes of madtoms is, at this time, probably a more realistic interpretation of their relationships than one based on meristic or morphological differences. I base this reliability on the following reasons: 1. M eristics (as seen in Taylor 1969) show much greater in tra sp e cific variation than chromosomes in madtoms. For many of Taylor's meristics, although modal values may differ slightly, ranges for most species overlap quite a bit so that distinct character states are hard to define. I t might be mentioned once again that Hudson (1976) suggests great intraspecific variation of FN's within ictalurids. As discussed above, I feel that th is is more apparent than re a l, but even so, I have emphasized 2N changes over FN changes. 2. Meristic values are subject to much greater environmental modification than are karyotypes. For example, i t is a well known phenomenon that the number of many m eristic elements is a function of the temperature at which development takes place. Individuals from cooler waters tend to have more rays, scales, vertebrae, etc. than those from warmer climates. Thus m eristic values may often reflect adaptation to specific environ ments rather than phylogenetic relationships. Such would not seem to be as lik e ly with karyo types. Meristic differences in obviously related species are at times greater than between more d is ta n tly related species in the same subgenus. The sim i larity in body form, identity of the karyotype, terminal mouth, extremely reduced body la te ra l line and allopatric but contiguous distributions point strongly to an intimate relationship between gyrinus and lachneri; an allopatric, sister species relationship is apparent between the two. In a number of meristic characters that might poten tia lly be used as indicators of relationship, they d iffe r from each other more than from less related species in Noturus (Noturus) . For example: lachneri gyrinus pectoral mode 8 7 anal rays 16-18 14-16 branchiostegal rays 11 9 caudal vertebrae 30 26-27 Total Vertebrae 37-38 34-35 Morphological data for all species are not avail able for use as potential tests of the phylogenetic 128 hypothesis presented here. For example, data on branchiostegal rays, caudal and precaudal vertebrae are not given in the literature for flavipinnis or taylori. 5. Data available on meristic characteristies are not consistent in the lite ra tu re (see, fo r example, discussions o f precaudal and caudal vertebrae above). These observations suggest that comparative morphology of Noturus species is too poorly known to use most available data as potential falsifiers (in the sense of Wiley 1975) of the phylogenetic hypothesis presented in this study of Noturus. Consequently, the phylogeny herein suggested must be considered as a f i r s t e ffo rt in the generation o f a s o lid ly based, em pirically produced and testable phylogenetic hypothesis fo r Noturus. In th is lig h t, i t appears that the next lo gical step and most f r u it f u l area o f study would be an exhaustive analysis of Noturus osteology much like that carried out by Lundberg (1970) fo r other ic ta lu rid s . Since Lundberg's study con tained only three species of Noturus (one representative of each of the then recognized subgenera) i t shed l i t t l e new lig h t on intrageneric relationships of madtom species. Such a study of all Noturus, with sufficient data, should provide a number of independent tests which could subsequently be applied to the phylogenetic hypothesis presented here to ascertain whether or not it will be corroborated. Prelimi nary consideration of 4 selected morphological characters (pelvic rays, pectoral rays, precaudal vertebrae and caudal vertebrae) indicates 129 that evolutionary trends in morphological traits can be fitted into a chromosomally deduced phylogenetic hypothesis. SUMMARY The chromosome complements o f 26 species of North American catfishes were studied. With previously published information, this brings the total number of ictalurids for which karyotypes are known to 30. Diploid numbers in the fam ily ranged from 40 in N_. ta y lo ri to 66-72 in N_. a lb a te r. Noturus albater was found to be polymorphic in its diploid number, with specimens possessing modal 2N's of 66-72. All other Noturus had 2N's of 40-54, with every even 2N between 40-54 repre sented in one or more species. Pylodictis possessed a karyotype composed of 56 chromosomes. Species curre n tly placed in the genus Ictalurus had diploid numbers of 48 (in I_. catus) to 62 {in JL natalis and brunneus) . The members of the subgenus Amiurus had 2N's e ith e r greater than or less than 58, while both I_. ( Ic ta lu ru s ) punctatus and furcatus possess 2N's of 58. Based primarily on the distribution of chromosomal states in the various lineages of ictalurids as well as in peripherally related siluriform families {e.g., Bagrldae, Pimelodidae and Ariidae) an ancestral ictalurid karyotype of 2N=58 with a relatively high FN (>80) has been postulated. This karyotype may have been ancestral for all siluriforms. Ictalurids (and siluriforms in general?) have karyotypes 130 131 characteristically possessing many chromosomes with identifiable second arms (ranging from minute to large) on most or all of their elements. Robertsonian rearrangements, including primarily centric fusions but also centric dissociations, appear to have been important in the evolution of ictalurid karyotypes. Interspecific karyotype differences cannot, however, be to ta lly explained by Robertsonian rearrangement; additional rearrangements such as non-Robertsonian translocations and/or pericentric inversions must be invoked to explain the karyotypic differences. While it is possible that karyotypic differences in ictalurids are sufficiently great to serve as postmating isolating mechanisms by inducing ste rility or reduced fe rtility in F-j's, evidence suggests that they are not great enough to prevent the formation of F-j's. The paucity of Noturus hybrids is probably a function of strong premating (ethological or ecological?) isolating mechanisms. It is suggested that Robertsonian rearrangements alone are insufficient to induce hybrid s t e r ilit y between species; however, i f additional rearrangements become fixed in each species, former linkage groups would more lik e ly be disrupted to the point that synapsis and segregation difficulties would produce s t e r ilit y in F-j hybrids. There does not appear to be any greater tendency toward karyo typic dissimilarity in sympatric vs. allopatric species pairs of Noturus. It is possible that chromosomal differences developed in allopatric madtom stocks during their early divergence and served as postmating isolating mechanisms upon return to sympatry. This 132 would prevent Introgression while more efficient premating mechanisms developed in sympatry. Karyotypic data presently available provide little insight into intergeneric relationships or intrageneric relationships in Ictalurus. However, Noturus karyotypes do provide the basis for a new hypothesis on intrageneric phylogeny of the madtoms. Chromosomal sim ilarities to SchiIbeodes as well as the lack of sufficiently unique morphological characters suggests the need for consolidation of the previously recognized subgenera Noturus and SchiIbeodes into a single subgenus Noturus (Noturus) . There has been a parallel trend in the evolution of karyotypes within N. (Noturus) (.s.K) and N^. (Rabida). This is characterized by stepwise reductions of 2N from 54 to 42 in N_. (Noturus) and 52 to 40 in N_. (Rabida) . A chromosomal race of N_. (Noturus) flavus was found in the Clinch River headwater trib u ta ry Copper Creek. This form was characterized by a diploid number of 50, while flavus from the Lake Erie and northern Ohio River Basin had 2N=48. The difference is apparently explainable as a Robertsonian rearrangement since 2N=48 flavus have a large distinct metacentric pair not seen in 2N=50 flavus. The hypothetical common ancestor of the Noturus-Pylodictis lineage probably possessed a karyotype similar to that seen in Pylodictis with 2N=56 and a FN of 82+. Early stages of Noturus divergence may have involved in itia l reduction of the 2N to 54 with a high FN retained; such a karyotype is seen in Noturus (Noturus) g ilb e rti and is closely approached by Noturus fla v ip in n is in Rabida 133 (2N=52, FN=82). Within N_. (Noturus) the ancestral condition is most closely approached by insignis and e x i1 is in addition to g ilb e rti (a ll 2N=54). This group o f species probably most closely resembles the basal stock from which the remaining members of that subgenus diverged. Within Rabida, hL flavipinnis and miurus (2N=52 and 50 respec tively) appear to be closest to the ancestral karyotype state in that subgenus. Noturus flavater, formerly aligned with those two species appears to be more closely related to the furiosus species group since it has a low 2N (44) and FN (62) similar to those seen in stigmosus and munitus (2N=42, FN=62). The sharing o f a common 2N (46) and sim ila r karyotype in hildebrandi and elegans supports a close relationship between those two species. The position of N^ albater remains unclear. Noturus ta y lo ri appears to be the most derived species o f Rabida since i t possesses the lowest 2N (40), a low FN (62) and a sex correlated chromosomal heteromorphism with females XX and males XY. More emphasis is placed on transformation of 2N's than FN's in assessing relationships since difficulties are apparent in the ascertainment of FN's. This is due to the large number of chromosomes with small to large second arms in the ictalurid karyotype as well as many borderline cases between the sm and st categories. Because sufficiently detailed morphological studies of Noturus have not been carried out, data are not available for strenuous testing of a chromo- somally based hypothesis against evolutionary trends in morphology. LITERATURE CITED Aghasadeh, H. and H. R itte r. 1971. Polyploidization in the fish family Cyprlnidae. Duplication of the gene loci for NAD- dependent malate dehydrogenase. Humangenetik 11:91-94. Bogart, J. P. 1969. Chromosomal evidence for evolution in the genus Bufo. Unpublished Ph.D. Dissertation, University of Texas, TT6pp. Chen, T. R. 1971. A comparative chromosome study of twenty k illifis h species of the genus Fundulus (Teleosteii: Cyprinodontidae). Chromosoma 32:456-^53, Chen, T. R. and F. H. Ruddle. 1970. A chromosome study o f four species and a hybrid of the k illifis h genus Fundulus (Cyprinodontidae). Chromosoma 29:255-267. C h ia re lli, B. and E. Capanna. 1973. Checklist of fish chromosomes. IN: Cytotaxonomy and Vertebrate Evolution, Chiarelli and Capanna (eds.), p p .'206-232 . Academic Press, New York. Davisson, M. T., L. E. Wright and L. M. Atherton. 1972. Centric fusion and trisomy for the LDH-B locus in brook trout, Salvelinus fo n tin a lis . Science 178:992-994. Denton, T. E. 1973. Fish Chromosome Methodology. C. C. Thomas, Springfield, Illinois, 166 pp. Dobzhansky, T. 1970. Genetics of the Evolutionary Process. Columbia Univeristy Press, New York, 505 pp. Douglas, N. H. 1972. Noturus tay1o r i, a new species of madtom (Pisces, Ictaluridae) from the Caddo River, southwest Arkansas. Copeia 1972:785-789. Gold, J. 1977 (In press). Fish cytogenetics. IN: Fish Physiology, Volume 8, Hoar and Randall (eds.), Academic Press, New Vork. Gold, J. and J. C. Avise. 1976. Spontaneous trip lo id y in the C alifornia roach Hesperoleucus syimetricus (Pisces:Cyprinidae). Cytogenet. Cell Genet. 17:144-149. 134 135 Gosline, W. A. 1975a. The palatine-maxillary mechanism in catfishes, with comments on the evolution and zoogeography o f modern s ilu ro id s . Occ. Pap. C a lif. Acad. Sci. No. 120, 31 pp. Gosline, W. A. 1975b. A reexamination of the sim ilarities between the freshwater fishes of Africa and South America. XV11e Congres International de Zoologie, Theme No.,1, Giogeographie et Liaisons Intercontinentales au cours du Mesozoique. Mem. Mus. Nat. H ist. N at., ser A, Zool. 88:146-154-. Greenwood, P. H ., 0. E. Rosen, S. H. Weitzman and G. S. Myers. 1966. Phyletic studies of teleostean fishes with a provisional class ific a tio n o f liv in g forms. American Museum of Natural H istory Bulletin 131:341-455. Hinegardner, R. T. and D. E. Rosen. 1972. C e llu la r DNA content and the evolution of teleostean fishes. American N a tu ra list 106:621-644. Hubbs, C. L ., R. R. M ille r and L. H. Hubbs. 1974. Hydrographic history and relict fishes of the north-central Great Basin. California Academy of Sciences Memoir 7:1-259. Hudson, R. G. 1976. A comparison of karyotypes and erythrocyte DNA quantities of several species of catfish (Siluriformes), with phylogenetic implications. Unpublished Ph.D. Dissertation, North Carolina State University, 209 pp. Jenkins, R. E. 1976. A lis t of undescribed freshwater fish species of continental United States and Canada, with additions to the 1970 checklist. Copeia 1976:642-644. John, B. and K. R. Lewis. 1969. The chromosome complement. Protoplasmatologia Via. Springer-Verlag, Vienna, 206 pp. Khuda-Bukhsh, A. R. and G. K. Manna. 1974. Somatic chromosomes in seven species of teleostean fishes. Chrom. Inf. Serv. No. 17:5-6. Kirpichnikov, V. S. 1973. On karyotype evolution in Cyclostomata and Pisces. Ichthyologia 5:55-77. Klose, H., U. Wolf, H. Hitzeroth and H. Ritter. 1969. Polyploidization in the fis h fam ily Cyprinidae, Order Cypriniformes. Humangenetik 7:245-250. Kobayasi, H. 1976. Comparative study of karyotypes in the small and large races of spinous loaches (Cobitis biwae). Zool. Mag. (Tokyo) 85:83-87. LeGrande, W. H. 1975. Karyology of six species of Louisiana fla t fishes (PIeuronectiformes :0steichthyes). Copeia 1975: 516-522. 136 LeGrande, W. H. and J. M. Fitzsimons. 1976. Karyology of the mullets Mugil curema and M. cephalus (Perciformes:Mugi1idae) from Louisiana. Copeia 1976:388-391. Levan, A., K. Fredga and A. Sandberg. 1964. Nomenclature fo r centromerlc position on chromosomes. Hereditas 52:201-220. Levin, C. B. 1973. Folio P-16 in F. L. Roberts (ed.), The Atlas of Chromosomes, Volume 2. Lundberg, J. G. 1970. The evolutionary h isto ry of North American catfishes, family Ictaluridae. Unpublished Ph.D. Dissertation, University of Michigan, 524 pp. Lundberg, J. G. 1975. The fossil catfishes of North America. Papers on Paleontology, University of Michigan Museum of Paleontology, No. 11, 51 pp. Lundberg, J. G. and J. N. Baskin. 1969. The caudal skeleton of the catfishes, order silu rifo rm e s. Novitates, American Museum of Natural History No. 2398, 49 pp. Manna, G. K. and R. Prasad. 1971. A new perspective in the mechanism o f the evolution o fchromosomes in fishes. Proc. First All India Congr. Cytol. and Genet., J. Cytol. and Genet. Congr. Suppl. :237-240. Manna, G. K. and R. Prasad. 1974. Cytological evidence fo r two forms o f Mystus v itta tu s (Bloch) as two species. The Nucleus 17:4-8. McPhail, J. D. and R. Jones. 1966. A simple technique for obtaining chromosomes from te le o st fishes. J. Fish. Res. Bd. Canada 23:767-768. Menzel, B. W. and E. C. Raney. 1973. Hybrid madtom c a tfis h , Noturus gyrinus X H. miurus, from Cayuga Lake, New York. American Mid1and N a tu ra list 90:165-176. M iller, R. R. 1966. Geographical distribution of Central American freshwater fishes. Copeia 1966:773-802. Muramoto, J. S., S. Ohno and N. B. Atkin. 1968. On the d ip lo id state of the fish order Ostariophysi. Chromosoma 24:59-66. Natarajan, R. and K. Subrahmanyam. 1974. A karyotype study o f some teleosts from Portonovo waters. Proc. Indian Acad. Sci., Sect. B, 79:173-196. Nayyar, R. P. 1966. Karyotype studies in thirteen species of fishes. Genetica 37:78-92. 137 Nikol'skiy, G. V. and V. P. Vasil'yev. 1973. Some features of the distribution of chromosome numbers in fish. J. Ichthyology 13:1-19. Nogusa, S. 1960. A comparative study of the chromosomes in fishes with particular considerations on taxonomy and evolution. Mem. Hyogo Univ. Agric. 3:1-62. Post, A. 1965. Vergleichende Untersuchungen der Chromosomenzahlen bei Subwasser-Teleosteern. Z. f. zool. Syst. u. Evol. 3:47-93. Prasad, R. 1971. On the controversy of somatic and germinal chromosomes in the fish Clarias batrachus. Proc. First All India Congr. Cytol. and Genet., J. Cytol. and Genet. Suppl. :208-212. R ishi, K. K. 1973. Somatic karyotypes o f three teleosts. Genen en Phaenen 16:101-107. Roberts, F. L. 1964. A chromosome study of twenty species of Centrarchidae. J. Morphology 115:401-418. Roberts, F. L. 1973. Folio P-17 in F. L. Roberts (ed.), The Atlas o f Chromosomes, Volume 2. Ross, M. R. 1973. A chromosome study o f fiv e species of etheostamine fishes (Percidae). Copeia 1973:163-165. Scheel, J. J ., V. Simonsen and A. 0. Gyldenholm. 1972. The karyotype and some electrophoretic patterns of fourteen species of the genus Corydoras. Z. f. zool. Syst. u. Evol. 10:144-152. Scott, W. B. and E. J. Crossman. 1973. Freshwater Fishes of Canada. Fisheries Res. Bd. Canada B u lle tin No. 184, 965 pp. Setzer, P. Y. 1970. An analysis of a natural hybrid swarm by means o f chromosome morphology. Trans. Amer. Fish. Soc. 99:139-146. Srivastava, M. D. L. and B. Das. 1969a. Somatic chromosomes of teleostean fish. J. Heredity 60:57-58. Srivastava, M. D. L. and B. Das. 1969b. Somatic chromosomes of Clarias batrachus (L.) (ClariidaeiTeleostomi). Caryologia 21 :349-352. Taylor, W. R. 1954. Records o f the fishes in the John N. Lowe collection from the Upper Peninsula of Michigan. Misc. Pub!., Mus. Zool. Univ. Mich. No. 87, 50 pp. 138 Taylor, W. R. 1969. A revision of the catfish genus Noturus Rafinesque with an analysis of higher groups in the Ictaluridae. B u lle tin , U. S. National Museum No. 282, 315 pp. Taylor, W. R., R. E. Jenkins and E. A. Lachner. 1973. Rediscovery and description of the ictalurid catfish, Noturus flavipinnis. Proc. Biol. Soc. Wash. 83:469-476. Thompson, K. W. 1976. Some aspects o f chromosomal evolution o f the Cichlidae (Teleostei:Perciformes) with emphasis on Neotropical forms. Unpublished Ph.D. Dissertation, University of Texas, 133 pp. Trautman, M. B. 1948. A natural hybrid catfish, Schilbeodes miurus x Schilbeodes m ollis. Copeia 1948:166-174. Trautman, M. B. 1957. The fishes of Ohio. Ohio State University Press, Columbus, Ohio, 683 pp. Uyeno, T. and R. R. M ille r, 1971. M ultiple sex chromosomes in a mexican cyprinodontid fish. Nature 231:452-453. Uyeno, T. and R. R. M ille r. 1972. Second discovery o f multiple sex chromosomes among fishes. Experientia 28:223. Uyeno, T. and G. R. Smith. 1972. Tetraploid origin of the karyotype of catostomid fishes. Science 175:644-646. White, M. J. D. 1973. Animal Cytology andEvolution. Cambridge Press, London, 961 pp. Wiley, E. 0. 1975. Karl R. Popper, systematics, and classification: A reply to Walter Bock and other evolutionary taxonomists. Syst. Zool. 24:233-243. Wilson, A. C. 1975. Evolutionary importance of gene regulation. Stadler Symposium 7:117-133. Wilson, A. C., G. L. Bush, S. M. Case and M. -C. King. 1975. Social structurin g of marmialian populations and rate o f chromosomal evolution. Proc. Nat. Acad. Sci. USA 72:5061-5065. Yerger, R. W. and K. Relyea. 1968. The flatheaded bullheads (Pisces:Ictaluridae) of the southeastern United States, and a new species of Ictalurus from the Gulf Coast. Copeia 1968: 361-384.