Rivuline Karyotypes and their Evolution (Rivulinae, Cyprinodontidae, Pisces)

180 /. /. Scheel

homologues aux dents de la .pla ue molairep des Parabathynellidae. Chez cette dernibre famille une mandibule dont la *p?aque molairew se compose d’une nombre klevke de dents arrangies sur un lobe nettement visible doit Etre regardie comme primitive. Au cours de I’Cvo- lution le nombre de dents se diminue et, simultaniment, le lobe est entiirement rkduit. A la fin, la <plaque molairew n’est plus reprksentie que par 3 dents minuscules. Dans cette lignCe kvolu- tive la mandibule des Bathynellidae occupe une position intermkdiaire et n’est point primitive.

Literatur

BOAS, J. E. V., 1883: Studien iiber die Verwandtschaftsbeziehungen der Malakostraken. Mor- phol. Jahrb. 8,485-579.

COINEAU, N., 1964: Recherches sur la faune des Iles MCditerranCennes. I. Le genre Para- bathynella (Crust. Syncar.) en Corse. Vie Milieu 15, 993-1005.

GORDON, I., 1964: O n the mandible of the Stygocaridae (Anaspidacea) and on some other Eumalacostraca, with special reference to the lacinia mobilis. Crustaceana 7, 150-157.

GREEN, J., 1964: Two new species of Parabathynella (Crustacea: Syncarida) from Lake Albert, Uganda. Proc. 2001. Soc. Lond. 142, 585-592.

HUSMANN, S., 1968: Ukologie, Systematik und Verbreitung zweier in Norddeutschland sympatrisch lebender Bathynella-Arten (Crustacea, Syncarida). Intern. Journ. Speleol. 3,

JAKOBI, H., 1954: Biologie, Entwicklungsgeschichte und Systematik von Bathynella natans

MANTON, S. M., 1928: On some points in the anatomy and habits of the lophogastrid Crustacea. Trans. Roy. SOC. Edinb. 56, 103-1 19.

NOODT, W., 1964: Natiirliches System und Biogeographie der Syncarida (Crustacea, Mala- costraca). Gewass. Abwass. 37-38, 77-186.

NOODT, W.; GALHANO, M. H., 1969: Studien an Crustacea Subterranea (Isopoda, Syncarida, Copepoda) aus dem Norden Portugals. Publ. Inst. Zool. ,Dr. A. Nobre“, 107,9-75.

SCHMINKE, H. K., 1972: Beitrage zur Evolution, zum System und zur Verbreitungsgeschichte der Parabathynellidae (Bathynellacea, Malacostraca). Verh. Acad. Wiss. Lit. Mainz, Math.- nat. KI. ( Im Druck).

111-145.

VEJD. 2001. Jb. (Syst.) 83, 1-62.

Anschriff des Verfassers: Dr. HORST KURT SCHMINKE, Zoologisches Institut der Universitat D-23 Kiel, Hegewischstr. 3

Danmarks Akvarium, Charlottenlund, Denmark

Rivuline Karyotypes and their Evolution (Rivulinae, Cyprinodontidae, Pisces)

By J. J. SCHEEL

Receipt of Ms. I. January 1972

Introduction

The present paper deals with 127 karyotypes disclosed in 75 different major phenotypes of rivuline fishes (Rivulinae, Cyprinodontidae, Pisces).

Haploid numbers for rivuline fishes were first published by POST (1965, p. 60-62) who studied 21 different phenotypes. The present author (1966 and 1968 a) published

2. zool. Syst. Evo1ut.-forsb. 10 (1972) 180-209 @ 1972 Verlag Paul Parey, Hamburg und Berlin

Rivuline Karyotypes and their Evolution 181

the haploid number for 47 rivuline phenotypes and tried (1968 a, p. 70-71) to construct the karyotypes for 32 phenotypes on the basis of spermatogonial mitotic metaphases.

Material and Methods

The present study of rivuline karyotypes has been based primarily on live material collected by the author during three zoological expeditions to tropical Atlantic Africa (1966, 1968, 1969). Live material collected by D. BLAIR, L. CHRISTENSEN, H . S. CLAUSEN, U. HANNERZ, A. RADDA, E. ROLOFF and A. SCHIDTZ has also been used. For all these samples the exact origin in nature is known. For the additional study, specimens supplied from killi fans or from the trade were used. The exact origin in nature of these samples is rarely known.

The technique used for the study of the gonads has been published previously (SCHEEL 1966).

The morphology of the chromosomes was studied on the basis of metaphases from fin tissues using the following technique: The hindmost part of the caudal fin was removed and the specimen was kept with live food for two days. After this time the hindmost millimeter of the fin was removed. The piece of fin was placed in a diluted Ringer solution (1 :4) with about 0.05°/o of colchicine for 2 hours at about 25OC. Next the piece of fin was transferred to 50°/o acetic acid with 0.05-O.l0/o orcein. After 24 hours of staining the best results were obtained. The slime covering the regenerated tissce usually contained a sufficient number of metaphases which were photographed at nominal 1250x using phase contrast. In copies (4000~) each chromosome arm was measured, not taking into consideration distinct heterochromatic seg- ments. On the basis of these measurements the homologues were paired. The relative length of the individual chromosome arm (tables 1-4) is given as the percentage of the total length of all arms. The results are given in half per cents.

In the tables 1-4 the different major phenotypes are indicated by certain codes (SCHEEL 1968a, p. 7-11) composed of three capital letters. Certain phenoty es consist of a number of allopatric populations which have different karyotypes. These diPferent populations are indicated by codes composed of two capital letters and referring to the origin of the strain.

The Genera Aplocheilus and Pachypanchax

The genus Aplocheilus McClelland, 1839 (including the genera Epiplatys Gill, 1862, Aphyopla- tys Clausen, 1967, and Pseudepiplatys Clausen, 1967) is the most widespread rivuline genus of the Old World. The majority of the phenotypes inhabit the forests of the Atlantic drainages of tropical Africa. Two savannah dwelling phenotypes tend eastwards and reach the Nile. The five Asian phenotypes inhabit the India-Ceylon area, and only one tends beyond this area and reaches the Indonesian islands. All the Asian and most of the African phenotypes were studied.

The genus Pachypanchax Myers, 1933 is very close to Aplocheilus (SCHEEL 1968a, p. 441) and inhabits Madagascar, the Seychelles Islands and Zanzibar. Two major phenotypes are known.

A N N : Aplocbeilus annulatus Boulenger, (SCHEEL 1968a, p. 84-90), ranges from the western lowlands of Guinbe to western Liberia and inhabits swamps and the like. Material from Kasewe, Sierra Leone and from Monrovia, Liberia was studied. - B A R : Aplocbeilus barrnoiensis (Scheel 1968~1, p. 102-103, 453-545), ranges from western Sierra Lcone to western Liberia, probably mostly in swamps. Material from the type locality (Barmoi, Sierra Leone) and from Monrovia, Liberia was used. - BIF: Aplocbeilus bifasciatus Steindachner, (SCHEEL 1968a; p. 112-116), ranges all over the savannahs of West Africa and reaches the Nile. Material from an almost fully fertile Volta/Niger hybrid strain was used. - BLO: Aplocbeilus blocki Arnold, (SCHEEL 1968a, p. 127-131), is an aquarium fish from southern India. - CHA: Aplocheilus chaperi Sauvage, (SCHEEL 1968a, p. 149-154), inhabits the forests of Ghana and Ivory Coast and reaches the N’ZCrCgor6 area, High GuinCe. Material representing three different subphenotypes from Ghana was used. - DAG: Aplocbeilus dageti (Poll), (SCHEEL 1968a, p. 301-303). ranges from southwestern Ghana to western Liberia and inhabits the sedimentary soils. Material from Monrovia, Liberia was used. - DAY: Aplocbeilus dayi Steindachner, (SCHEEL 1968a, p. 176-177), is an aquarium fish from Ceylon. Plate I E. - DUB: Aplocbeilus duboisi (Poll), (SCHEEL 1968a, p. 179-182), inhabits the Central Congo. Material from Stanley Pool was used. - ESE: Aplocheilus eskanus (Scheel) (1968a, p. 185-187, 454-455), inhabits a very

Tabl

e 1

The

kar

yoty

pic

char

acte

rs fo

r 18

dif

fere

nt m

ajor

phe

noty

pes

of t

he g

ener

a A

ploc

heilu

s an

d Pa

chyp

anch

ax o

f th

e O

ld W

orld

PLA

DAG

WE

R

ANN

DA

Y

FAS -

-

BL

O

LIN

-

PA

N

CHA

DUB

SEX

GR

A

-

-

ESE

BA

R

BIF

SA

N

SPI

24

G

F 25

G

F

24

-

F 25

G

F

24

G

F 20

-

F 19

G

F

18

G

F 24

G

F

25

G

-

24

-

F 19

-

F 18

G

-

24

G

-

25

G

F

24

G

F 24

G

F

21

G

F 17

G

F

20

G

F

23/2

4 G

-

24

G

-

17

G

-

48

46

44

43

43

38

38

38?

35 ?

30

28 ?

26

26

25

25 ?

25

25

24 ? ?

11

16

11

10 3 5 6 7 1 ? 0 6 ? 0 0 0 0 2-

3 4 6 4 0

7

<1

<1

<1

<1

<1

<1 1 1 1 ? 1 1 ?

>1 1

>1

>1 1

>1

>1 1 1 1

1.7

2.5

2.0

2.2

2.5

3.0

3.3

4.0

2.5 ? 2.5

3.1 ? 2.2

2.2

2.2

2.2

4.0

4.6

4.2

3.3

2.4

3.3

1.7

2.5

2.0

2.2

2.5

1.7

17

2.

0 2.5 ? 1.7

1.9 ? 1.8

2.2

1.8

1.8

3.2

2.2

1.6

1.8

2.4

1.8

3 2112

2 2'1

2 2 3 3 2'1

2 21

12

? 3 3!/2

? 3 21

12

3 3 1 '1

2

2112

3 3 2 3

Non

e N

one

Non

e N

one

n M

arke

rs

G

F A

I M

EI

I

R1

R2

I

~

Tabl

e 3

Kar

yoty

pic

char

acte

rs fo

r 37 d

iffe

rent

maj

or p

heno

type

s of

the

gen

era

Aph

yose

mio

n an

d N

otho

bran

chiu

s (K

IR, P

AL

, RA

C, T

HI)

of A

frik

a

19

G

20

G

19

G

19

G

19

G

18

G

16/1

8 G

17

G

17

G

17

G

17

G

13

G

20

G

19

G

20

G

19

G

18

G

17/1

8 G

16

G

15

G

13

G

11

G

10

G

10

G

17

G

17

G

16

G

16

G

15

G

14

/15

G

12/1

3 G

10

G

9

G

15

-

-

F F F F F F F F F F F F F F F F F F F F F F F F F F F F F -

- - -

36

27

25

20

19

27

27

26

25

25

19

22

22

34

33

31

31

22

22

19

19

20

22

20

20

22

23

21

23

23

22

22

20

18

1011

2 8'/

z 11

2 0 1 1%

20

1 15%

? 7%

15

25

l/r

10%

13

%

13%

14

14

17

Ih

13%

14

lIr

27

44l/2

45

45

%

1 O'/Z

10

15

16

21

%

26

38

44

42l/2

1.9

2.0 1.9

1.8

2.0

1.7

1.9 ? 2.1 ? 1.

9 1.

8 1.

9 1.

8 1.

7 1.

4 1.

4 1.

3 1.

3 1.

5 2.

0 1.8

1.4

1.6

1.8

1.5

1.7

1.8

1.5

1.4

1.7

1.5

1.7

1.6

3.3

5.5

5.3

5.0

4.2

6.0

7.2

7.3

6.4

5.5

5.6

5.2

4.4

4.8

3.8 3.0

4.8

3.3

4.4 3.6

3.1

4.3

2.2 1.7

2.1

2.7

2.7

4.4

2.9

3.1

4.5

2.7

2.1

2.6

Mar

kers

(Vo

) co

~

- 3 2 2 2 2%

21

12

2%

2 2112

2 2%

2%

2%

2 2%

3 2%

3l

/2

2%

3%

4 3 3

+3

3 4

- 4%

31

/2+4

1/r

3I/z

3I

/z

3%

3l/z

4 3 3

3% + 3

v2

2l/2

+ 3

10,

4Il2

4-5,

8,

7I/z

, 21

/r+51

!z,

5, 5

, l1

/z+3

'/2

11,

1+71

/2,

3lIn

4-4,

6V

t, 1/

z+51

/2,

1+4

2%+8

, 9I

/z,

8, 7

, 6I

/z,

6I/z

, 6

2+8,

9I

/z,

8, 7

, 6%

, 6l

/o,

6 10

1/z,

9lI2

, 8,

6%

, 5l

/z,

5I/a

, 5I

/z

51/2

+91/

z, 2

l/2+7

, 7l

/z,

2l/2

+4,

7, 6

, 1 +

4%,

5'12

71/z

+101

/2,

24-8

, 2+

6,

(24-

6,

4+5)

, 11

/z+4

, 5I

l2

4%+1

0,

54-6

, 2'

/z+8

'/r,

8lh,

1+5

'/r,

6, 5

'12

41/t+

l11/

s,

1 +8l

/t,

4+4I

/e,

l'/n+

61/z

. 11

/z+6

, 7,

6

2+10

, 11

, 24

-8,

11/2

+71/

z, 3+

4,

5%,

1+

4

4+10

, 31

/z+8

1/r,

101/

o, 6

lh,

6, 6

, 5V

z 4+

9,

31/z

4-91

/n,

4+41

/2,

11/t+

61/t,

24-

5, 7

, 6I

l2,

6 21

/2+8

1/r,

5+5'

/2,

10,

3'/e

+5%

, 4+

4%,

3%+

5 9%

, 24

-6,

7l/2

, 1 +

51/t,

6,

21/

e+31

/z,

5lh,

'/t+

4'h

9l/t,

9,

24-6

, 1+

61/t,

1+

51/z

, 5%

, 14

-4,

24-3

l'/

r+71

/~,

lll2

+P

l2,

2+5I

/r,

1 +5,

I'

/s+

~~

/z, 1'/2

+4

54-7

, 11

/r+6

, 11

/z+6

, 1+

41/t

, 1+

41/r

, 14

-4,

1+3l

/t

4'1z4

-7,

2lI2

-l-7,

41

/r+41

/z,

2I/z

+4,

6, 5

l/t

4% + 6

lh,

3 4- 7

, 2+

6I/e

, 4 +

4%,

6, 5

Il2,

(4 + 5

%)

4W

-8,

41/~

+8,

41/s

+41/

t, 9,

8%

, 6,

5%

4+

8%,

101/

t, 9V

2, 9,

44-

4,

31/n

+31/

n,

7, 3

+3l/2

3+

10,

5l/2

+7,

64-6

, 4+

4,

44-4

, 7l

/2,

3lIn

4-4,

6

6+7,

4l

/t+

5,

34-6

1/z,

4+5

I/z,

3+

61/e

, 4+

5,

34-6

6%

+ 6%

, 6%

+ 6%

, 5 +

6I/n

, 4I

lz-k

5Vz,

4% + 5

, 4%

+4'

h 8+

8I/t

, 4+

71/n

, 4+

6lh,

54

-5,

4l/

~+

5, 4

1/z+

41/z

2+

7'/2

, 11

/r+71

/2,

1'/~

+7,

8%

, 7I

/z,

3+31

/z,

6, 5

1/n

21/~

+7,

9, 1

1/z+

71/2

, 8,

14-

6, 3

+3,

1/

e+51

/r,

6, 6

7'

/e+8

, 9%

, 9,

8,

31/z

+31/

~, 7

, 51

/2 34

-7,

4+51

/o,

8, 1

1/r+

6,

7Il2

, 3+

4,

14-6

, 1+

5l/2

5+

71/2

, 4+

5l/r

, 21

/t+61

/2,

2+6,

7,

3+

4,

11/e

+5,

6 51

/z+8

, 2'

/2+7

, 4l

/r+

5,

2+6%

, 24

-6,

8, (

4'/,4

-5)

7+8,

5+

7,

5+51

/z,

2Ve-

I-6,

2+6l

/e,

44-4

, (5

+6)

6l

/z+8

, 5+

9,

5+8l

/2,

41/r

+5,

4l/r

+4l/z

, 4+

4'/r

, 4+

4l/z

6I

l2 + 8

, 7 +

7Ve,

5 + 8, 4

% 4

- 7,

5 + 6l/s

, 3I

/o + 8

, 5 4

- 5%

PH

IJ

N

S

NN

YA

L

O

MO

ES KI

BI

G1

SA

EC

B

A

ND

NN

NS

MB

S

A

1J

CA

B

E

KI

co

MO

LN

MI

JU

MA

K

O

NG

L

W

TY

A

Q

Tabl

e 3 c

ontin

ued

Mar

kers

(Vo

) co

in

1

GI

Fl

.J

sy

~ 1

1 id

s

j 1

CIN

20

G

F

30

13

2.0

4.6

2%

l+lO

1/e,

2l

/2+8

, 7I

/e,

6, 2

+4l!t

, 5,

2+

3,

1+

4

TY

C

OG

13

/15

G

F 18

24

'/e

2.0

4.5

4 71

/d-1

01/e

, 31

/r+7

, 7,

21/

r+31

/r,

6, (

7+7I

/e.

3Iln

i-4)

A

Q

EL

E

10

-

F 18

37

2.

1 3.

5 2

+3

6+

11%

, 64

-10,

6+

6,

54-6

, 4+

5,

4+5,

4+

4l/e

, 6%

A

Q

FIL

18

G

-

24

131/

e 1.

9 5.

0 2l

/2

21/2

+10,

2l/e

+lO

, 8,

7,

7, 3

4-3,

5l/2

, 21

/n+2

1/e

IJ

-

15/1

6 G

F

21

19l/s

2.

3 6.

4 2l

/2

3l/

~+

12

~/e

, 3+10

1/e,

41

/~+

5, 3+

31/e

, 7%

, 7,

(4+

5)

AQ

-

15/1

6 -

F 20

19

Vr

2.5

4.5

3 13

%,

4I/r

4- 5

%,

7I/e

, 7I

/z,

7, 6

, (5

+ 7*/

2, 4

I/o + 5

l/2)

AQ

FRA

11

G

F

22

46

1.5

1.9

3+31

/e

41/n

+8,

6+6,

4+

6,

5+5,

41

/2+5

, 4+

41/t,

4+

41/r

T

Y

GA

R

20

G

F 27

8

1.7

5.3

2 1'1

z4-9

, 8I

/t,

11/z

+6,

1+5,

1+4

1/z,

1/2

+5,

'/e+5

, 5l

/2

EY

-

18/2

0 -

F 28

18

%

2.2

6.7

2 2+

11'/2

, '/t

+9,

l'

/~+

4~

/2, 14

-5,

5I/e

, (5

+5,

4+

5)

EG

-

18

G

F 27

13

1/e

1.7

5.5

2 1'

/~+

9~/n

, 3+8I

/r,

11/e

+6,

21/n

+4,

14-5

, 1+

41/2

, 'Ir

4-5

OW

EX

1 18

G

F

36

33l/n

1.

2 2.

8 1 +

2 2 +

6I/n,

1'/z

+61/

2,

2+ 5

v2,

2 + 4%

, ll/

o + 5

, 3 +

3 N

G

GE

R

20

G

F 21

Z1

/r 1.

5 2.

7 3

8, 7

, 7,

6%

, 5V

2, 5l

/2,

21/2

+21/

1 *Q

GU

L

16

G

F 16

0

2.0

3.1

3I/z

11

, 10

, 9,

8,

7, 6

1/e,

6I/e

, 6,

5I/e

. ~

I/P

IJ

G

UI

19

G

F 24

14

1.

5 4.

5 3

5I/t

+8,

31

/t+6,

3'

/2+4

I/r,

7, 6

, 2+

31!~

, 5I

/r

LA

KIR

18

-

F 29

17

1.

6 4.

2 2%

2+

81/t.

, 11

/n+8

, 2+

6l/e

, 21

/n+5

1/e,

1/2

+6,

2+4,

1+

41/e

A

Q

LA

B

11

G

F 25

30

1.

8 5.

5 1%

+l1

/e

5%+

9,

3+9%

, 41

/e+7

1/~,

4+51

/2,

11/e

+5,

21/~

+4,

6%

AQ

L' - ,,

13

-

F 25

37

I/n

1.5

4.8

3 6l

/2 + 8

, 5%

+ 8,

4% 4

-6,

2 4- 7

I/e,

21/n + 5

, 2l

/2 +

41

/~

NS.

4

LIB

21

G

F

22

3 1.

6 3.

4 2%

L

I1/t,

7I/

t, 7,

34-

3, 6

, 5'1

2 T

Y

LO

U

10

G

-

20

42%

1.

4 1.

7 2+

51/n

6+

61/2

, 5+

7,

4+71

/n,

54-6

, 5+

5, 4

+6,

4+

5.

MI

ME

T

15

-

F 18

ll

l/e

1.8

3.2

3 41

/e+5

, 9I

/e,

9, 8

, 8,

4+

4,

7, 3

+3I/

e,

6l/e

, 6

TY

M

IR

16/1

9 -

F 30

28

1/r

1.9

6.5

l/e+

ll/e

2+

10'/r

, 3+

31/e

, 11

/e+5

, (4

+9,

5+

5l/r

, 31

/2+5

1/e)

T

Y

ND

I 20

G

-

22?

21

2.3?

6.

0 2

12,

9I/z

, 7,

6I/n

, 6,

6,

5I/z

T

Y

occ

23

G

F 44

(3

0 1.

3 3.8

2%

. 1 +

7, ll

/r +

4%,

'/e +

5%,

1 + 4%

A

Q

OG

O

20

G

-

36

21

1.4

2.5

3 2+

5'/e

, 7,

11/

n+5,

6l/z

, 11

/e+5

, 11

/e+4

1/e,

11/

e+4

AQ

PA

L

17

-

F 23

17

1/n

1.6

4.0

3I/n

6+

8,

5+51

/r,

9, 7

, 6I

/a,

6, 2

+311

2 19

57

-

17/1

8 -

F 20

9'/

2 1.

6 4.

0 3I

/n

6+8,

(2

+8,

11

/e+5

1/2)

, 9,

7lh

, 5I

/r

1956

P

ET

20

-

F 28

16

1.

1 3.

2 2l

/2

31/r

+41/

r, 31

/n+4

, 31

/2+3

1/r,

3+31

/r,

21/t+

4,

6 A

Q

RA

C

9G

- 1

5 34

l/2

1.8

2.9

5%

7% + 8

%,

6% +

9%,

7 4- 7

%,

6 + 7%

, 3%

+ 8, 4

+ 5%

, 7%

A

Q

RO

L

21

G

F 23

5

1.6

3.4

2%

8I/z

, 2+

5*/r

, 7,

3+

3,

6, 5

I/r

AQ

SC

E

20

G

F 35

17

1.

9 4.

4 2%

1+

10,

1+71

/n,

11/2

+5%

, 11

/r+5

%,

1%+4

%,

1+41

/s

AQ

SC

H

11

-

F 18

32

1.

5 2.

6 6

84-8

, 4+

7%,

5+5%

, 5+

5'/r

, 3+

7,

4+4,

8,

7,

6%

AQ

SE

M

14

G

F 24

13

1.

8 4.

2 3'

/r 6'

/e+8

, l1

/e-!

-9'/r

, 9I

/z,

1+7,

8ll

2,

7I/s

, 7,

b1/2

T

Y

SIS

20

G

F 20

7

2.0

4.2

2%

101/

e, 8

, 1+

61/r

, 6I

/e,

1+5,

1+

4I/e

, 5,

1+

4

TY

SJ

O

20

G

F 20

0

1.9

5.0

2 10

, 7,

6,

6, 6

, 6,

5I/z

A

Q

TH

I 21

/22

G

F 24

6l

/2

1.6

4.3

2 tI1

/2,

(4+

4),

7'12,

6, 5

lh,

5l/2

, 1

+2

A

Q

TO

D

23

G

F 43

31

0.

9 3.

0 21

/2

21/2

+5,

1'/~

+4~

/n, 21

/2+2

1/z,

2+

3%,

24-3

, l'/

n+31

/e,

5 T

Y

VA

L

18

G

F 24

6

1.8

3.0

3%

1 +91

/e,

1 +91

/r,

1% + 5

%,

% +6

1/z,

6,

+ 5

%,

5%

AQ

a

A

Sr.

2.

hi s 2. 3

196 J. J. Scbeel

v1

1

w E

C

m . * + s -l. ul ul

*" 111-

u 9) + u u u = 0 u u mi 0 u u u u u u u c c c * c c c + c c c c c 0 c 0

Z Z m Z Z Z ~ Z Z Z ~ Z Z Z Z Z Z Z Z 8 8; 0 0 o+ 0 0 0 0 0 0 0 0 0 0 0

001 I I I I I I I I I I I l u l u l

restricted area north of Eseka, East Cameroon. Ma- terial from the type locality was used. - FAS: Aplocbeilus fasciolatus Gunther, (SCHEEL 1968a, p. 194-203), inhabits the forest and derived savannahs of Lower GuinCe, Sierra Leone and Liberia. Specimens from a number of populations of Sierra Leone and from Monrovia, Liberia was used. Plate 111 B. - G R A : Aplocbeilus grabami Boulenger, (SCHEEL 1968a, p. 231-235), ranges from southwestern Nigeria to the Mouth of the Congo, generally in swamps. Mate- rial from the Douala area, East Cameroon (n=23/24), and from Rio Benito drainage, Rio Muni was used. Plate I1 B. - L I N : Aplocbei- lus lineatus (C and V), (SCHEEL 1968a, p. 266-269), is an aquarium fish from southern India. Two different strains were used. - PAN: Aplocbeilus pancbax (Ha- milton-Buchanan), (SCHEEL 1968a, p. 347-352), ranges from India to Thailand and Indonesia. Material from Bangkok, Thailand (n = 19) and from an aquarium strain said to have originated from Thailand (n- 18) was used. - PLA: Pachy- pancbax playfairi (Gunther), (SCHEEL 1968a, p. 371-374), is an aquarium fish probably from Zanzibar. Plate I1 A. - S A N : Aplocbeilus sangmeli- nensis (Ahl), (SCHEEL 1968a, p. 371-374), inhabits the forest of the inland plateau of East Cameroon and Rio Muni. Material from Sang- melima, E. Cameroon was used. - SEX: Aplocbeilus sex- fasciarus (Gill), (SCHEEL 1968a, p. 380-395), ranges from eastern Ghana to Ga- bon in the western lowlands, Material from Ghana, southern Nigeria, Fernando Po, the Cameroons, and Rio Muni was used. - SPI: Aplo- cheilus spilargyreius (Dume- ril), (SCHEEL 1968a, p. 375- 378), ranges over the savannahs of West Africa and reaches the Nile and the Central Congo. Material


from Kontagora, northern Nigeria was used. - WER: Aplocheilus werneri Meinken, (SCHEEL 1968a, p. 431-432), is an aquarium fish, probably from Ceylon.

The Tribe Rivulidi

The phenotypes of the tribe Rivulidi Hoedeman are characterized by the frontal scalation (SCHEEL 1970b, p. 360-361). The posterior part of the A-scale not covered by the anterior part of the B-scale. The many phenotypes of the genus Rivulus Poey are non-annual rivulins and are widespread over South and Central America, the Antilles, Mexico and Florida, USA. The phenotypes of the genus Pterolebias Garman are highly annual rivulins inhabiting inland temporary waters in S. America. The names used by killi fans are attached to the phenotypes of the genus Rivulus. See also SCHEEL 1970b, p. 358-370, 403-413). The exact identification of strains of Rivulus is not possible for the time being.

C Y L : Rivulus cylindraceus Poey. (Cuba). - D O R : Rivulus dorni Myers. (Rio de Janeiro, Brazil). - H A R : Rivulus harti (Boulenger). (Northeastern S. America). - HOL: Rivulus holmiae (Eigenmann). (Northeastern S. America). - LON: Pterolebias longipinnis Garman. (Lower Amazon). Plate V F. - MAC: Pterolebias maculipinnis Radda. (Venezuela). - MAR: Rivulus marmoratus (Poey). (Florida, Cuba ect.). - M I L : Rivulus milesi Fowler. (Colombia). Plate I1 C. - OCE: Rivulus ocellatus Hensel. (Brazil). - ORN: Rivulus ornatus Hensel. (Lower Amazon). Plate I11 C. - P E R : Pterolebias peruensis Myers. (Upper Amazon). - STR: Rivulus strigatus Regan. (Amazon). - URO: Rivulus urophthalmus Giinther. (Lower Amazon).

The Tribe Cynolebiatidi

The tribe Cynolebiatidi Hoedeman is characterized by the frontal scalation (SCHEEL 1970b, p. 360-361, 384-390). The posterior part of the A-scale is covered. The phenotypes of this tribe range from the Pacific coast of Colombia to Argentina, along the Atlantic coast. Most populations inhabit the coastal lowlands.

BEL: Cynolebias bellotti Steindachner. (La Plata). - DOL: Austrofundulus dolichopterus Weitzman and Wourms. (Venezuela). - MEL : Cynopoecilus melanotaenia Regan. (Brazil). - N I G : Cynolebias nigripinnis Regan. (La Plata). - TRA: Austrofundulus transilis Myers. (Vene- zuela). Plate I11 A. - W H I : Cynolebias whitei Myers. (Rio de Janeiro, Brazil).

The Genus Aphyosemion

The genus Aphyosemion Myers is the major object of the author’s study of the rivulins. The many phenotypes of this genus inhabit the forests of tropical Atlantic Africa. Only two forms are exclusively restricted to the savannahs. Several of the phenotypes are annual rivulins. A complete description of the karyotypes is given in subsequent papers dealing with the detailed study of some of the phenotypes and with the phylogeny of the genus.

A R N : Aphyosemion arnoldi (Boulenger), (SCHEEL 1968a, p. 91-97), inhabits the delta of the Niger River. Material from Port Harcourt was used. - B I V : Aphyosemion bivittatum (Lonnberg), (SCHEEL 1968a, p. 116-127), ranges from southern Togo to southern Rio Muni. The author’s study of the rivulins is centered in this phenotype. Live material from twenty- three populations was studied together with the results of twenty-five different crossings. A detailed report of this study will be published by the MusCe Royal de I’Afrique Centrale, Tervuren, Belgium subsequently. The following strains have been studied : BI : Lower Rio Benito, Rio Muni. EC: Rio Ecucu, Rio Muni. ES: Eseka, E. Cameroon. G I : Wuri drainage, E. Cameroon. IJ: Ijebu-Ode, Nigeria, KI: Kienke drainage, E. Cameroon. LO: Lob6 drainage, E. Cameroon. MO: Lokundje drainage, E. Cameroon. N N : Lower Nyong, E.Cameroon. NS: Lower Nyong, E. Cameroon. SA: Sanaga drainage, E. Cameroon. YA: Dibamba drainage, E. Cameroon and have different karyotypes. Plate I A-C. and fig. 1. - B U A : Aphyosemion buala- num (Ahl), (SCHEEL 1968a, p. 134-139) inhabits the savannahs of East Cameroon and the adjacent parts of West Cameroon and the Central African Republique. Two strains: BA: Bamkin, East Cameroon, and ND: Ndop, West Cameroon. - C A L : Aphyosemion calliurum (Boilenger), (SCHEEL 1968a, p. 80-84, 97-101, 140-145) ranges from southwestern Nigeria to the Lowcr Congo in the lowlands. The different subphenotypes are not well separated. The detailed study of eleven different populations subsequently will be published. BE: Benito drainage, Rio Muni. CA: an aquarium fish, probably from Gabon. CO: Longji, E. Cameroon. IJ: Ijebu Ode,

188 1. /. Scheel

Plate I . Somatic metaphases of: A=The YA population; B=The NS PO ulation; C=The NN population of Aphyosemion bivittatum (Loennberg); D =The LW femaL of Aphyosemion

All chromosome photos are 2000 X. cameronense (Boulenger) with 2n=25; and E= Aplocheilus dayi


Plate I I . Somatic meta hases of A=Pachypanchax playfairi; B= Aplocheilus grahami; C = Rivulus m i i s i ; D = Aplocheilichthys normani; E = Aphyosemion toddi

190 J . /. Scheel

Plate I l l . Somatic metaphases of A = Austrofundulus transilis; B= Aplocheilus fasciolatus; C = Rivulus ornatus; D = Aphyosemion petersii; E = Aphyosemion exiguum


Plate IV. Somatic metaphases of A=Nothobrmchius thierryi; B= Aphyosemion cinnamomeum; C = Aphyosemion gulare; D = Aphyosemion calliurum (KI); E = Aphyosemion calliurum (MO)

192 1. J . Scbeel

Plate V. Somatic metaphase of A=Aphyosemion seymoxri, B=Aphyosemion cognatum (2n= 29); C=Aphyosemion cbristyi; D=Aphyosemion aff. cognatxm (NS 4); E=Aphyosemion

gardneri (EG); F= Pterolebias longipinnis


Nigcria. KI: Kienke drainage, E. Cameroon. MB: Wuri drainage, E. Cameroon. MO: Memt drainage. W. Camercon. N N : Lower Nyong drainage E. Cameroon. NS: Lower Nyong drainage, E. Cameroon. SA: Sanaga drainage, E. Cameroon. Fig. 4 and plate IV D-E. - CAM: Aphyosemion cameronense (Boulenger), (SCHEEL 1968a, p. 145-148) replaces A. calliururn on the basement complex generally. This phenotype ranges from the Mungo drainage, W. Ca- mercon to Gaboon. A detailed report of the study of ten populations subsequently will be published. BE: Benito drainage. Rio Muni, EC: Ecucu drainage, Rio Muni. JLJ: KelC drainage, E. Cameroon, KO: Nyong drainage, Lolodorf Hills, E. Cameroon, LN: Lokundje drainage, Lolodorf Hills, E. Cameroon. LW: Lokundje drainage, Lolodorf Hills, E. Cameroon. MA: Mungo drainage, W. Cameroon. MI: Nyong drainage, Lolodorf Hills, E. Cameroon. NG: Nyong drainage, Lolodorf Hills, E. Cameroon. YA: Inland plateau, west of Yaunde, E. Came-

YA

(11111111 GI I 1 Irr I!! Fig. 1. The probable evolution of the karyotypes of four populations of Aphyosetnion bivitta- tr?m of East Cameroon. The N N and the YA karyotypes probably derived from the NS karyotyFe after a minimum of five and six pericentric inversions respectively. These inversions transformed two-armed chromosomes into telocentrics. The G I karyotype probably derived

from the YA karyotype after two centric fusions of telocentrics.

194 J. J. Scheel

roon. The karyotypes of the BE, EC, LN and YA populations were similar. The BE and EC strains were, however, reproductively isolated by sterility of hybrids. Plate I D. - CEL: Aphyosemion celiae Sheel, (1971a, p. 48-66), appears to be endemic to the Upper Mungo, W. Cameroon. Material from the type locality was studied. - CHR: Aphyosemion christyi (Bou- !enger), (SCHEEL 1968a, p. 158-166), inhabits the Central Congo. Several aquarium strains were studied. Plate V C. - CIN: Aphyosemion cinnamomeum Clausen, (SCHEEL 1968a, p. 166-

I

I I I I I I F

Fig . 2. Schematic presentation of the principles of the hypothetical evolution of the rivuline karyotype by alternating series of pericentric inversions converting metacentrics into telocentrics followed by centric fcsions of the telocentrics. The first, second, and third generation

of two-armed and one-armed elements are figured out for the n=24 karyotype.


168), appears to be restricted to the Upper Mungo, W. Cameroon. Material from the type locality and from Badouma, closer to Kumba, was studied. Plate IV B. - COG: Aphyosemion cognatum Meinken, (SCHEEL 1968a, p. 170-172), inhabits the Central Congo. Several aquarium strains, probably from the Stanley Pool area were studied. Plate V B and V D. - ELE: Aphyose- mion ufl. elegans (Boulenger), (SCHEEL 1968a, p. 183-184), inhabits the Central Congo. Aquarium material was used. - EXI: Aphyosemion exiguum (Boulenger), (SCHEEL 1968a, p. 188-192), lives in the forests of the inland plateau of East Cameroon and in the adjacent parts of Gabon and Rio Muni. Material from Sangmelima and from the Lolodorf Hills, E. Came- roon was studied. PI. 111 E. - FIL: Aphyosemion filamentosurn (Meinken), (SCHEEL 1968a, p. 202-210), lives in southwestern Nigeria. Material from Ijebu-Ode (IJ) and from several aquarium strains was studied. - FRA: Aphyosemion franzwerneri Scheel, (1971a, p. 48-66), lives in the forests south of Yabassi, E. Cameroon. Material from the type locality was used. - G A R : Aphyosemion gardneri (Boulenger), (SCHEEL 1968a, p. 215-228), inhabits the forests and derived savannahs on the basement complex in Nigeria and West Cameroon. EG: Lake Ejagham, W. Cameroon. EY: Eyomojok, W. Cameroon. OW: OWO, southwestern Nigeria. Plate V E. - GER: Aphyosemion geryi Lambert, (SCHEEL 1968a, p. 228-231), is known from

F i g . 3 . Four extremes of the n=20 karyotype of the genus Aphyosemion. A=A. ogoense, 36 arms, index 1.4, 7l/z% (the longest element); B=A. petersii, 25 arms, index 1.1, 8%; C=A. bittutum (IJ), 25 arms, index 2.0, 11 z; D = A . sjoestedti, 20 arms, index 1.9, 10%. Arms less

than ' 12 % are not considered

196 J . J . Scheel

Lower GuinCe and Sierra Leone. Aquarium material was esed. - G U I : Aphyosemion guineense Daget, (SCHEEL 19681, p. 241-244), is restricted to the mountains of northern Sierra Leone and GuinCe. Material from Lago, Sierra Leone and from the aquarium was used. - GUL: Aphyosemion gulare (Boulenger), (SCHEPL 1968a, p. 244-248), ranges over southwestern Nigeria. Material from Ijebu-Ode, Nigeria Plate IV C. - LAB: Aphyosemion labarrei Poll, (SCHEEL 1968a, p. 258-262), is an aquarium fish from the Lower Congo. - L I B : Aphyosemion liberiense (Boulenger), (SCHEEL 1968a, p. 139-140, 264-266), lives near Mon- rovia, Liberia. Material from this area was used. - LOU: Aphyosemion louessense (Pellegrin), (SCHEEL 1968a, p. 277-279), inhabits the upper reaches of the Kouilou River, Brazzaville- Congo. Material from Mindouli was used. - MET: Aphyosemion melanopteron Goldstein and Ricco, (G and R: 1970, p. 8-11), is an aquarium fish, probably from the Lower Congo. Aquari- um material was used. - MIR: Aphyosemion mirabile Radda, (RADDA 1970a, p. 58-61), inhabits the slopes bordering the eastern part of the Mamfe plains, W. Cameroon. Material was kindly offered by Radda. The large variation of the karyotype in this material indicate a hybrid origin of the strain. Three subphenotypes are known. - NDI: Aphyosemion ndianum Scheel (1968a, p. 310-316, 455-456), lives near Osomba, the Oban Hills, East Nigeria. Material from the type locality was used. - occ: Aphyosernion occidentale Clausen, (SCHEEL 1968a, p. 328-

F i g . 4 . Four of the eight known karyotypes of Aphyosemion callicrum from East Cameroon and Rio Muni. A=The N N strain from the lowlands just north of the Nyong River. B=The SA strain from the lowlands north of the Sanaga River. C=The BE strain from the lowlands north of the Benito River. D=The C O strain from Longji, north of Kribi, E. Cameroon

Kivufine Karyotypes and their Evolution 197

337), lives in southeastern Sierra Leone. Aquarium material was used. - OGO: Aphyosemion ogoense (Pellegrin), (SCHEEL 1968a, p. 280-;83, 338-339, 417-418), is an aquarium fish from the Central Congo. Fig. 3 A. - PET: Aphyosemion petersii (Sauvage), (SCHEEL 1968a, p. 354- 357), inhabits the forests of Ghana and Ivory Coast. Material from Ivory Coast was used. Plate 111. D. and fig. 3 B. - ROL: Aphyosemion rofoffi Roloff, (SCHEEL 1968a, p. 366-367), lives in Sierra Leone. Aquarium material was used. - SEM: Aphyosemion seymouri Blair and Loiselle, (B and L 1972, 1-11) was discovered near the Mouth of the Volta. Material from the type locality was studied. Plate V A. - SIS: Aphyosemion santaisabellae Scheel, (1968b, p. 332-342), is endemic to Fernando Po. Material from the type locality was used. - SCE: Aphy- osemion scheefi Radda, (RADDA 1970b, p. 117-180), is an aquarium fish, probably from southeastern Nigeria (SCHEEL 1968a, p. 139). Formerly called the “Burundi-Aphyosemion”. - SCH: Aphyosemion schoutedeni (Boulenger), (SCHEEL 1968a, p. 157, 374-375), inhabits the Central Congo. Aquarium material was used. - SJO : Aphyosemion sjoestedti (Lonnberg), (SCHEEL 1968a, p. 400&407), is an aquarium fish from southwestern Nigeria. This variable phenotype ranges to West Cameroon. Fig. 3 D. - TOD: Aphyosemion toddi Clausen, (SCHEEL 1968a, p. 425-427), represents the western subphenotype of the A . occidentafe complex in Sierra Leone. - WAL: Aphyosemion walkeri (Boulenger), (SCHEEL 1968a, p. 430-431), inhabits the forests of Ghana and Ivory Coast. Aquarium material was used.

The Genus Notobranchius

The phenotypes of the genus Nothobranchius Peters inhabit temporary fresh water in tropical Africa. The subgenus Nothobranchius generally is restricted to East Africa with one phenotype near the Lake Chad. The scbgenus Fundufosorna Ah1 lives in West Africa. The phenotypes of this genus have not recently been revised, and appear to grade into the genus Aphyosemion.

K I R : Nothobranchius kirki is an aquarium fish from Rhodesia. - PAL: Nothobranchius pafm- questi (Lonnberg), (SCHEEL 1968a, p. 342-347), is a variable aquarium fish, probably from Tanzania. Two phenotypes labelled “1956” and “1957” were studied. - RAC: Nothobranchius racbovii Ahl, (SCHEEL 1968a, p. 360-363), is an aquarium fish from the Beira area of Mozam- bique. - T H I : Nothobranchius (Fundufosoma) thierryi (Ahl), (SCHEEL 1968a, p. 421-424), inhabits the savannahs of West Africa. Two strains from Ghana were studied. One of these originated from the Mouth of the Volta. Plate IV A.

Additional Atheriniform Karyotypes

Nineteen non-rivuline, atheriniform karyotypes were studied to cast light on the composition of the karyotypes of forms more or less closely related to the rivulins.

Cyprinodontidae

Subfamily Procatopodinae A B E : Procatopus aberrans Ahl, (SCHEEL 1970a, p. 4-14, 1970b, p. 439-440), inhabits the soils of the basement complex in Nigeria and West Cameroon. Material from Eyomojok (P. nigromarginatus Clausenj and from Mamfe (P. roseipinnis Clausen) was used. - MAO: Aplo- cheifichtbys macrophthafmus Meinken, (SCHEEL 1970b, p. 436), is an aquarium fish from southwestern Nigeria. The phenotype ranges from southern Togo to Rio Muni with three subspecies. - NOR: Apfocbeificbthys normani Ahl, (SCHEEL 1970b, p. 436-437), inhabits permanent freshwater on the savannah of West Africa and reaches the Nile. Material from Monrovia, Liberia ( A . macrurus manni Schultz) was studied. Plate I1 D. - SIM: Procatopus similis Ahl, (SCHEEL 1970a, p. 439-440), inhabits sedimentary soils in Nigeria and all types of soil in the Cameroons and Fernando Poo. Material frcm the Lower Wcri and from Fernando Po0 was studied. - SPI: Aplocbeificbthys spifauchen (Dumeril) inhabits brackish water frcim the Mouth of the Sene- gal River to the Mouth of the Congo River. Aquarium material was used.

Subfamily Fundulinae CON: Fundufus conf2uentus Goode and Bean, (SCHEEL 1970b, p. 423). An aquarium strain was used. The karyotype is very similar to those which were studied by CHEN and RUDDLE (1970, p. 255-267). - MUL: Adinia muftifasciatus Girard. An aquarium strain was used. - OLI: Fundu- Ius ofivaceus (Storer). Aquarium material was uscd.

198 J . J . Scbeel

Subfamily Cyprinodontinae

FLO: Jordaneffa floridae Goode and Bean, (SCHEEL 1970b, p. 426-427). Aquarium material was used.

Subfamily Aphaniinae VAL: Valencia hispanica (C and V), (SCHEEL 1970b, p. 432). Aquarium material was used.

Subfamily Oryziatinae

MEL: Oryzias mefastigma (McClelland). Aquarium material was used.

Poeciliidae

(The gambusinos)

HEL: Xiphophorus hefferi Heckel. Material from the University of Giessen, Germany. - MAC: Xiphophorus macufatus (Giinther). A number of strains from the University of Giessen. Ger- many. - S P C : Gambusia species. Introduced mosquito fish in Tenerife. - RET: Poecilia reticufatus (Peters). Aquarium material.

Exocoetoidei

(The halfbeaks) PUS : Derrnogenys pusiffus Hasselt. Aquarium material. (E. Asia).

Atherinoidei

(The silversides)

LAD: Tefmaterina fadigesi Ahl. (Celebes). Aquarium material. - G E A : Bedotia geayi Pellegrin. (Madagascar). Aquarium material. - MCC: Mefanotaenia maccufocbi Ogilby. (Australia). Aquarium material.

Rivuline Karyotypes

The karyotypic characters (tables 1-4) have been based primarily on metaphases from fin tissues of both sexes. In most cases the haploid number had been found in advance in primary meiotic metaphases from males. No morphologically dissimilar sex-chromosomes were found. The haploid number (gonads) corresponded to the diploid number (fins).

Rivuline karyotypes are thought to be characterized by the following major distin- guishing features:

1. The haploid number 2. The symmetry of the chromosomes 3. The symmetry of the complement 4. The marker chromosomes

These characters are given in a schematic outline in the tables 1-4. 1. The haploid number (n) directly found in gonads of males (G) or indirectly (halved

diploid number) in fins (F).


2. The symmetry of the chromosomes. Metacentrics represent symmetrical elements, telocentrics the extreme of asymmetry. For all karyotypes the number of chromosome arms (A) independent of their length is given. For the karyotypes of the primary study (the genera Aphyosemion and Nothobranchius) the chromosome symmetry also is expressed by the total length of the short arms (SY, %) of two-armed elements. Arms of less than 1/z% (the error of measurements) are not considered. For the karyotypes of the additional study (tables 1, 2 and 4) the number of symmetrical elements (ME) is given. In some karyotypes one or more chromosome arms conspicuously exceed the length of the largest chromosome of the basic rivuline karyotype (p. 201). The presence of such “oversized” arms is indicated by the index (I), which gives the number of times the longest chromosome arm makes the length of the longest chromosome of the basic karyotype.

3. The symmetry of the complement. Karyotypes with chromosomes of similar size are considered to represent symmetrical complements. R1: gives the number of times which the longest chromosome makes the length of

the shortest. R2: gives the number of times which the longest, apparently basic, chromosome

makes the length of the shortest. This character cannot be given for most karyotypes of the primary study (table 3) because in these highly specialized complements there is usually no distinct grouping by length of the basic elements on one hand and the elements resulting from centric fusions on the other hand. gives the relative length (%) of the shortest element of the complement and indicates the relative constancy of the length of this particular element.

4. The marker chromosomes. Rivuline chromosomes in the absence of secondary con- strictions are not well suited as markers. Large, probably derived, two-armed elements, oversized telocentrics (I > 1) and small symmetrical elements are referred to the markers. Where the karyotype contains many such elements only the largest and most characteristic ones are given.

(CO) in table 3 refers to the different strains of the primary study. The abbreviations are explained in p. 187-197. (AQ) stands for an aquarium strain, usually of unknown origin. (TY) stands for a strain from the type locality, the type area or for a strain of aquarium fishes from which the types originated.

Heterozygous karyotypes were found in some of the strains. The variation of the haploid number is indicated (n=17/18 in the CAL-SA strain indicates that 2n= 34-35-36 was found in that strain). The two-armed elements which occur in the heterozygous condition are given in brackets under the markers. The symmetry index (SY) refers to the lowest haploid number of heterozygous complements.

S:

Discussion

A. The most primitive Karyotype

MYERS (1 955) considered the rivulins to represent the most generalized cyprinodonts and among the rivulins he pointed out the genus Aplocheilus as representing the most generalized taxon.

On the basis of an extensive osteological study SETHI (1960, p. 208) said, that “the aplocheilids (= the rivulins of the present study), or a stock similar to them almost certainly represent the ancestral stock from which all other major groups of both the oviparous and the viviparous cyprinodont fishes have originated.. . . Aplocheilus of

200 J. J. Scheel

Asia, Rivulus of South and Central America and probably Epiplatys of Africa (included in Aplocheilus in the present study) appear to be the most generalized of the aplocheilids.. . . Except for the cyprinodontids (the subfamily Cyprinodontinae of the present study) which probably originated from fundulids (Fundulinae) all other groups of oviparous cyprinodonts have evolved from the aplocheilids or a stock similar to them”.

The genus Pachypanchax appears to be very close to Aplocheilus (SCHEEL 1968 a, p. 441) and may be tentatively included among the generalized cyprinodonts.

The most generalized karyotype is consequently likely to be present in the genera Aplocheilus and Riuulus and probably in Epiplatys and Pachypanchax as well. Thirty- one karyotypes referrable to these four genera were studied (tables 1 and 2). These karyotypes are very different. The haploid number ranges from 17 to 25, the number of chromosome arms (A) from 24 to 48, and the ratio of lengths (Rl) from 1.7 to 4.6.

The karyotypes which exhibit the largest variation of the length of the chromosomes also have the lowest haploid number and the largest elements are almost symmetrical. These large and symmetrical elements thus appear to represent the results of centric fusions. When these large elements are thought to represent two basic chromosomes the variation of the haploid number is reduced to n=23-25.

I t has been assumed (EBELING and CHEN 1970, p. 135) that 2n=48 acrocentric chromosomes constitute the “basic” modal complement of cyprinodontoid fishes and most other teleost groups. These authors also assumed that any deviation from the “nombre fondamentale” of 48 (diploid) chromosome arms might be construed as a derived karyotype.

The thirty-one karyotypes of the apparently most generalized rivuline genera differ markedly by the number of chromosome arms. Even when the short arms are not considered the nombre fondamentale (NF) reaches at least 70-82 in the diploid complements of the four phenotypes first mentioned in table 1. Furthermore, the chromosomes of these four karyotypes are very small and difficult to observe, (pl. I1 A). Thus these four karyotypes could not have derived from a doubling of the number of chromosomes of one of the karyotypes with NF = 24-25.

When the elements of the fourteen karyotypes with n=24-25 are arranged according to increasing length and without regard to the position of the centromeres, these fourteen karyotypes are very similar or almost identical, apart from some variations of the shortest and the longest elements. Most of the chromosomes of these karyotypes thus appear to represent homologues which differ only by the position of the centromeres. The replacement of the centromere by pericentric inversions is a reversible process in two-armed elements. Consequently the original position of the centromere cannot be pointed out. When, however, the centromere accidentally comes into a terminal position, after a pericentric inversion, no type of chromosome mutation is known at present which is able to move the centromere towards a more median position. The telocentric condition thus appears to represent an irreversible condition. When the telocentric condition has been reached one chromosome arm appears to have been definitely lost.

Thus it seems that the n=24-25 karyotype with a maximum of chromosome arms represents the most generalized rivuline karyotypes.

To cast mcre light on the problems concerninp the basic cyprinodontoid karyotype the cytological study was directed towards non-atheriniform fishes. More than 150 different phenotypes representing more than thirty different families of teleosts were studied on the basis of metaphases from fin tissues. I t was realized that karyotypes with haploid numbers from 23 to 46 and with a total, or almost total, of distinctly two-armed elements and with a variation of the length of the chromosomes similar to that of the assumed basic rivuline karyotype were much more common among these


fishes than among the rivulins. The results of this additional study will be published subsequently.

I t has generally been assumed (POST 1965, OHNO and ATKIN 1966, ROBERTS 1967 and 1970, MURAMOTO et al. 1968, EBELING and CHEN 1970), that n=24 is the modal or basic number of teleost fishes.

At least eight of the karyotypes of the genus Aplocheilus (table 1) have n=25 or appear to have derived from karyotypes having this haploid number by centric fusions. Pterolebias pevuensis (table 2) has n = 27 and 45 chromosome arms.

If the shortest diromosomes of the basic rivuline karyotype become one-armed and fuse, the resulting element will not exceed the length of the longest basic element (characters R1 and R2 in the tables 1 and 2) generally. In this way a cryptic reduction of the haploid number may take place. The karyotype of Rivulus wilesi (table 2 and pl. I1 C) has n=23 and the karyotype contains no conspicuously enlarged elements. The variation of the length of the elements of this karyotype corresponds to that of Rivulus dorni with n=24 and to that of the assumed basic rivuline karyotype. The karyotype of R. milesi may have undergone a cryptic reduction by the fusion of two of the shortest elements. The conditions in Melanotaenia macculochi and in Poecilia reticulatus are similar, (table 4). Both karyotypes have 23 strictly one-armed elements, whereas other karyotypes of the corresponding families have n = 24. Probably these two n=23 karyotypes have undergone a cryptic reduction of the haploid number similar to that of R. milesi and later on, the derived two-armed element became telocentric by pericentric inversions.

Thus the basic haploid number of the subfamily Rivulinae was probably larger than 24. The basic haploid number of the other cyprinodont subfamilies, of the family Poeciliidae and of the suborders Exoetoidei and Atherinoidei probably is 24.

The shortest element of the assumed basic rivuline karyotype is about half the length of the longest element and the elements are graded evenly in length. This means that the shortest and the longest elements are 2/3 and 4/3 of the mean basic length or about 2.7% and 5.3 % respectively in the n=25 karyotype.

Chromosomes shorter than 2 % of the total length of all chromosomes are very rare in rivuline karyotypes and actually were found only in one karyotype (GRA. 23/24 in table 1). The short (lt/z%) element of this karyotype may represent the remains of an unequal translation of the “Robertsonian” type heterozygous to this karyotype. The short element may be lost in subsequent generations. Elements shorter than 2 % probably are not able to produce sufficiently stable chiasmata to secure proper segre- gaticn of the short elements at anaphases. Chromosomes which are more than S1/z% (+ 1/z % for errors in measurement) may consequently represent derived elements resulting from centric fusicns, tandem fusions or unequal translocaltions. Pterolebias peruensis (table 2) with n=27 has four large metacentrics. The karyotype also contains four small similar elements (11/2+2%, I1/2+11/z%, 11/z+I1/z%, 1 +I1/z%). These eight metacentrics probably represent the result of four unequal translocations and the basic haploid number is probably 27 and not 31.

A close study of the marker chromosomes and the length of the longest element of the karyotype (S x R1 and S x R2) shows that (tables 1, 2 and 4) chromosome arms exceeding 6% (5 l /z%+l /z%) are very rare in these karyotypes and that arms exceeding 7 % occur only in Rivulus ornatus and in Pterolebias longipinnis. Arms larger than 6% occur, hcwever, in almost all the many karyotypes of the primary study (table 3). The term “index” (p. 199) was introduced to point out the karyotypes with “oversized” chromosome arms.

202 /. /. Scheel

B. Evolutionary Mechanisms

Two types of chromosome mutations were probably common during the evolution of the rivuline karyotype: pericentric inversions and centric fusions.

1 . Pericentric inversions

Autosomal polymorphism due to pericentric inversions has been described and an- alyzed by MATTHEY (1966) and by OHNO et al. (1966). Such polymorphism was not disclosed in the heterozygous state in any natural strain. Some of the aquarium fishes, probably of hybrid origin, exhibited such polymorphism, however. Autosomal polymorphism due to pericentric inversions appears to be present in some of the populations of Aphyosemion bivittatum (BIV in table 3) from East Cameron, (fig. 1 and plate I. A-C).

The NS population inhabiting the soils of the basement complex in the lowlands of the southern part of the Nyong drainage has n = 19 and 25 chromosome arms. Six ringbivalents, one large and five small, are present in meiotic metaphases of males.

The N N population inhabits the sedimentary soils just north of the Nyong River, and 30 kms north of the NS locality. n = 19 and 20 chromosome arms. Only one ring- bivalent (large) in primary meiotic metaphases of males.

The YA population inhabiting the soils of the basement complex of the Dibamba drainage, 30 kms north of the N N locality, has n=19 and 19 arms. No ringbivalents occur in primary meiotic metaphases of males.

When the chromosomes of these three karyotypes are arranged according to decreas- ing lengths full agreement (within or less of the total length) is obtained between the apparent homologues, (fig. 1). The NN karyotype appears to have developed from the NS karyotype after five pericentric inversions which transformed the five short two-armed elements of the NS to telocentrics. The YA karyotype may have developed from the NN karyotype after one pericentric inversion which transformed the largest and two-armed (2+8%) chromosome of the NN (and NS) karyotype into a telocentric.

2. Centric fusions

Numerous examples of polymorphism due to centric fusions have been described in literature (WHITE 1954, p. 135-155; JOHN and LEWIS 1968, p. 76-87).

Similar polymorphism was disclosed in at least eleven different strains of the primary study (table 3) and was present in seven natural populations.

Such polymorphism was found in two of the five strains of Aphyosemion cameronense from the Lolodorf Hills, E. Cameroon. The specimens were collected over a distance of only 25 kms in a rather hilly landscape situated between the Nyong River (the ferry south of Eseka) and the town Lolodorf. N o river was crossed within the 25 kms. Only small brooks and swamps were observed. The fauna of freshwater fishes of this part of E. Cameroon was much reduced and apart from A . cameronense only A . exiguum and Ctenopoma nanum were observed. The haploid number of the former varies from 12 to 17 in this geographical area. Outside this particular area the karyotype of the A . cameronense phenotype appears to be rather stable, with n = 17. No karyotypic differences were found during the study of two reproductively isolated populations from Rio Muni, one population from the Yaunde area of E. Cameroon and one of the populations of the Lolodorf area (LN). The marked reduction of the haploid number within the Lolodorf Hills thus probably is due to centric fusions of telocentrics. The male from the NG locality (the village Ngoyang) was heterozygous for one centric fusion (41/n+5%) of telocentrics and the corresponding trivalent was


present in primary meiotic metaphases. Somatic metaphases from the fins of the female had 2n=29 and the basic elements which had fused were strictly one-armed. Such fusions of telocentrics recently were discussed by JOHN and HEWITT (1968) who pointed out that such fusions have taken place in the Orthoptera. The conditions in the LW strain from the western outshirts of Lolodorf were similar. The male was homozygous (n=12), but the female was heterozygous with 2n=25 (pl. I. D). Nine different metaphases from this female were studied and all were similar. A fusion of two telocentrics (5 % and 6 %) was present in the heterozygous condition. The karyotypic situation within the populations of A . cameronense of the Lolodorf Hills probably reflect a labile condition after a recent invasion in the absence of severe competition. The karyotypic differences cannot, however, be explained by a simple pattern of centric fusions. Translocations and pericentric inversions probably also took place.

The probable evolution of the YA karyotype of Aphyosemion bivittatum on the basis of the NS and NN karyotypes has been figured out (fig. 4). Only 12 kms north of the YA locality the GI population inhabits a small river of the Wuri drainage. The GI karyotype resembles the YA karyotype by the low number of chromosome arms (A = 19) and by the index. The GI karyotype has two large two-armed elements and the corresponding elements are telocentric in the YA karyotype. The expected two trivalents were present in the primary meiotic metaphases of the hybrid males which exhibited a normal spermatogenesis. The GI karyotype probably originated from the YA karyotype after two centric fusions of telocentrics.

C. T h e hypothetical evolution of the rivuline Karyotype

The basic rivuline karyotype probably had 25-27 haploid two-armed chromosomes with median or submedian centromeres. The chromosomes graduated evenly in length and the shortest element was about half the length of the longest.

Examples: Aplocheilus dageti (n =25, 46 arms) (SCHEEL 1968a, Spermatogenial metaphase in p. 302). The karyotype of Pachypanchax playfairi n = 24, 48 arms, pl. 11. A) and of Rivulus rnilesi (n=23, 46 arms, pl. 11. C) are very similar but have less chromosomes.

I. The first generation of one-armed elements

During the initial evolution of the rivuline karyotype the centromeres were replaced along the axis of the chromosomes by pericentric inversions. Accidentally the chromosomes became terminal centromeres and the number of chromosome arms decreased. An irreversible condition was probably established in the one-armed elements. The replacement for the centromeres in different chromosomes was independent and karyotypes exhibiting a mixture of chromosomes with median, submedian, subterminal and terminal centromeres developed.

Examples: Aplocheilus dayi (n=24, 43 arms, pl. I. E), and A . grahami (n=24, 25 arms, pl. 11. B) represent two extremes.

In some non-rivuline, atheriniform karyotypes, i. e. the family Poeciliidae, Oryzias melastigma, Melanotaenia macculochi, Aplocheilichthys rnacrophthalmus, and A. normani (pl. 11. D) all chromosomes became telocentric (table 4).

2. The second generation of two-armed elements

Centric fusions of telocentrics (and subtelocentrics?) took place in some karyotypes but not in others. The second generation of two-armed elements developed and the haploid number decreased. In most cases the derived two-armed elements were symmetrical

204 J. J. Scheel

(rables 1-4) and were composed of arms of supermedian length in the genus Aplochei- /us (table 1). The relative length of the longest chromosome arm was maintained in (almost) all the karyotypes of the tables 1 , 2 and 4.

Examples: Aphyosemion toddi (n=23, 43 arms, pl. 11. E), Austrofundulus transilis (n=22, 40 arms, pl. 111. A), Aplocheilus fasciolatus (n=20, 38 arms, pl. 111. B), Riwu- /us ornatus (n=2O, 33 arms, pl. 111. C), and Aphyosemion petersii (n=20, 28 arms, pl. 111. D).

3. The second generation of one-armed elements

Pericentric inversions replaced the centromeres of the two-armed elements of the second generation frcm a median towards a terminal position and chromosome arms which markedly exceeded the relative length of the longest element of the basic karyotype (oversized arms) developed. The corresponding increase of the diameter of the cells probably took place in advance (SCHEEL 1968a, p. 62-64). Accidentally the centromere of one or more of the derived two-armed elements came into a terminal position and the number of chromosome arms decreased. The second generation of telocentrics had developed.

Examples: Aphyosemion exiguum (n=18, 36 arms, index 1.2 or 6l/2%;, pl. 111. E), Nothobranchius thierryi (2n=43, 24 arms, index 1.6 or 8l/a%;, pl. IV. A), Aphyosemion cinnamomeum (n=20, 30 arms, index 2.0 or 101/~%, pl. IV. B) and Aphyosemion gulare (n=16, 16 arms, index 2.0 or 11%, pl. IV. C).

4 . The third generation of two-armed elements

Oversized telocentrics (or subtelocentrics?) fused and the third generation of two- armed elements developed. In some phenotypes symmetrical karyotypes were reestab- lished on the basis of asymmetrical complements. Examples:

Symmetrical karyotypes: Aphyosemion calliurum KI (n= 11, 22 arms, R1: 2.2, pl. IV. D), A. calliurum MO (n= 10, 20 arms R1: 2.1, pl. IV. E), and A. christyi (n=9, 18 arms, R1: 2.6, pl. V. C.). Maximum of symmetry was reached in A . louessense (n=10, 20 arms, R1: 1.7 spermatogonial metaphase in SCHEEL 1966, pl. VI, fig. 2c).

Asymmetrical karyotypes: Aphyosemion gardneri EG (2n =27, 28 arms, R1: 6.7, pl. V. E), A. cognatum (2n=29, 18 arms, R1: 4.5, pl. V. B), A. a f f . cognatum NS. 4 (n=15, 25 arms, R1: 4.8, pl. V. D), and A,seyrnouri (n=14, 24 arms, R1:4.2, p1.V.A).

5 . The third generation of one-armed elements

The centromeres of the two-armed elements of the third generation were replaced from a median towards a terminal position during increases of the oversized chromosome arms. Accidentally the centromeres came into a terminal position and the number of chromosome arms decreased. The telocentrics of the third generation had developed.

Example: Pterolebias longipinnis (n = 10, 10 arms, index 3.3, R1=2.9, pl. V. F). The longest telocentric (1 7l/z%) undoubtedly contains four or more basic elements.

Two-armed elements which might be classified as belonging to the fourth generation of such elements have not yet been disclosed in rivuline karyotypes. The Pt. longipinnis organism is able, however, to handle huge telocentrics during anaphases. Conse- quently fusions may occur, reducing the haploid number to five and concentrating in the largest element (1 5 + 17l/r %) one third of the chromosome material. Theoretically the haploid number may be reduced to three by further reorganization of the material. In this way the haploid number of Pt. longipinnis may be reduced to one-ninth of the number of the closely related Pt. peruensis, which maintains the highest haploid number known in the Rivulinae.


D. Additional aspects

The hypothesis concerning the evolution of the rivuline karyotype postulates that only two different types of chromosome mutations, i. e. pericentric inversions which transformed two-armed chromosomes into telocentrics and centric fusions which produced derived two-armed elements on the basis of telocentrics (and subtelocentrics?) were common. Under this simple system similar karyotypes may develop independently in different lines of descent (parallel evolution of the karyotypes).

Within the genus Aphyosemion symmetrical and very similar karyotypes with much reduced haploid numbers were disclosed in seven populations (table 3: CAL-KI, CAL- CO, CAL-MO, CEL, CHR, FRA and LOU). Asymmetrical karyotypes with higher haploid numbers were disclosed in one of these phenotypes (the other populations of CAL). Except for the karyotypes of CAL-KI (n=11) and CAL-CO (n=lO) which may be separated only by three chromosome mutations (two pericentric inversions and one centric fusion) the other karyotypes probably developed independently from asymmetrical karyotypes.

Because of the apparent disagreement between the karyotypes of the CAL-CO and CAL-MO a crossing was prepared. The hybrids were very feeble and only two males could be raised to sexual maturity. No sperm cells were present in live gonad tissue and spermatogenesis was arrested at the first meiotic metaphase. The chromosomes decomposed at this phase. In metaphases which had not started decomposition no ringbivalents were present and in the most distinct divisions one long chain of pairing chromosomes was present. The chiasmata were quite distinct and probably the twenty arms of each of the karyotypes were quite similar. The combination of these arms was different in the two karyotypes. The KI strain ( n = l l , symmetrical karyotype) and the SA strain (2n=35 in the specimen used, asymmetrical karyotype) of A. calliurum were also crossed. The hybrids were feeble, but four males and three females could be raised to sexual maturity. The gonads of the females were normal and contained many eggs. Those of the males contained a normal concentration of fast moving, apparently normal, spermatozoa and the spermatogenesis appeared to be normal, except for the frequent occurence of small univalents in primary meiotic metaphases. These two crossing experiments demonstrated that the CO and the MO strains were not more closely related than the KI and the SA strains, in spite of the similarities of the karyotypes. The CO and the KO karyotypes probably developed independently on the basis of one of the asymmetrical karyotypes of Cameroonian A. calliurum.

The karyotypes of Adinia rnultifasciatus (n= 16, 32 arms) o i the subfamily Fundu- linae and of Aplocheilus fasciolatus (n = 18, 38 arms) of the subfamily Rivulinae, are very similar and are composed of a mixture of two-armed elements of the first and the second generation. These two karyotypes undoubtedly developed their characteris- tics independently by an evolution which favoured centric fusions.

The karyotypes of most phenotypes of the family Poeciliidae, of Dermogenys pusillus, Oryzias melastigma, Fundulus confluentus, Aplocheilichthys normani, and Aplocheilus grahami are very similar. All have 24 haploid elements and all or most are telocentric. These phenotypes belong, however, to different lines of descent of atheriniform and cyprinodontoid fishes, and the similarities of their karyotypes are probably due to a parallel evolution. The close affinities of the karyotypes of Poecilia reticulatus of the suborder Cyprinodontoidei and Melanotaenia macculochi of the suborder Athe- rinoidei have been mentioned. Karyotypes which resemble the assumed basic rivuline karyotype are present also in the suborder last mentioned, i. e. Bedotia geayi (n=24, 36 arms, R1=1.6) and Telmatherina ladigesi (n=24, 43 arms, R1=1.6).

OHNO (1970) published the photographic presentation of three n =24, 24 arms

206 /. /. Scheel

karyotypes representing three different orders of teleosts (Heterosomata, Percomor- phi and Zsospondyli). Probably all these n = 24, 24 arms karyotypes developed independently from n = 24,48 arms karyotypes.

E. Karyotype and speciation

A close study of the many karyotypes of the genus Aphyosemion (table 3) shows that karyotypes having about twenty haploid chromosomes are common in this taxon. Furthermore, one karyotype having n=19-21 was disclosed in almost all the major phenotypes. In eleven of these n = 20 was found.

These n=20 karyotypes differ, however, markedly by the number of chromosome arms (A=20-36), by the index (1=1.1-2.0), and by the relative length of the longest element (7l/~%-l 11/~%). Four of these karyotypes, representing extremes, are shown in fig. 3.

These findings indicate that the differentiation which produced the major phenotypes of this genus took place by an almost fixed haploid number and was characterized by many pericentric inversions and by some unequal translocations. In some of the widespread major phenotypes a profound reorganization of the karyotype subsequently took place in some populations, but not in others. Within the geographically rather limited western lowlands of East Cameroon the karyotype of A . calliurum exhibits a remarkable evolution with n = 10-20 and A=20-31 (fig. 4).

F. Evidence from other sources

STEBBINS (1958, referred in JOHN and LEWIS 1968, p. 175) studied the chromosome conditions in the tribe Cichorieae and argued that in plants three types of karyotypic changes produced linkage (of genes): a. reduction of the haploid number; b. change from a median to a subterminal position of centromeres; c. a trend towards increasing size differences between members of the same complement. The net result of these changes is to convert a symmetrical karyotype (chromosomes all metacentric and about the same size) into asymmetrical ones (many chromosomes with subterminal centromeres and with a wide range of size). He argued that in the Cichorieae the original members of the tribe had x = 9 and karyotypes with chromosomes of nearly equal size an all, or nearly all, with median or submedian centromeres. In many genera within the tribe increasing asymmetry appears to have begun through the occurence of pericentric inversions which converted metacentrics into acrocentrics and telocentrics. After a considerable proportion of the chromosomes had acquired subterminal centromeres a second type of asymmetry developed through the occurence and establishment of unequal translocations between non-homologous chromosomes.

The chromosome evolution observed by STEBBINS resembles that of the rivulins. The assumed basic karyotypes are similar and the transformation of the karyotype of the Cichorieae, first by pericentric inversions, later on by unequal translocations resembles the condition in the genus Aplocheilus. In other rivuline genera the unequal translocations or centric fusions tcok place at a phase when the majority of the basic elements still were metacentric or submetacentric. Even within the genus Aplocheilus (table 1) in which most karyotypes with two-armed chromosomes of the second generation occur together with a total or almost total of telocentric basic chromosomes the transformation of the latter into the one-armed condition may indeed have taken place subsequently to the centric fusions.

The present author (1971b) described an evolutionary system similar to that of the rivulins within the Anura. Eight basic and two-armed chromosomes of the common (n= 13) karyotype of Rana appear to have carried through an evolution which result-


ed in the production of two very large metacentrics in the karyotype of the African frog Arthroleptis with n = 7.

LOWE et al. (1970) studied the chromosomes of the reptile genus Cnemidophorus. They concluded that the karyotype of C. deppei with 26 haploid and telocentric chromosomes represented the ancestral karyotype which gave rise to the karyotypes of the other species groups. The karyotypes of the other species groups are characterized by the presence of one or more large two-armed elements and reduced haploid numbers, corresponding to the number of derived two-armed elements. The karyotypes with reduced haploid numbers have, however, markedly more chromosome arms than the assumed ancestral karyotype. There seems to be 38 arms in the C. tigris and C. tesselatus groups, 28-39 arms in the C. sexlineatus and C . lemniscatus groups against only 26 arms in the C. deppei group. The authors considered the increase of the number of chromosome arms as a result of unequal pericentric inversions which transformed telocentric chromosomes into two-armed elements. They did not, however, discuss the problems concerning the acting of pericentric inversions in strictly one- armed elements.

The hypothetical evolution of the Cnemidophorus karyotype may be explained under the assumption of the existence in the past of an ancestral karyotype in which most of the 26 haploid elements were two-armed. The karyotype of the C. deppei group seems to represent a separate line of evolution characterized by pericentric inversions which transformed the two-armed elements into telocentrics and no centric fusion took place (phase C.2. of the hypothetical evolution of the rivuline karyotype). In the other lines of descent which produced the other species groups, one or more centric fusions took place, but some (C. sexlineatus and C. lemniscatus) or many (C. tigris and C. tesselatus) of the basic elements remained two-armed (phase C.3. of the evolution of the rivuline karyotype).

WAHRMAN et al. (1969) studied the geographical variation of the karyotype of the genus Spalax in Israel. They argued that the site of the ancestral Spalax may have been in southeastern Europe or Asia Minor and that the karyotype of this Proto-Spalax had a relatively small number of chromosomes and rather many chromosome arms. Yugo- slavia and Bulgaria: 2n = 48-56, 80-94 arms. During the southwards expansion within Israel the diploid number increased and the number of arms decreased from 2n=52 and 84 or more arms, through 2 = 54 and 82 or more arms, 2n = 56 and 82 arms, 2n = 58 and 76 arms, to 2n=60 and 76 arms. The increase of the diploid number was caused by dissociations of large metacentrics into telocentrics and the decrease of the number of arms was caused by pericentric inversions which replaced the centromeres of small two-armed elements into a terminal position. The process last mentioned resembles that of the ancestral rivuline karyotype, wheras centric fisions do not seem to have represented an important evolutionary mecanism during the evolution of the rivuline karyotype.

Acknowledgements

I am indepted to my advisors, Professor Dr. CURT KOSSWIG and Professor, Dr. BENT CHRISTEN- SEN, for advice and support throughout the study of rivuline fishes, to the Carlsberg Foun- dation for grants, which made possible three expeditions to tropical Africa, and to the Danish State Research Foundation for microscopical equipment. I also wish to thank zoologists and killi fans who kindly supplied live fishes for the present study.

Summary

A hypothesis concerning the probable evolution of the karyotype of rivuline fishes is put for- ward on the basis of a study of 127 different rivuline karyotypes. I t is assumed that the basic karyotype had 25-27 haploid metacentric chromosomes, the longest of which was twice the

208 J. J. Scheel

length of the shortesr, the other elements graduating evenly between these extremes. By pericentric inversions an increasing number of elements became telocentric and centric fusions took place in some taxcns. In this way karyotypes with almost exclusively telocentric elements resulted without any marked change in the grading of chromosome lengths. In other taxons the haploid number was reduced to 17 by centric fusions, generally of chromosomes of supra- median size, but the relative length of the longest chromosome arm of the complement did not increase beyond the length of the longest chromosome of the basic karyotype. In other, more specialized, karyotypes, Fericentric inversions acting on the large derived two-armed elements replaced centrcmeres to a more terminal position and oversized chromosome arms resulted. This increase of the chromosome arms went hand in hand with the corresponding increase in diameter of the cells of the organism. By this process some of the derived two-armed elements of the second generation of twc-armed elements became telocentric and fusions with similar elements took place, producing huge two-armed elements of the third generation of such elements. In one phenotype these huge two-armed elements became one-armed by pericentric inversions and a karyotype having ten strictly one-armed elements resulted.

Resume Caryotypes rivulines et leur Cvolution

L’Ctude de 127 caryotypes rivulines a permis d’Ctablir une hypothhse s x I’Cvolution probable du caryotype des poissons rivulines. On suppose que le premier caryotype avait 25 a 27 chro- niosomes haploi’des et metacentriques, le plus long d’entre eux ayant une langueur double de celle du plus court, la longueur des autres ClCments variant rCguli6rement entre ces deux ex- t r h e s . A la suite d’inversions piricentriques, un nombre croissant de chromosomes devinrent tklocentriques, et des fusions centriques eurent alors lieu dans quelques taxons. De cette faSon fut crCC des caryotypes composCs en grande majorit6 d’klkments tkloccntriques sans que ceci entraina des modifications sensibles sur la variation rkgulikre de la longeur des chromosomes. Dans d’autres taxons le nombre des haploi’des fut reduit, aprks fusions centriques, B 17, ceci concerne en majorid des chromosomes de taille supCrieure B la moyenne, mais la longeur relative du bras du chromosome le plus long ne dCpassa pas la longeur du plus long des chromosomes du caryotype primitif. Chez des caryotypes plus particuliers les centromkres se dC- plachrent vers une position e x t r h e par suite d’inversions pericentriques, il en rCsulta des bras de chromosomes de dimensions supkrieures. Cet allongement des bras des chromosomes se pro- duisit parall6lement B un agrandissement du diamhre des cellules de l’organisme. Par ce pro- cessus, certains klkments B deux bras appartenant a la seconde gknCration, devinrent tClocen- triques, et par fusions avec des klCments identiques de tr& grands ClCments B deux bras de la t ro i s ihe gknkration se dCvCloppirent. Dans un caryotype particulier ces t r b grands ClCments A deux bras devinrent tklocentriques aprks inversion pbricentriques.

Zusarnrnenfassung Karyotypen der Rivulinae und ihre Evolution

Eine vergleichende Untersuchung von 127 verschiedenen Karyotypen rivuliner Zahnkarpfen wurde durchgefiihrt. Dieses Material dient der Unterstiitzung einer Hypothese iiber die mut- madliche Evolution der Karyotypen dieser Fische. Es wird angenommen, dad die Basiszahl des haploiden Satzes dieser Fische aus 25-27 metacentrischen Chromosomen bestand; dabei diirfte das langste Element etwa doppelt so grod wie das kleinste gewesen sein, die iibrigen Chromo- somen des haploiden Satzes diirften zwischen diesen beiden Extremen eine graduierende Reihe gebildet haben. Durch perizentrische Inversionen wurde eine grodere Zahl dieser urspriinglich metazentrischen Chromosomen telozentrisch. Durch zentrische Fusionen dieser telozentrischen Elemente wurde die Zahl der Chromosomen in einigen Taxa vermindert. Bei der Umwandlung metazentrischer in telozentrische Chromosomen wird die Lange derselben nicht bemerkenswert verandert. Wo in bestimmten Taxa durch zentrische Fusionen die Haploidzahl auf 17 vermindert wurde, bercht dies auf der Vereinigung von Chromosomen von mehr als mittlerer GroBe, aber so, dad die relative Lan e des langsten Chromosenarms im Komplement nicht mehr wird als die Lange des langstens &ornosoms des urspriinglichen Satzes. In anderen, hoher speziali- sierten Karyotypen wurden die sekundar entstandenen, zweiarmigen Elemente durch erneute perizentrische Inversionen verandert, indem die Centromeren eine mehr terminale Lage bezogen. Die Folge sind uberlange Chromosomenarme. Die Verlangerung bestimmter Chromosomenarme ging Hand in Hand mit einer Vergroderung des Zelldurchmessers. Wahrend dieses Prozesses wurden einige der sekundar fusionierten Chromosomen schliedlich telozentrisch und anschlie- Bend konnten erneut Fusionen erfolgen, die zu besonders groden Elementen fiihrten. In einem Fall wurden infolge perizentrischer Inversionen diese groden Elemente einarmig, so daii die haploide Chromosomengarnitur nur noch aus 10 streng einarmigen Elementen besteht.

Rivufine Knryotypes and their Ecofution 209

Literature

CEHN, T. R.; RUDDLE, F. H., 1970: A chromosome study of four species and a hybrid of the

EBELING, A. W.; CHEN, T. R., 1970: Heterogamety in Teleostean Fishes. Trans. Amer. Fish.

GOLDSTEIN, R. J.; RICCO, J. F., 1970: A description of a new species of Aphyosemion, with

JOHN, R.; HEWITT, G. M., 1968: Patterns and Pathways of Chromosome Evolution within the

TOHN. R.: LEWIS. K. R.. 1968: The Chromosome Complement. Handbuch der Protoplasma-

killifish genus Fundufus (Cyprinodontidae). Chromosoma (Berl.), 29, 255- 267.

SOC. 99, 131-138.

notes on its maintenance in aquaria. J. Amer. Killifish Ass. 7, 8-11.

Orthoptera. Chromosoma (Berl.), 25, 40-74. “

foisch.’ 6. A. 266. pp. LOISELLE, P. V.; BLAIR, D., 1972: A new species of Aphyosemion (Tefeostomi: Cyprinodon-

tidae: Rivufinae) from Ghana, and a redefinition of subgenus Fundufopanchax Myers, 1924. J. Amer. Killifish Ass. 8, 1-11.

LOWE, C. H.; WRIGHT, J. W.; COLE, C. J.; BEZY, R. L., 1970: Chromosomes and evolution of the species groups of Cnemidophorus (Reptifia, Teiidae). Syst. Zool. 19, 128-141.

MATTHEY, R., 1966: Une inversion pbricentrique a l’origine d’un polymorphisme chromosomi- clue non-robertsonien dans une population de Mastomys (Rodentia, Murinae). Chromosoma . - - . (Bed), 18, 188-200.

MURAMOTO. 1.: OHNO. S.: ATKIN. N. B., 1968: On the diploid state of the fish order Osta- riophysi: Chromosoma (Berl.), 24, 59-66.

MYERS. G. S.. 1955: Notes on the Classification and Names of Cyprinodont Fishes. The Trop- . _ ical Fish Magazine, March 1955.

OHNO. S.. 1970: The Enormous Diversity of Genome Sizes of Fish as a Reflection of Nature’s Extensive Experiments with Gene D;plications. Trans. Amer. Fish. SOC. 99, 120-130.

OHNO, S.; ATKIN, N. B., 1966: Comparative DNA values and chromosome complements of eight species of fishes. Chromosoma (Bed.), 18, 455-466.

OHNO, S.; STENIUS, C.; WEILER, C.; POOLE, J.; CHRISTIAN, L., 1966: Autosomal polymorphism due to pericentric inversions in the deer mouse (Peromyscus manicufatus), and some evidence of somatic segregation. Chromosoma (Berl.), 18, 177-187.

OHNO, S.; MURAMOTO, J.; CHRISTIAN, L.; ATKIN, N. B., 1967: Diploid-Tetraploid relationship among Old World members of the family Cyprinidae. Chromosoma (Berl.), 23, 1-9.

POST, A., 1965 : Vergleichende Untersuchungen der Chromosomenzahlen bei SuBwasser-Tele- ostern. Zeitschr. Syst. Evol. Forsch. 3, 47-93.

RADDA, A., 1970a: Vorlaufige Beschreibung zweier neuer Unterarten einer neuen Art der Gat- tung Aphyosemion Myers aus Westkamerun. Aquaria 4, 58-61.

- 1970b: Vorlaufige Beschreibung von Aphyosemion scheefi nov. spec. Aquaria 4, 177-180. ROBERTS, F. L., 1967: Chromosome cytology of the Osteichthyes. Proc. Fish-Cult. 29, 75-83. - 1970: Atlantic Salmon (Salmo safar). Chromosomes and Speciation. Trans. Amer. Fish.

SCHEEL, J. J., 1966: Taxonomic Studies of African and Asian tooth-carps (Rivufinae) based on chromosome numbers, haemo lobin patterns, some morphological traits and crossing experiments. Vidensk. Medd. d a n k naturh. Foren. 129, 123-148. 1968a: Rivulins of the Old World. T. F. H. Publ. USA. 480 pp. 1968b: An new species of Aphyosemion (Cyprinodontidae, Pisces) from Fernando Po. Rev. Zool. Bot. Afr. 78, 332-342. 1970a: The genus Procatopus. Trop. Fish. Hobbyist. 18,4-14. 1970b: Familie Cyprinodontidae. In STERBA, G.: SuBwasserfische aus aller Welt. Teil 2, 358- 444. 1971a: Aphyosemion franzwerneri and Aphyosemion cefiae, two new rivulins from Came- roon. Trop. Fish. Hobbyist. 19, 46-66. 1971b: The seven chromosome-karyot pe of the African frog Arthroleptis, a probable

SETHI, R. P., 1960: Osteology and Phylogeny o r Oviparous Cyprinodont Fishes. (Order C y - prinodontiformes). Mic. 60-1909. Xerox. Univ. Microfilms. Ann Arbor Michigan. 274 pp.

WAHRMAN, J.; GOITEIN, R.; NEVO, E., 1969: Geogra hic variation of chromosome forms in Spalax, a subterranean mammal of restricted mobifity. Comp. Mamm. Cytogenetics. 1969,

SOC. 99, 105-111.

derivative of the thirteen chromosome- l aryot pe of Rana. Hereditas. 67, 287-290.

30-48. WHITE, M. J. D., 1954: Animal Cytology and Evolution. Cambridge Univ. Press. 454 pp.

Author’s address: J. J. SCHEEL, 95, Abrinken, DK-2830 Virum

Documents

Rivuline Karyotypes and their Evolution (Rivulinae, Cyprinodontidae, Pisces)