<%BANNER%>

Systematics of the African Catfish Family Amphiliidae (Teleostei

Permanent Link: http://ufdc.ufl.edu/UFE0045204/00001

Material Information

Title: Systematics of the African Catfish Family Amphiliidae (Teleostei Siluriformes).
Physical Description: 1 online resource (214 p.)
Language: english
Creator: Thomson, Alfred William
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2013

Subjects

Subjects / Keywords: amphiliidae -- amphilius -- anoplopterus -- catfish -- phylogeny -- siluriformes -- systematics -- taxonomy -- teleostei
Biology -- Dissertations, Academic -- UF
Genre: Zoology thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: The phylogenetic relationships of Amphiliidae are studied using Bayesian and Maximum Likelihood methods based on data from one mitochondrial locus and two nuclear loci from 12 of the 13 recognized amphiliid genera and more than half of the known species diversity of the family. The loci are analyzed separately and as a concatenated data set. All analyses find strong support two of the three currently recognized subfamilies but the subfamily Amphiliinae is consistently recovered as polyphyletic. The three loci are then employed in Bayesian and Maximum Likelihood analyses of the largest genus in the family, Amphilius. All analyses do not support the monophyly of Amphilius but the previously recognized High African and Low African morphological groups are consistently recovered as well supported clades. The name Anoplopterus Pfeffer,1889 is resurrected for the species of the High Africa group and all thespecies of the Low Africa group are retained in the genus Amphilius.  The Amphilius jacksonii complex is revised and five new species are described: A.n. sp. Ruzizi from the Ruzizi drainage and northeastern tributaries of Lake Tanganyika; A. n. sp. Malagarasi from the Malagarasi drainage, Lake Rukwa basin, and upper Great Ruaha River system,Rufiji basin; A. n. sp. Congo from the upper Congo basin; A. n. sp. Rufiji from the Rufiji and Wami basins; and A.n. sp. Lake Kyogo from the Lake Kyogo drainage, northeastern tributaries of Lake Victoria, and the Lake Manyara basin.Taxonomic work is also conducted on the High African Amphilius. The status of four nominal species (Chimarrhoglanis leroyi Vaillant 1897, A. hargeri Boulenger 1907, A. brevidorsalis Pellegrin 1919, and A. platychir var. cubangoensis Pellegrin 1936) that have long been placed in the synonymy of A. uranoscopus is addressed and two new species are described: A. n. sp. Ruvu from theRuvu basin, and A. n. sp. Mangulafrom the Rufiji basin.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Alfred William Thomson.
Thesis: Thesis (Ph.D.)--University of Florida, 2013.
Local: Adviser: Page, Larry M.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2013
System ID: UFE0045204:00001

Permanent Link: http://ufdc.ufl.edu/UFE0045204/00001

Material Information

Title: Systematics of the African Catfish Family Amphiliidae (Teleostei Siluriformes).
Physical Description: 1 online resource (214 p.)
Language: english
Creator: Thomson, Alfred William
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2013

Subjects

Subjects / Keywords: amphiliidae -- amphilius -- anoplopterus -- catfish -- phylogeny -- siluriformes -- systematics -- taxonomy -- teleostei
Biology -- Dissertations, Academic -- UF
Genre: Zoology thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: The phylogenetic relationships of Amphiliidae are studied using Bayesian and Maximum Likelihood methods based on data from one mitochondrial locus and two nuclear loci from 12 of the 13 recognized amphiliid genera and more than half of the known species diversity of the family. The loci are analyzed separately and as a concatenated data set. All analyses find strong support two of the three currently recognized subfamilies but the subfamily Amphiliinae is consistently recovered as polyphyletic. The three loci are then employed in Bayesian and Maximum Likelihood analyses of the largest genus in the family, Amphilius. All analyses do not support the monophyly of Amphilius but the previously recognized High African and Low African morphological groups are consistently recovered as well supported clades. The name Anoplopterus Pfeffer,1889 is resurrected for the species of the High Africa group and all thespecies of the Low Africa group are retained in the genus Amphilius.  The Amphilius jacksonii complex is revised and five new species are described: A.n. sp. Ruzizi from the Ruzizi drainage and northeastern tributaries of Lake Tanganyika; A. n. sp. Malagarasi from the Malagarasi drainage, Lake Rukwa basin, and upper Great Ruaha River system,Rufiji basin; A. n. sp. Congo from the upper Congo basin; A. n. sp. Rufiji from the Rufiji and Wami basins; and A.n. sp. Lake Kyogo from the Lake Kyogo drainage, northeastern tributaries of Lake Victoria, and the Lake Manyara basin.Taxonomic work is also conducted on the High African Amphilius. The status of four nominal species (Chimarrhoglanis leroyi Vaillant 1897, A. hargeri Boulenger 1907, A. brevidorsalis Pellegrin 1919, and A. platychir var. cubangoensis Pellegrin 1936) that have long been placed in the synonymy of A. uranoscopus is addressed and two new species are described: A. n. sp. Ruvu from theRuvu basin, and A. n. sp. Mangulafrom the Rufiji basin.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Alfred William Thomson.
Thesis: Thesis (Ph.D.)--University of Florida, 2013.
Local: Adviser: Page, Larry M.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2013
System ID: UFE0045204:00001


This item has the following downloads:


Full Text

PAGE 1

1 SYSTEMATICS OF THE AFRICAN CATFISH FAMILY AMPHIL IIDAE (TELEOSTEI: SILURIFORMES) By ALFRED WILLIAM THOMSON A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREM ENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2013

PAGE 2

2 2013 Alfred William Thomson

PAGE 3

3 To my parents, Alfred S. and Gery L. Thomson, who have always been there for me and have always suppor ted me

PAGE 4

4 ACKNOWLEDGMENTS I thank my advisor, Lawrence M. Page, as well as my committee members, Jon athan W. Armbruster, James D. Austin, Rebecca Kimball, and Colette M. St. Mary Funding for this study was provided by the All Catfish Species Inventory project funded by the U.S. National Science Foundation (DEB 0315963). For t issue gifts, specimen loans and access to institutional specimens I thank Barbara A. Brown and Melanie L. J. Stiassny of the American Museum of Natural History (AMNH), Jonathan Ar mbruster and David Werneke of the Auburn University Museum Fish Collection (AUM), Ralf Britz and James Maclain of the Natural History Museum (BMNH), William N. Eschmeyer, T omio Iwamoto and David Catania of the California Academy of Sciences (CAS), John Fri el and John P. Sullivan of the Cornell University Museum of Vertebrates (CU), Mary A. Rogers of the Field Museum of Natrual History (FMNH), Tom Geerinckx of Ghent University, Belgium, Andrew Bentley of the University of Kansas Natural History Museum and Bi odiversity Research Center (KU), Karsten Hartel of the Harvard Museum of Comparative Zoology (MCZ), Guy Duhamel and Patrice Pruvost of the Musum National d'Histoire naturelle (MNHN), Jos Snoeks, Mark Hanssen s and Miguel Parrent of the Royal Museum of Cent ral Africa (MRAC), Paul Skelton, Sherwyn C. Mack and Roger Bills of the South African Institute of Aquatic Biodiversity (SAIAB), Ray Schmidt of Tulane University (TU), Moritz Muschick and Walter Salzburger of Universitt Basel Switzerland, Robert H. Robin s of the Florida Museum of Natural History (UF), Aventino Kasangaki of Uganda, Jeffery T. Williams of the Smithsonian Institution National Museum of Natural History (USNM), and Dirk Neumann of the Zoologische Staatssammlung Mnchen (ZSM).

PAGE 5

5 TABLE OF CONTENT S page ACKNOWLEDGMENTS .................................................................................................. 4 LIST OF TABLES ............................................................................................................ 9 LIST OF FIGURES ........................................................................................................ 11 ABSTRACT ................................................................................................................... 14 CHAPTER 1 INTRODUCTION .................................................................................................... 16 2 MOLECULAR PHYLOGENY OF THE AFRICAN CATFISH FAMILY AMPHILIIDAE (TELEOSTEI: SILURIFORMES). .................................................... 19 Background ............................................................................................................. 19 Material and Methods ............................................................................................. 21 Acquisition of Specimens ................................................................................. 2 1 DNA Extraction, PCR amplification, and Sequencing ....................................... 22 Gene Tree Reconstruction ............................................................................... 23 Results .................................................................................................................... 24 Characteristics of the Sequence Data and Models of Sequence Evolution ...... 24 Phylogenetic Relationships: cyt b ..................................................................... 26 Ph ylogenetic Relationships: Rag 2 ................................................................... 27 Phylogenetic Relationships: S7 Intron 1 ........................................................... 28 Phylogenetic relationships: cyt b + Rag 2 + S7 Intron 1 ................................... 29 Discussion .............................................................................................................. 29 3 MOLECULAR PHYLOGENY OF THE AFRICAN CATFISH GENUS AMPHILIUS (SILURIFORMES: AMPHILIIDAE). ......................................................................... 46 Background ............................................................................................................. 46 Material and Methods ............................................................................................. 47 Results .................................................................................................................... 48 Characteristics of the Sequence Data and Models of Sequence Evolution ...... 48 Phylogenetic Relationships: cyt b ..................................................................... 49 P hylogenetic Relationships: Rag 2 ................................................................... 49 Phylogenetic Relationships: S7 Intron 1 ........................................................... 50 Phylogenetic Relationships: cyt b + Rag 2 + S7 Intron 1 .................................. 51 Discussion .............................................................................................................. 51 Taxonomic Descriptions .......................................................................................... 57 Amphilius Gnther, 1864 .................................................................................. 57 Diagnosis (based, in part, on Roberts, 2003 and Skelton, 2007b) ............. 57 Species included ........................................................................................ 57

PAGE 6

6 Distribution ................................................................................................. 58 Notes ......................................................................................................... 58 Anoplopterus Pfeffer, 1889 ............................................................................... 58 Diagnosis (based, in part, on Roberts, 2003 and Skelton, 2007b) ............. 58 Species included ........................................................................................ 59 Distribution ................................................................................................. 59 Notes ......................................................................................................... 59 4 TAXONOMY OF THE AFRICAN CATFISH GENUS ANOPLOPTERUS (SILURIFORMES: AMPHILIIDAE) WITH DESCRIPTIONS OF TWO NEW SPECIES AND REDESCRIPTIONS OF ANOPLOPTERUS URANOSCOPUS CHIMARRHOGLANIS LEROYI AMPHILIUS BREVIDORSALIS, AMPHILIUS HARGERI AND AMPHILIUS CUBANGOENSIS. .................................................... 79 Background ............................................................................................................. 79 M aterial and Methods ............................................................................................. 80 Taxonomic Descriptions .......................................................................................... 82 Anoplopterus uranoscopus Pfeffer, 1889 (Figure 42; Table 41) ..................... 82 Material Examined ..................................................................................... 82 Diagnosis ................................................................................................... 83 Description ................................................................................................. 84 Coloration ................................................................................................... 86 Distribution ................................................................................................. 86 Anoplopterus leroyi (Vaillant, 1897) (Figure 44; Table 44) ............................. 86 Material Examined ..................................................................................... 87 Diagnosis ................................................................................................... 87 Description ................................................................................................. 89 Coloration ................................................................................................... 91 Distribution ................................................................................................. 91 Anoplopterus hargeri (Boulenger, 1907) (Figure 45; Table 45) ...................... 91 Material examined ...................................................................................... 91 Diagnosis ................................................................................................... 92 Description ................................................................................................. 93 Coloration ................................................................................................... 95 Distribution ................................................................................................. 95 Anoplopterus brevidorsalis (Pellegrin, 1919) (Figure 47; Table 46) ............... 96 Material examined ...................................................................................... 96 Diagnosis ................................................................................................... 97 Description ................................................................................................. 98 Coloration ................................................................................................. 100 Distribution ............................................................................................... 100 Anoplopterus cubangoensis (Pellegrin, 1936) (Figure 48; Table 47) ........... 100 Material examined .................................................................................... 101 Diagnosis ................................................................................................. 104 Description ............................................................................................... 105 Coloration ................................................................................................. 107 Distribution ............................................................................................... 108

PAGE 7

7 Anoplopterus n. sp. Mangula new species ( Figure 49; Table 48) ................ 108 Non types ................................................................................................. 108 Diagnosis ................................................................................................. 108 Description ............................................................................................... 110 Coloration ................................................................................................. 111 Distribution ............................................................................................... 112 Anoplopterus n. sp. Ruvu new species (Figure 410; Table 49) .................... 112 Diagnosis ................................................................................................. 112 Description ............................................................................................... 113 Coloration ................................................................................................. 115 Distribution ............................................................................................... 115 Discussion ............................................................................................................ 116 Comparative Material Examined ........................................................................... 121 5 REVISION OF THE AMPHILIUS JACKSONII COMPLEX (SILURIFORMES: AMPHILIIDAE), WITH THE DESCRIPTION OF FIVE NEW SPECIES. ................ 147 Background ........................................................................................................... 147 Materials and Methods .......................................................................................... 148 Taxonomic Descriptions ........................................................................................ 150 Amphilius jac ksonii complex ........................................................................... 150 Amphilius jacksonii Boulenger 1912 (Figure 5 1; Table 51) .......................... 151 Material examined .................................................................................... 151 Diagnosis ................................................................................................. 154 Description ............................................................................................... 154 Coloration ................................................................................................. 156 Distribution ............................................................................................... 157 Amphilius n. sp. Ruzizi (Figure 5 3; Table 53) ............................................... 157 Non types ................................................................................................. 158 Diagnosis ................................................................................................. 158 Description ............................................................................................... 159 Coloration ................................................................................................. 161 Dis tribution ............................................................................................... 161 Amphilius n. sp. Malagarasi (Figure 5 4; Table 54) ....................................... 161 Non types ................................................................................................. 162 Diagnosis ................................................................................................. 166 Description ............................................................................................... 166 Coloration ................................................................................................. 168 Distributio n ............................................................................................... 169 Amphilius n. sp. Congo .................................................................................. 169 Non types ................................................................................................. 169 Diagnosis ................................................................................................. 171 Description ............................................................................................... 172 Coloration ................................................................................................. 174 Distribution ............................................................................................... 174 Amphilius n. sp. Rufiji (Figure 56; Table 56) ................................................ 175 Non types ................................................................................................. 175

PAGE 8

8 Diagnosis ................................................................................................. 176 Description ............................................................................................... 176 Coloration ................................................................................................. 178 Distribution ............................................................................................... 179 Amphilius n. sp. Lake Kyogo (Figure 5 7; Tables 57) .................................... 179 Non types ................................................................................................. 179 Diagnosis ................................................................................................. 180 Description ............................................................................................... 180 Coloration ................................................................................................. 182 Distribution ............................................................................................... 183 Discussion ............................................................................................................ 183 Comparative Material Examined ........................................................................... 186 6 CONCLUSIONS ................................................................................................... 203 LIST OF REFERENCES ............................................................................................. 206 BIOGRAPHICAL SKETCH .......................................................................................... 214

PAGE 9

9 LIST OF TABLES Table page 2 1 Locality, museum catalogue numbers, tissue numbers, and GenBank accession numbers for specimens used in the molecular analyses. Bold: Sequences from Genbank. CAR = Central African Republic; DRC = Democratic Republic of the Congo. .................................................................... 33 3 1 Locality, museum catalogue numbers, tissue numbers, and GenBank accession numbers for specimens used in the molecular analyses. Bold: Sequences from Genbank. CAR = Central African Republic; DRC = Democratic Republic of the Congo. .................................................................... 60 4 1 Morphometric data for Anoplopterus uranoscopus. .......................................... 125 4 2 Meristic traits, caudal fin shape, and coloration diag nostic for species of the Anoplopterus ................................................................................................... 126 4 3 Morphometric traits diagnostic for species of the Anoplopterus ....................... 128 4 4 Morphometric data for Anoplopterus leroyi Range and mean include the holotype. ........................................................................................................... 129 4 5 Morphometric data for Anoplopterus hargeri Range and mean include the holotype. ........................................................................................................... 130 4 6 Morphometric data for Anoplopterus brevidorsalis Range and mean include the holotype. ..................................................................................................... 131 4 7 Morphometric data for Anoplopterus cubangoensis Range and mean include the holotype. ..................................................................................................... 132 4 8 Morphometric data for Anoplopterus n. sp. Mangula. Range and mean include the holotype. ......................................................................................... 133 4 9 Morphometric data for Anoplopterus n. sp. Ruvu Range and mean include the holotype. ..................................................................................................... 134 4 10 Total vertebrae counts in species of the Anoplopterus .................................... 135 4 11 1st dorsal pterygiophore intercept counts in species of Anoplopterus ............. 135 4 12 Total gill raker counts in species of Anoplopterus ............................................ 136 4 13 Branched anal fin ray counts in species of Anoplopterus ................................ 136 5 1 Morphometric data for Amphilius jacksonii Range and mean include the holotype. ........................................................................................................... 188

PAGE 10

10 5 2 External traits diagnostic for species of the Amphilius jacksonii complex. ........ 189 5 3 Morphometric data for Amphilius n. sp. Ruzizi Range and mean include the holotype. ........................................................................................................... 190 5 4 Morphometric data for Amphilius n. sp. Malagarasi Range and mean include the holotype. ..................................................................................................... 191 5 5 Morphometric data for Amphilius n. sp. Malagarasi Range and mean include the holotype. ..................................................................................................... 192 5 6 Morphometric data for Amphilius n. sp. Rufiji. Range and mean include the holotype. ........................................................................................................... 193 5 7 Morphometric data for Amphilius n. sp. Lake Kyogo. Range and mean include the holotype. ......................................................................................... 194 5 8 Branchiostegal ray counts in A. jacksonii, A. n. s p. Ruzizi A. n. sp. Malagarasi A. n. sp. Congo, A. n. sp. Rufiji, and A. n. sp. Lake Kyogo ........... 195 5 9 Total gill raker counts in A. jacksonii, A. n. sp. Ruzizi A. n. sp. Malagarasi A. n. sp. Co ngo A. n. sp. Ruzizi and A. n. sp. Lake Kyogo ................................. 195 5 10 Branched pectoral fin ray counts in A. jacksonii, A. n. sp. Ruzizi A. n. sp. Malagarasi A. n. sp. Congo, A. n. sp. Ruzizi and A. n. sp Lake Kyogo .......... 195 5 11 Branched anal fin ray A. jacksonii, A. n. sp. Ruzizi A. n. sp. Malagarasi A. n. sp. Congo, A. n. sp. Ruzizi and A. n. sp. Lake Kyogo ..................................... 195

PAGE 11

11 LIST OF FIGURES Figure page 2 1 Majority rule consensus tree from the Bayesian analysis of the mitochondrial cyt b locus.. ......................................................................................................... 38 2 2 Majority rule consensus tree from the likelihood analysis of the mitochondrial cyt b locus.. ......................................................................................................... 39 2 3 Majority rule consensus tree from the Bayesian analysis of the nuclear Rag 2 gene. ................................................................................................................. 40 2 4 Majority rule consensus tree from the likelihood analyses of the nuclear Rag 2 gene.. .............................................................................................................. 41 2 5 Majority rule consensus tree f rom the Bayesian analysis of the nuclear S7 intron.. ................................................................................................................ 42 2 6 Majority rule consensus tree from the likelihood analysis of the nuclear S7 intron.. ................................................................................................................ 43 2 7 Majority rule consensus tree from the Bayesian analysis of the concatenated cyt b, Rag 2, and S7 data. .................................................................................. 44 2 8 Majority rule consensus tree from the likelihood analysis of the concatenated cyt b, Rag 2, and S7 data. .................................................................................. 45 3 1 Majority rule consensus tree from the Bayesian analysis of the mitochondrial cyt b locus. .......................................................................................................... 67 3 2 Majority rule consensus tree from the likelihood analysis of the mitochondrial cyt b locus. .......................................................................................................... 68 3 3 Majority rule consensus tree from the Bayesian analysis of the nuclear Rag 2 gene. .................................................................................................................. 69 3 4 Majority rule consensus tree of the from the Bayesian analysis of the nuclear Rag 2 gene showing only the Amphilius jacksonii complex. ............................... 70 3 5 Majority rule consensus tree of the from the Bayesian analysis of the nuclear Rag 2 gene showing only the high African Amphilius ........................................ 71 3 6 Majority rule consensus tree from the l ikelihood analyses of the nuclear Rag 2 gene. ............................................................................................................... 72 3 7 Majority rule consensus tree from the likelihood analyses of the nuclear Rag 2 gene showing only the Amphilius jacksonii complex.. ..................................... 73

PAGE 12

12 3 8 Majority rule consensus tree from the likelihood analyses of the nuclear Rag 2 gene showing only the high African Amphilius ................................................. 74 3 9 Ma jority rule consensus tree from the Bayesian analysis of the nuclear S7 intron. ................................................................................................................. 75 3 10 Majority rule consensus tree from the likelihood analysis of the nuclear S7 intron. ................................................................................................................. 76 3 11 Majority rule consensus tree from the Bayesian analysis of the concatenated cyt b, Rag 2, and S7 data. .................................................................................. 77 4 1 Species of the Amphilius uranoscopus group recognized by Thomson & Page (2010). ..................................................................................................... 137 4 2 Anoplopterus uranoscopus UF 170723 96.2 mm SL; lateral, dorsal and ventral view. ...................................................................................................... 138 4 3 Distribution of Anoplopterus n. sp. Ruvu A. n. sp. Mangula A. leroyi and A. uranoscopus .................................................................................................... 139 4 5 Anoplopterus hargeri SAIAB 34332 80.3 mm SL; lateral, dorsal and ventral view. ................................................................................................................. 141 4 6 Distributions of Anoplopterus brevidorsailis A. hargeri and A. cubangoensis 142 4 7 Anoplopterus brevidorsailis SAIAB 53691 XXX mm SL; lateral, dorsal and ventral view. ...................................................................................................... 143 4 8 Anoplopterus cubangoensis SAIAB 80539 XXX mm SL; lateral, dorsal and ventral view. ...................................................................................................... 144 4 9 Anoplopterus n. sp. Mangula, UF 170702 95.6 mm SL, holotype; lateral, dorsal and ventral view. .................................................................................... 145 4 10 Anoplopterus n. sp. Ruvu UF 170727 75.4 mm SL, holotype; lateral, dorsal and ventral view. ............................................................................................... 146 5 1 Amphilius jacksoni UF 110743 XXmm SL; lateral, dorsal and ventral view. .... 196 5 2 Distributions of species of the Amphilius jacksonii complex. ............................ 197 5 3 Amphilius n. sp. Ruzizi MRAC 9329493300 82.8 mm SL, Holotype; lateral, dorsal and ventral view. .................................................................................... 198 5 4 Amphilius n. sp. Malagarisi, CU 97334, 52.3 mm SL, Holotype; lateral, dorsal and ventral view. ............................................................................................... 199

PAGE 13

13 5 5 Amphilius n. sp. Congo CU 97335, 83.1 mm SL, Holotype; lateral, dorsal and ventral view. ............................................................................................... 200 5 6 Amphilius n. sp. Rufiji, UF 184237, 86.3 mm SL, Holotype; lateral, dorsal and ventral view. ...................................................................................................... 201 5 7 Amphilius n. sp. Lake Kyogo, UF 184238, 104.1 mm SL, Holotype; lateral, dorsal and ventral view. .................................................................................... 202

PAGE 14

14 Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy SYSTEMATICS OF THE AFRICAN CATFISH FAMILY AMPHILIIDAE (TELEOSTEI: SILURIFORMES) By Alfred William Thomson May 2013 Chair: Lawrence M. Page Major: Zoology The phylogenetic relationships of the A mphiliidae are studied using Bayesian and Maximum Likelihood methods based on data from one mitochondrial locus and two nuclear loci from 12 of the 13 recognized amphiliid genera and more than half of the known species diversity of the family. The loci are analyzed separately and as a concatenated dataset. All analyses find strong support for two of the three currently recognized subfamilies but the subfamily Amphiliinae is consistently recovered as polyphyletic. The three loci are then employed in Bayesian and Maximum Likelihood analyses of the largest genus in the family, Amphilius No analyses support the monophyly of Amphilius but the previously recognized High African and Low African morphological groups are consistently recovered as well supported cl ades. The name Anoplopterus Pfeffer 1889 is resurrected for the species of the High African group, and all species of the Low African group are retained in Amphilius The Amphilius jacksonii complex is revised, and five new species are described: A. n. s p. Ruzizi from the Ruzizi drainage and northeastern tributaries of Lake Tanganyika; A. n. sp. Malagarasi from the Malagarasi drainage, Lake Rukwa basin, and

PAGE 15

15 upper Great Ruaha River system, Rufiji basin; A. n. sp. Congo from the upper Congo basin; A. n. sp. Rufiji from the Rufiji and Wami basins; and A. n. sp. Lake Kyogo from the Lake Kyogo drainage, northeastern tributaries of Lake Victoria, and the Lake Manyara basin. Taxonomic work is also conducted on the High African Amphilius The status of four nomina l species, Chimarrhoglanis leroyi Vaillant 1897, A. hargeri Boulenger 1907, A. brevidorsalis Pellegrin 1919, and A. platychir var. cubangoensis Pellegrin 1936, that have long been placed in the synonymy of A. uranoscopus is addressed, and two new species are described: A. n. sp. Ruvu from the Ruvu basin, and A. n. sp. Mangula from the Rufiji basin.

PAGE 16

16 CHAPTER 1 INTRODUCTION Catfishes of the family Amphiliidae (Teleostei: Siluriformes) are small to moderatesized fishes native to small streams throughout subSaharan Africa. The phylogenetic relationships, monophyly and limits of Amphiliidae have been the subject of considerable debate. T hirteen genera are currently recognized in three subfamilies. Eighty five species are currently recognized, but a taxonomic r evision of this family is badly needed. Most of the species have very restricted distributions and show a high level of endemicity to river drainages and possibly to single river systems and rivers. The several recognized species with large distributions are thought to be species complexes with many populations worthy of taxonomic recognition. Additionally, recent collection efforts have yielded additional undescribed taxa that are awaiting formal taxonomic description. He et al. (1999) and Diogo (2003) con ducted phylogenetic analyses of the family using morphological data and reached different conclusions on the monophyly of the family. He et al. (1999) found that the genera of the subfamily Doumeinae and the genus Zaireichthys were more closely related to the Asian catfish family Sisoridae than to the genera of the subfamily Amphiliinae but Diogo (2003) found strong support for the monophyly of the family. In addition to Diogos study, support for the monophyly of Amphiliidae was found in several broader st udies on the phylogenetic relationships of the Siluriformes. Mo (1991) and Diogo (2005) both included representatives of all three recognized amphiliid subfamilies in their morphological phylogenetic studies of the Siluriformes and found strong support for the monophyly of Amphiliidae. The molecular phylogenetic analysis of the Siluriformes by Sullivan et al. (2006), which included four

PAGE 17

17 amphiliid genera including representatives of each of the three subfamilies, also found strong support for the monophyly o f Amphiliidae. All phylogenetic studies on amphiliids have been limited in taxon sam pling so the monophyly of each of the 13 recognized genera are untested and species relationships within genera are unknown. The most speciose and most widely distributed g enus in the family is Amphilius As currently recognized it includes 29 species and is distributed throughout Low Africa (northern and western Africa in which most of the land is at elevations between 500 and 1000 meters ) and High Africa (southern and east ern Africa in which most of the land is at elevations well above 1000 meters ( much above 4000 meters ). Skelton (1984) identified two characters that distinguished most of the species from Low Africa from most of the species from High Africa. The Low African species have an epidermal fold at the base of the caudal fin and 6 + 7 or 7 + 8 principal caudal fin rays while the High African species lack the epidermal fold and have 8 + 9 principal caudal fin rays The Low and High African Amphilius were later recognized as two different morphological groups by Thomson & Page (2010). Within both the Low and High African morphological groups a considerable amount of undescribed diversity exits. In the Low African group, one of the most widely distributed species is A mphilius jacksonii Amphilius jacksonii was described from Uganda but has since been recorded from throughout much of eastern Africa. There is a substantial amount morphological difference between populations from different drainages indicating the material identified as Amphilius jacksonii represents more than one species.

PAGE 18

18 Amphilius uranoscopus is the most widely distributed species in the High African group. Thomson & Page (2010) removed two nominal species from the synonymy of Amphilius uranoscopus and described a population from the Athi River Kenya as a new species. Four other nominal species that have long been placed in the synonymy of A. uranoscopus were not addressed by Thomson & Page (2010) and their status needs to be addressed before additional species of High African Amphilius can be described. In this dissertation, the phylogenetic relationships of Amphiliidae are analyzed based on data from one mitochondrial locus and two nuclear loci. Seq uence data from 12 of the 13 currently recognized gener a and more than half of the known species diversity of the family are included in the study allowing not only the monophyly of the subfamilies but also the monophyly of many of the genera to be tested. A phylogenetic analyses of Amphilius is also conducted employing the same loci but with more extensive taxon sampling. The analysis provides strong support for the monophyly of both the High and Low African morphological groups and also indicates a substantial amount of undescribed diversity. The Amphilius ja cksonii complex is revised and five new species in the complex are described. Taxonomic work is also conducted on the High African Amphilius The status of four nominal species that have long been placed in the synonymy of A. uranoscopus is addressed and t wo new species are described.

PAGE 19

19 CHAPTER 2 MOLECULAR PHYLOGENY OF THE AFRICAN CATFISH FAMILY AMPHILIIDAE (TELEOSTEI: SILURIFORMES). Background The family Amphiliidae was erected by Regan (1911), who divided it into two subfamilies The genera Amphilius and Paramphilius were placed in the subfamily Amphiliinae, and the genera Andersonia Belonoglanis Doumea, Paraphractura, Phractura and Trachyglanis were placed in the subfamily Doumeinae. David and Poll (1937) proposed that the genus Leptoglanis which was previously placed in the family Bagridae, was closely related to species of Amphiliidae and included it in the family. Harry (1953) rejected the inclusion of Leptoglanis in the family and placed Paraphractura in the synonymy of Phractura. Bailey and Stewart (1984) transferred Leptoglanis back to Amphiliidae and included Zaireichthys a genus described by Roberts (1967) and thought to be closely related to Leptoglanis The inclusion of Leptoglanis and Zaireichthys in Amphiliidae was strongly supported in the phylogenetic analysis of Siluriformes by Mo (1991) who listed 14 derived features shared by these two genera and all or some amphiliids. Additionally, Mo (1991) identified two synapomorphies for the family Amphiliidae: two prominent spinelike processes o n the posterior palatine, and presence of a hook like dorsal laminar process on the pelvic girdle. He et al. (1999) conducted a phylogenetic analysis that included Zaireichthys ( Leptoglanis camerunensis = Zaireichthys camerunensis ) but not Leptoglanis and concluded that Amphiliidae was not monophyletic. Their results indicated that the genera of the subfamily Doumeinae and Zaireichthys were more closely related to the sisorid genera Euchiloglanis and Glyptothorax and should be

PAGE 20

20 recognized in their own famil y, Doumeidae, with Amphiliidae being restricted to Amphilius and Paramphilius Diogo (2003, 2005) proposed relationships among the genera of Amphiliidae that strongly supported the inclusion of Leptoglanis and Zaireichthys in Amphiliidae, and the subfamil y Leptoglanidinae was erected independently by Diogo (2003) and Roberts (2003) for the inclusion of these genera in Amphiliidae. Roberts (2003) also described three new genera in Leptoglanidinae: Dolichamphilius Psammphiletria and Tetracamphilius Diogo ( 2005) diagnosed Amphiliidae by 12 characters, three of which are auta po morphies. The three auta po morphies included one character inspired by Mo (1991) (autopalatine markedly bifurcate posteriorly), and two newly discovered synapomorphies: arrector ventrali s markedly bifurcate mesially, and posterior portion of autopalatine significantly expanded dorsoventrally due to the presence of ventral and dorsal laminar expansions. The phylogenetic analysis of Siluriformes by Sullivan et al. (2006) based on R ag 1 and R ag 2 nuclear gene sequence data included four amphiliid genera including representatives of each of the three subfamilies. Their results strongly supported the monophyly of Amphiliidae in the parsimony and likelihood analyses with 100% bootstrap support and in the Bayesian analysis with a 100% posterior probability. To date, all analyses have agreed on the monophyly of the three subfamilies of the family, but the available morphological and molecular data give conflicting hypotheses of the relationships a mong subfamilies. The morphological hypotheses proposed by He et al. (1999) and Diogo (2005) have the subfamily Leptoglanidinae as sister to the subfamily Doumeinae, while the molecular results of Sullivan et al. (2006)

PAGE 21

21 have the genus Amphilius of the subf amily Amphiliinae as sister to the two doumeine genera included in the analysis. Taxon sampling was too limited in all analys e s to test the monophyly of the genera of Amphiliidae. Presented here are the results of a phylogenetic analysis of the family base d on data from one mitochondrial ( cytochrome b cyt b) locus and two nuclear (recombination activating gene subunit 2, R ag 2 and the first intron of the nuclear ribosomal S7 gene S7) loci from 12 of the 13 recognized amphiliid genera and more than half of the known species diversity of the family. While the analyses find support for two of the three currently recognized subfamilies, the subfamily Amphiliinae is consistently recovered as polyphyletic Additionally, several currently recognized genera are not recovered as monophyletic. Material and Methods Acquisition of Specimens Tissues for DNA analysis were obtained from the Genetics Research Repository at the Florida Museum of Natural History or borrowed from other institutions (Table 21 ). Materials exami ned in this study are deposited in the following institutions: the American Museum of Natural History, New York, New York ( AMNH ), the Natural History Museum, London ( BMNH ), the Cornell University, Vertebrate Collections, Ithaca, New York ( CU ), the University of Kansas Natural History Museum Lawrence, Kansas ( KU ), the National Museums of Kenya Nairobi, Kenya ( NMK ), the South African Institute for Aquatic Biodiversity Grahamstown, South Africa ( SAIAB), and the Florida Museum of Natural History, Gainesville Florida ( UF, FLMNH for tissue samples ), Zoologische Staatssammlung Mnchen, Munich, Germany ( ZSM ). Sequences for outgroup taxa (species of Mochokidae and Malapteruridae) and some ingroup taxa were obtained from GenBank (GenBank numbers provided in Table 2 1 ).

PAGE 22

22 DNA Extraction, PCR amplification, and Sequencing Genomic DNA was extracted from fin or muscle tissue using a 5.0% Chelex solution with 3 u L of P roteinase K and digestion at 55 C overnight or by using a Qiagen DNeasy Blood and Tissue Kit (Qiagen, Inc .) and following the manufacturers instructions. The extracted DNA was used as template for amplification of three markers by polymerase chain reaction (PCR). The complete cytochrome b gene (cyt b) was PCR amplified from the mitochondrial genome using previously published primers (GLU 2 and Pro R1; Hardman and Page, 2003). The complete recombination activating gene subunit 2 (Rag 2) was PCR amplified using previously published primers (MHFA and MHR1) from Hardman (2004). The first intron of the nuclear ribosomal S7 gene was amplified for a subset of the same individuals and species, using previously published primers (S7RP EX 1 F and S7R PEX2R) of Chow and Hazama (1998). The cycling protocol for the cyt b locus was as follows: initial denaturation at 94C for 30 s followed by 35 cycles of 94C for 30 s, 56C for 30 s, and 72C for 45 s, and a final extension of 72C for 10 min. The cycling protocol for the Rag 2 gene was: initial denaturation at 94C for 3 min followed by 35 cycles of 94C for 30 s, 56C for 30 s, and 72C for 1 min 30 s, and a final extension of 72C for 5 min. The cycling protocol for the S7 intron was: initial denaturation at 98C for 2 min followed by 33 cycles of 94C for 1 min, 53C for 45 s, and 72C for 1 min, and a final extension of 72 C for 8 min. Primers used in PCR amplification of the three markers were als o used in sequencing reactions performed by the Interdisciplinary Center for Biotechnology Research at the University of Florida. Sequences were edited in CodonCode Aligner version 3.7.1 (CodonCode Corporation, Dedham, MA, USA), and initially aligned using

PAGE 23

23 SEAVIEW v.4.1 (Galtier et al., 1996) using the MUSCLE algorithm (Edgar, 2004) Final alignments were inspected by eye for any obvious misreading Pairwise sequence divergences for all taxa were generated for each molecular marker using PAUP* 4.0b10 (Swofford, 2003) using the appropriate evolutionary models. A ppropriate evolutionary models for each marker were estimated by Modeltest 3.7 (Posada and Crandall, 1998) The best fit mode l and its parameters selected under the Akaike Information Criterion (AIC) were implemented in the respective Bayesian and Likelihood analyses. Gene Tree R econstruction Each marker and a concatenated dataset that included all three markers were analyzed with Bayesian methods implemented in Mr. Bayes version 3.1.2. In the concatenated analysis, partitions were set for each marker and model parameter values were estimated for each partition independently. Ten million generations of Markov chain Monte Carlo (M CMC) were performed using a random starting topology with trees sampled every 1000 generations. Four Markov chains with default priors were used in all analyses, and two replicate runs were conducted to ensure the MCMC went through a sufficient number of i terations to allow convergence in the estimations of tree topology with the best posterior probability. Tracer 1.4 (Rambaut and Drummond, 2004) was used to examine the distribution of log likelihood scores in order to determine stationarity and to determine the burnin of the MCMC analysis. In each of the analyses, the first 1 million generation (10%) were discarded as burnin. The remaining trees were used to calculate a 50% majority rule consensus topology. Each marker and the concatenated dataset were al so analyzed under a maximum likelihood optimality criterion as implemented in the software GARLI 2.0 (Zwickl, 2006) using the Garli web service:

PAGE 24

24 http://www.molecu larevolution.org/software/phylogenetics/garli/garli_create_job In the concatenated analysis, the partitions were the same as in the Bayesian analysis and model parameter values were again estimated for each partition independently. Topological robustness was investigated using 1000 nonparametric bootstrap replicates. Results Characteristi cs of the Sequence D ata and Models of Sequence E volution The length of the nucleotide sequences generated for the cyt b locus ranged from 1050 to 1158 bp. Of the 1158 nucleotides sites, 578 were variable and 530 were parsimony informative. Sequence divergence (TamuraNeicorrected distances) between species of Amphiliidae ranged from 0.2% to 22.4%. Within the subfamily Doumeinae, sequence divergence between species ranged from 0.4% to 20.1%, and in the subfamily Leptoglanidinae, sequence divergence between species ranged from 7.6% to 18.5%. Sequence divergence ranged from 0.2% to 20.7% between species of Amphilius 3.3% to 11.0% between species of Doumea, 0.4% to 11.6% bet ween species of Phractura and 0.8% to 17.8% between species of Zaireichthys Sequence divergence between Amphilius and species of Doumeinae ranged from 12.4% to 20.4%, between Amphilius and species of Leptoglanidinae, 16.0% to 20.4%; between Amphilius and Paramphilius 16.4% to 20.6%; between species of Doumeinae and Leptoglanidinae, 15.8% to 20.1%; between Doumeinae and Paramphilius 16.4% to 18.4%; and between Leptoglanidinae and Paramphilius 18.1% to 20.9%. The best fit model of sequence evolution selected for cyt b was GTR+I+G. The length of the nucleotide sequences generated for the Rag2 gene ranged from 917 to 945 bp. Of the 945 nucleotides sites, 378 were variable and 242 were

PAGE 25

25 parsimony informative. Sequence divergence (TamuraNeicorrected distances) between species of Amphiliidae ranged from 0.1% to 11.0%. Within the subfamily Doumeinae, sequence divergence between species ranged from 0.2% to 9.3% and in the subfamily Leptoglanidinae, sequence divergence between species ranged from 1.1% to 9.1%. Sequence divergence ranged from 0.1% to 5.2% between species of Amphilius 0.8% to 1.5% between species of Doumea, 0.5% to 2.5% between species of Phractura, and 1.1% to 5.7% between species of Zaireichthys Sequence divergence between Amphilius and species of Doumeinae ranged from 1.9% to 6.0%; between Amphilius and species of Leptoglanidinae, 6.1% to 14.1%; between Amphilius and Paramphilius 3.9% to 6.3%; between species of Doumeinae and Leptoglanidinae, 6.0% to 9.7%; between Doumeinae and Paramphilius 3. 4% to 5.0%; and between Leptoglanidinae and Paramphilius 7.3% to 10.1%. The best fit model of sequence evolution selected for the cyt b gene was HKY+I+G. The length of the nucleotide sequences generated for the S7 gene ranged from 455 to 591 bp and of the 680 nucleotides sites; 480 were variable and 366 were parsimony informative. Sequence divergence (TamuraNeicorrected distances) between species of Amphiliidae ranged from 0.2% to 35.4%. Within the subfamily Doumeinae, sequence divergence between species ranged from 0.4% to 11.1% and in the subfamily Leptoglanidinae, sequence divergence between species ranged from 3.5% to 20.8%. Sequence divergence ranged from 0.2% to 19.2% between species of Amphilius 2.6% to 7.5% between species of Doumea, 0.4% to 11. 2% between species of Phractura, and 3.6% to 14.3% between species of Zaireichthys Sequence divergence between Amphilius and species of Doumeinae ranged from 7.4% to 21.6%;

PAGE 26

26 between Amphilius and species of Leptoglanidinae, 23.2% to 35.4%; between Amphiliu s and Paramphilius 15.2% to 21.9%; between species of Doumeinae and Leptoglanidinae, 23.6% to 32.1%; between Doumeinae and Paramphilius 14.8% to 18.6%; and between Leptoglanidinae and Paramphilius 28.5% to 33.3%. The best fit model of sequence evolution selected for cyt b was GTR+I+G. Phylogenetic R elationships: cyt b The results of the Bayesian and likelihood analyses are shown in Figures 2 1 and 2 2. Both the Bayesian and likelihood the analyses recovered Amphiliidae as a well supported clade (1.0/86). In both analyses, the subfamilies Doumeinae and Leptoglanidinae were also recovered as clades with high support (Doumeinae: 1.0/89; Leptoglanidinae: 1.0/100). Amphiliinae was not recovered as a monophyletic group in either analysis with Paramphilius baudo ni recovered as sister to all remaining amphiliids (Doumeinae + Leptoglanidinae + Amphilius ) in both analyses. The likelihood analysis did not provide support for the monophyly of Amphilius but the genus was recovered as a monophyletic group in the Bayesian analysis with a post erior probability of 0.62. The H igh African morphological group of Amphilius (as recognized by Thomson & Page, 2010) was recovered in as a well supported clade in both analyses (1.0/96), but the Low African group was not recovered as a monophyletic group in either analysis. In both analyses, the genus Congoglanis was recovered as a well supported clade (1.0/100), and was recovered as the sister group to a clade that included all the other species of the subfamily Doumeinae. Doumea was recovered as a monophyletic group in the Bayesian analysis with a posterior probability of 0.82, but the likelihood analysis did not provide support for its monophyly. The genera Phractura and

PAGE 27

27 Trachyglanis were found to be polyphyletic in both analyses. I n the subfamily Leptoglanidinae, Zaireichthys was not recovered as a monophyletic group in either analysis. The H igh African species of Zaireichthys were recovered as a well supported clade (1.0/100) that was sister to a clade (0.92/64) that included Doloc amphilius Tetracamphilius Leptoglanis and two species of Zaireichthys from the Congo River basin. Phylogenetic R elationships: Rag 2 The results of the Bayesian and likelihood analyses for Rag 2 are shown in Figures 2 3 and 24. Both the Bayesian and lik elihood analyses recovered Amphiliidae as a well supported clade (1.0/100). In both analyses, the subfamilies Doumeinae and Leptoglanidinae were also recovered as clades with high support (Doumeinae: 1.0/85; Leptoglanidinae: 1.0/100). As in the cyt b analy ses, Amphiliinae was not recovered as a monophyletic group in either of the Rag 2 analyses. Paramphilius baudoni was again recovered as sister to all remaining amphiliids (Doumeinae + Leptoglanidinae + Amphilius ) in both of the Rag 2 analyses. The likelihood analysis did not provide support for the monophyly of Amphilius but the genus was recovered as a clade with weak support in the Bayesian analysis. The High African and L ow African morphological groups of Amphilius were recovered in well supported clades in both analyses (High African group: 1.0/99; Low African group: 0.99/83). In both analyses, the genus Congoglanis was recovered as a well supported clade (1.0/99), and was recovered as the sister group to a clade that included all the others species of the subfamily Doumeinae. The genera Doumea, Phractura and Trachyglanis were found to be polyphyletic in both analyses. In the subfamily Leptoglanidinae, Zaireichthys was not recovered as a monophyletic group in either analysis. As in the cyt b analysis, the H igh

PAGE 28

28 African species of Zaireichthys were recovered as a well supported clade (1.0/100) that was sister to a clade (1.0/99) that included Dolocamphilius Tetracamphilius Leptoglanis and two species of Zaireichthys from the Congo River basin. Phylogenet ic R elationships: S7 Intron 1 The results of the Bayesian and likelihood analyses are shown in 2 5 and 26. As in the cyt b and Rag 2 analyses, the Bayesian and likelihood analyses of the first intron of the S7 gene recovered Amphiliidae as a well supported clade (1.0/100), and both Doumeinae (1.0/89) and Leptoglanidinae (1.0/100) as well supported clades. Again, Amphiliinae was not recovered as a monophyletic group in either analysis with Paramphilius baudoni recovered as sister to all remaining amphiliids (Doumeinae + Leptoglanidinae + Amphilius ). Neither the Bayesian nor the likelihood analysis provided support for the monophyly of Amphilius but the High African and L ow African morphological groups of Amphilius were recovered in well supported clades in both analyses (High African group: 1.0/100; Low African group: 1.0/89). In both analyses, Congoglanis was again recovered as the sister group to a clade that included all other species of the subfamily Doumeinae, and the genera Doumea, Phractura and Trachy glanis were found to be polyphyletic. In the subfamily Leptoglanidinae, Zaireichthys was not recovered as a monophyletic group in either analysis. As in the cyt b and Rag 2 analyses, the H igh African species of Zaireichthys were recovered as a well support ed clade (1.0/100) that was sister to a clade (0.97/99) that included Dolocamphilius and Tetracamphilius and the one species of Zaireichthys from the Congo River basin that was included in the analysis.

PAGE 29

29 Phylogenetic relationships: cyt b + Rag 2 + S7 Intron 1 The results of the concatenated Bayesian and likelihood analyses of the cyt b, Rag 2, and the S7 intron are shown in Figures 2 7 and 28. In the analyses, Amphiliidae, Doumeinae, Leptoglanidinae, Amphilius L ow African Amphilius Congoglanis and H igh African Zaireichthys are all recovered as well supported. The phylogenetic position of Paramphilius baudoni is the same as in the analyses of the individual genes, sister to all remaining amphiliids (Doumeinae + Leptoglanidinae + Amphilius ). Congoglanis wa s again recovered as the sister group to a clade that included all other species of the subfamily Doumeinae, and the genera Doumea, Phractura and Trachyglanis were found to be polyphyletic. The H igh African species of Zaireichthys were again sister to a cl ade that included Dolocamphilius Tetracamphilius Leptoglanis and the two species of Zaireichthys from the Congo River basin. Discussion Amphiliidae was recovered as a well supported clade in all analyses, consistent with the results of Mo (1991), Diogo (2003, 2005), and Sullivan et al. (2006). The subfamilies Doumeinae and Leptoglanidinae were also recovered as well supported clades in all analyses. A sister group relationship between Doumeinae and a monophyletic Amphilius was recovered with strong supp ort in the concatenated analyses but was not well supported in any of the single genes analyses. The support for the monophyly of Doumeinae is consistent with the results of He et al. (1999) and Diogo (2003, 2005) which found strong support for the monophyly of the subfamily. The genus Congoglanis was recovered as sister to a clade formed by the remaining doumeine genera as hypothesized by Ferraris et al. (2011) who described

PAGE 30

30 the genus for species that lack many of the derived characters that had been trad itionally used to define Doumeinae. The relationships recovered for the other doumeine taxa in this study indicates that the remaining genera, as currently defined, do not accurately reflect phylogenetic relationships. The genus Phractura is not recovered as a monophyletic group, and Doumea was recovered as monophyletic only in the Bayesian analysis of the cyt b gene. Taxon sampling was limited for both of these genera, and notably missing were P hractura bovei the type species of Phractura, and species of the Doumea chappuisi complex, recognized by Ferraris et al. (2010) for three derived species of Doumea The current study included species from four of the five genera of Leptoglanidinae recognized by Roberts (2003), and the subfamily was recovered as a st rongly supported clade in all analyses. Leptoglanidinae is primarily distributed in the Congo basin with four of the five genera in the subfamily endemic to the basin. All analyses recovered two well supported clades. One clade corresponded to species endemic to the lower or middle Congo basin, and the other clade to the Zaireichthys rotundiceps complex, which is found in the upper Congo basin and other eastern and southern African basins. All analyses also recovered Paramphilius baudoni as sister to all re maining amphiliids (Doumeinae + Leptoglanidinae + Amphilius ), and consequently rendered the subfamily Amphiliinae as currently recognized polyphyletic. Paramphilius was erected by Pellegrin (1907) for Paramphilius trichomycteroides, and the genus was subse quently placed in the subfamily Amphiliinae with the genus Amphilius by Regan

PAGE 31

31 (1911). Paramphilius baudoni was described as Amphilius baudoni by Pellegrin (1928) and transferred to Paramphilius by Skelton (2007a ). The first phylogenetic study to include Pa ramphilius and Amphilius was that of He et al. (1999). Amphiliinae appeared as a monophyletic group in this work, but the authors were not able to find any synapomorphies to define the subfamily and stated that the reason for placing Amphilius and Paramphi lius together is the lack of synapomorphies shared by other genera. In contrast, the results of Diogo (2005) concluded that Amphiliinae was clearly a monophyletic taxon diagnosed with 10 synapomorphic characters, four of which he considered derived, homopl asy free autapomorphies. The results of the current study are not directly comparable to the studies of He et al. (1999) or Diogo (2005) because the species of Paramphilius used in the current study is different from the species used by He et al. (1999) and Diogo (2003, 2005), and Paramphilius has never been demonstrated to be a monophyletic group. He et al. (1999) and Diogo (2005) used Paramphilius trichomycteroides in their studies, the type species of Paramphilius which Skelton (1989) considered to be, along with P. firestonei the most derived species of the genus. In the current study, Paramphilius was represented by P. baudoni which was considered by Skelton (1989) to be the primitive, sister species of all upper Guinean Paramphilius Because P. baudoni is not the type species of Paramphilius and its relationships to the upper Guinean Paramphilius is not confirmed, revising the classification of the genus based on the results of the current study would be premature. The extensive taxonomic sampling in the genus Amphilius makes it possible to test the monophyly of the genus and the monophyly of the High African and Low African

PAGE 32

32 morphological groups. Amphilius was recovered as a well supported clade in the cytochrome b Bayesian analysis and the concatenat ed analysis, and recovered as a clade with weak support in the Rag 2 Bayesian analysis. The genus was not recovered as a monophyletic group in any of the likelihood analyses or the S7 Bayesian analysis. Amphilius can be divided into two morphological groups, a Low African group of species that have an epidermal fold at the base of the caudal fin and 6 + 7 or 7 + 8 principal caudal fin rays and a High African group of species that lack the epidermal fold and have 8 + 9 principal caudal fin rays (Skelton, 19 84; Thomson & Page, 2010). High Africa refers to parts of southern and eastern Africa where most of the land is at elevations well above 1000 ft. and much of it is at elevations above 4000 ft. Low Africa is northern and western Africa where most of the land is at elevations well between 1000 ft. (Roberts, 1975). The species of the High African morphological group were recovered as a well supported clade in all analyses. The species of the Low African morphological group were recovered as a well supported cl ade in the all of the nuclear gene analyses and the concatenated analyses, but not in either of the cytochrome b analyses. The taxonomic implications of this will be addressed following more extensive analyses of the phylogenetic relationships of the genus (Chapter 2).

PAGE 33

33 Table 2 1 Locality, museum catalogue numbers, tissue numbers, and GenBank accession numbers for specimens used in the molecular analyses. Bold: Sequences from Genbank. C AR = Central African Republic ; DRC = Democratic Republic of the Congo. Taxon Collection location Voucher catalog # Tissue # GenB ank nos. cyt b Rag 2 S7 Amphilius atesuensis Loh River, Guinea AMNH 248702 T5658 KC331868 JX573425 JX944623 Amphilius brevis Amba River, DRC ZSM 39281 DRC 2009/0885 KC331844 Amphilius c f. jacksonii Ivi River, Uganda UF 169238 FLMNH 2008 0370 KC331864 JX573390 Amphilius cf. lentiginosus Lucala River, Angola SAIAB 85100 SA85100 2 KC331867 JX573416 JX944607 Amphilius cf. zairensis Wagenia Falls, DRC CU 95906 CU 1852 KC331846 JX573417 JX 944608 Amphilius chalei Little Ruaha River, Tanzania UF 170728 FLMNH 2007 0903 KC331847 JX573430 JX944609 Amphilius cryptobullatus Lwela River, Zambia CU 91059 KC331848 JX573424 Amphilius cubangoensis Okavango River, Namibia SAIAB 96484 KC331849 JX5 73374 JX944610 Amphilius jacksonii Ishasha River, Uganda UF 169242 FLMNH 2008 0355 KC331866 JX573380 JX944622 Amphilius jacksonii Idete River, Tanzania UF 170729 FLMNH 2007 0961 KC331865 JX573389 JX944621 Amphilius kivuensis Ihihizo River, Uganda Vouch er lost FLMNH 2008 0341 KC331852 JX573391 JX944613 Amphilius krefftii Una River, Tanzania UF 170724 FLMNH 2007 0780 KC331853 JX573426 JX944614 Amphilius lampei Ukuma River, Ethiopia KU 29940 KU T1389 KC331854 JX573392 Amphilius laticaudatus Rotanda St ream, Mozambique SAIAB 67569 KC331855 JX573393

PAGE 34

34 Table 21 Continued Taxon Collection location Voucher catalog # Tissue # Ge nB ank nos. cyt b Rag 2 S7 Amphilius longirostris Cameroon AMNH 249566 KC331870 JX573364 JX944625 Amphilius maesii Mbali R iver, DRC AMNH 251819 KC331871 JX573363 JX944626 Amphilius natalensis Mahlabane River, Swaziland SAIAB 67360 SAIAB 67360a KC331857 JX573395 Amphilius nigricaudatus Cameroon AMNH 249567 KC331858 Amphilius nigricaudatus Dzousnou River, Rep. Congo AMNH 253311 JX573397 JX944627 Amphilius platychir Koumbouya River, Guinea AMNH 250575 RS4 KC331873 JX573400 JX944628 Amphilius rheophilus Finton River, Guinea AMNH 250584 RS11 KC331872 JX573401 JX944629 Amphilius uranoscopus Mbulumi River, Tanzania C U 93741 FLMNH 2007 0872 KC331862 JX573402 JX944619 Andersonia leptura Gribingui River, CAR CU 91439 KC331874 JX573361 Andersonia leptura Ethiopia CU 94684 KC331875 JX573396 Belonoglanis tenuis Gumi River, DRC AMNH 241576 KC331877 JX573366 Belo noglanis tenuis Mambili River, Rep. Congo CU 88147 KC331878 DQ492376 JX944630 Belonoglanis tenuis Lekeni River, Rep. Congo CU 92915 JX573375 Congoglanis alula Lower Congo River, Rep. Congo AMNH 240438 KC331879 JX573404 JX944631 Congoglanis alula Mbourou River, CAR CU 91611 KC331880 JX573358 Congoglanis alula Wagenia Falls, DRC CU 95908 KC331881 JX573405 JX944632 Congoglanis sagitta Mulembo River, Zambia CU 91042 KC331882 JX573406

PAGE 35

35 Table 21 Continued Taxon Collection location Voucher cat alog # Tissue # GenB ank nos. cyt b Rag 2 S7 Dolocamphilius sp. Lulua River, DRC AMNH 252785 AMNH t 080 7918 KC331900 JX573407 JX944642 Doumea gracila Lob River, Cameroon CU 90224 KC331883 JX573419 JX944633 Doumea sanaga Djerem River, Cameroon CU 93195 KC331884 JX573408 JX944634 Doumea typica Gabon UF 182240 FLMNH 2011 0484 KC331885 JX573427 JX944635 Doumea typica Gabon UF 182240 FLMNH 2011 0485 KC331886 JX573409 Leptoglanis xenognathus N'Sele River, DRC AMNH 249925 AMNH t 068 6731 KC331901 JX573423 Malapterurus beninensis Ogoou River, Gabon CU 80367 DQ492384 Malapterurus beninensis Kribi, Cameroon CU 93177 KC331841 JX573371 JX944603 Malapterurus tanganyikaensis Lake Tanganyika, Tanzania CU 88625 KC331842 JX573367 Paramphilius baudoni Ngampoku River, DRC AMNH 251816 KC331843 JX573410 JX944605 Phractura intermedia N'Sele River, DRC AMNH 250958 AMNH t 069 6855 KC331894 JX573372 JX944636 Phractura lindica Mambili River, Rep. Congo CU 88146 KC331891 JX573411 Phractura lindic a Biaro Creek, DRC CU 96007 JX573412 JX944638 Phractura longicauda Bal Creek, Gabon CU 81153 DQ492377 Phractura longicauda Ndingui Creek, Cameroon CU 90221 KC331893 JX573420

PAGE 36

36 Table 21 Continued Taxon Collection location Voucher catalog # Ti ssue # GenB ank nos. cyt b Rag 2 S7 Phractura longicauda Bal Creek, Gabon CU 92300 KC331892 JX573381 Phractura scaphyrhynchura N'Sele River, DRC AMNH 250919 KC331897 JX573368 JX944639 Phractura sp. Lower Congo River, DRC AMNH 241577 KC331887 JX573413 Phractura sp. Lekenie River, Rep. Congo CU 89214 KC331888 JX573377 Phractura sp. Mbudei stream, DRC ZSM 39413 DRC 2009/0857 KC331890 JX573369 Phractura sp. Amba River, DRC ZSM 39497 DRC 2009/0907 KC331889 JX573362 Phractura tenuicauda Mbourou River, CAR CU 91601 KC331896 JX573421 Phractura tenuicauda Kotto River, CAR CU 91606 KC331895 JX573376 JX944637 Synodontis afrofischeri Lake Albert, Uganda BMNH 2007.8.29.31 JJD 5547 FM87891 7 Synodontis batesii Bal Creek, Gabon CU 8112 9 DQ492382 Synodontis polli Lake Tanganyika BMNH 2006.3.6.21 JJD 5100 DQ886645 Tetracamphilius notatus Liwa River, CAR CU 91423 KC331902 JX573365 JX944643 Tetracamphilius pectinatus Oubangui River, CAR CU 93012 KC331903 JX573378 JX944644 Tra chyglanis cf. ineac Bakere River, DRC ZSM 39343 DRC 2009/0794 KC331898 JX573373 JX944640 Trachyglanis cf. sanghensis Luilaka River, DRC AMNH 239576 KC331876 JX573385

PAGE 37

37 Table 21 Continued Taxon Collection location Voucher catalog # Tissue # GenB ank nos cyt b Rag 2 S7 Trachyglanis cf. sanghensis Lower Congo River, DRC ZSM 37683 DRC 2008/406 KC331899 JX573382 JX944641 Zaireichthys cf. flavomaculatus Tshina River, DRC AMNH 247475 AMNH t 055 5436 KC331904 JX573357 Zaireichthys cf. rotundiceps Mo yowosi River, Tanzania CU 90424 KC331907 DQ492379 JX944645 Zaireichthys cf. rotundiceps Malaba River, Uganda UF 169234 FLMNH 2008 0298 KC331906 JX573379 Zaireichthys mandevillei Tomi River, CAR CU 91421 KC331905 JX573415 JX944648 Zaireichthys sp. Lu fubu River, Zambia CU 91033 CU 0162 JX573428 JX944646 Zaireichthys sp. Lubulafita Stream, Zambia CU 91037 CU 0241 JX573360 Zaireichthys sp. Ngona River, Zambia CU 91038 CU 0245 KC331908 JX573383 JX944647 Zaireichthys sp. Mutoloshi River, Zambia C U 91039 CU 0329 JX573384 Zaireichthys sp. Kanchibiya Stream, Zambia CU 91040 CU 0378 JX573422 Zaireichthys sp. Mumbuluma Falls, Zambia CU 91091 CU 0101 JX573414 Zaireichthys wamiensis Wami River, Tanzania UF 170710 FLMNH 2007 0828 KC331909 J X573429

PAGE 38

38 Figure 21 Majority rule consensus tree from the Bayesian analysis of the mitochondrial cyt b locus. Numbers on nodes are posterior probabilities. Asterisks indicate posterior probability values of 1.0. Catalog numbers of voucher specimens or tissue numbers follow the species name and correspond to data in Table 2 1.

PAGE 39

39 Figure 22 Majority rule consensus tree from the likelihood analysis of the mitochondrial cyt b locus. Numbers on nodes are bootstrap support values. A sterisks indicate boots trap support values of 100. Catalog numbers of voucher specimens or tissue numbers follow the species name and correspond to data in Table 2 1.

PAGE 40

40 Figure 23 Majority rule consensus tree from the Bayesian analysis of the nuclear Rag 2 gene. Numbers on nod es are posterior probability values. A sterisks indicate posterior probability values of 1.0. Catalog numbers of voucher specimens or tissue numbers follow the species name and correspond to data in Table 2 1.

PAGE 41

41 Figure 24 Majority rule consensus tree fro m the likelihood analyses of the nuclear Rag 2 gene. Numbers on nodes are bootstrap support values. A sterisks indicate bootstrap support values of 100. Catalog number of voucher specimen or tissue number of tissue samples follows species name and correspond to data in Table 2 1.

PAGE 42

42 Figure 25 Majority rule consensus tree from the Bayesian analysis of the nuclear S7 intron. Numbers on nodes are posterior probability values. A sterisks indicate posterior probability values of 1.0. Catalog numbers of voucher s pecimens or tissue numbers follow the species name and correspond to data in Table 2 1.

PAGE 43

43 Figure 26 Majority rule consensus tree from the likelihood analys i s of the nuclear S7 intron. Numbers on nodes are bootstrap support values. A sterisks indicate bootstrap support values of 100. Catalog numbers of voucher specimens or tissue numbers follow the species name and correspond to data in Table 2 1.

PAGE 44

44 Figure 27 Majority rule consensus tree from the Bayesian analysis of the concatenated cyt b, Rag 2, and S 7 data. Numbers on nodes are posterior probability values. A sterisks indicate posterior probability values of 1.0. Catalog numbers of voucher specimens or tissue numbers follow the species name and correspond to data in Table 2 1.

PAGE 45

45 Figure 28 Majority rule consensus tree from the likelihood analys i s of the concatenated cyt b, Rag 2, and S7 data. Numbers Numbers on nodes are bootstrap support values. A sterisks indicate bootstrap support values of 100. Catalog numbers of voucher specimens or tissue numbers follow the species name and correspond to data in Table 2 1.

PAGE 46

46 CHAPTER 3 MOLECULAR PHYLOGENY OF THE AFRICAN CATFISH GENUS AMPHILIUS (SILURIFORMES: AMPHILIIDAE). Background The African catfish genus Amphilius includes soft bodied species of Amphiliidae tha t lack spines in the dorsal and pectoral fins and are adapted for fast flow and rocky habitats (Skelton 1986) The genus is widely distributed throughout subSaharan Africa and can be divided into two morphological groups, a Low African group of primarily western species that have an epidermal fold at the base of the caudal fin and 6 + 7 or 7 + 8 principal caudal fin rays and a High African group of primarily eastern and southern species that lack the epidermal fold and have 8 + 9 principal caudal fin rays (Skelton, 1984; Thomson & Page, 2010). High Africa is southern and eastern Africa in which most of the land is at elevations well above 1000 ft. and much of it is at elevations above 4000 ft. Low Africa is northern and western Africa in which most of the land is at elevations between 500 and 1000 ft. (Roberts, 1975). In a study of the phylogenetic relationships of the family Amphiliidae (Chapter 2 ), Amphilius was not recovered as a monophyletic group in individual likelihood analyses of the mitochondrial ( cytochrome b, cyt b) locus and two nuclear (recombination activating gene subunit 2, R ag 2 and the first intron of the nuclear ribosomal S7 gene, S7) loci. The genus was recovered as a well supported clade in a concatenated Bayesian analysis of all three g enes and in a Bayesian analysis of the cyt b locus and was also recovered as a clade, but only with weak support in a Bayesian analysis of the Rag 2 locus and was not recovered as a monophyletic group in the Bayesian analysis of the S7 locus. In contrast to the genus as a whole, species of the High Africa n morphological group were recovered in a well supported clade in all analyses

PAGE 47

47 and the species of the Low African morphological group were recovered as a well supported clade in all of the nuclear gene a nalyses and the concatenated analyses, but not in either of the cytochrome b analyses. R esults of a n expanded phylogenetic analysis of the genus including sequences from 10 of the 18 currently recognized species of the Low Africa n group and 10 of the 11 currently recognized species of the High African group, provide additional support for the recognition of each morphological group as a separate genus. The name Anoplopterus Pfeffer 1889 is resurrected for the species of the High African group and all of the species of the Low Africa n group are retained in Amphilius The analyses also include extensive sampling of the Amphilius jacksonii complex and of species in the High Africa n group The analyses indicate that there is a great amount of undescribed dive rsity in both groups. Material and Methods Tissues for DNA analysis were obtained from the Genetics Research Repository at the Florida Museum of Natural History or borrowed from other institutions (Table 3 1). Materials examined in this study are deposite d in the following institutions: the American Museum of Natural History, New York, New York ( AMNH ), the Cornell University, Vertebrate Collections, Ithaca, New York ( CU ), the University of Kansas Natural History Museum Lawrence, Kansas ( KU ), the National Museums of Kenya, Nairobi, Kenya ( NMK ), the South African Institute for Aquatic Biodiversity Grahamstown, South Africa ( SAIAB), and the Florida Museum of Natural History, Gainesville, Florida ( UF, FLMNH for tissue samples ), Zoologische Staatssammlung Mnc hen Munich, Germany ( ZSM ). Additional sequences for outgroup taxa and some ingroup taxa used in this study are listed in Chapter 2.

PAGE 48

48 Methods of DNA extraction, PCR amplification, sequencing, and gene tree reconstruct ion are described in Chapter 2. Results Characteristics of the Sequence D ata and Models of Sequence E volution The length of the nucleotide sequences generated for the cyt b locus ranged from 906 to 1158 bp. Of the 1158 nucleotides sites, 548 were variable and 489 were parsimony informative. Seq uence divergence (TamuraNeicorrected distances) ranged from 0.0% to 20.8% within Amphilius 0.0% to 13.1% within the High African group, 0.0% to 20.8% within the Low African group, and 0.0% to 9.9% within the Amphilius jacksonii complex. The best fit mod el of sequence evolution selected for cyt b was GTR+I+G. The length of the nucleotide sequences generated for the Rag 2 locus ranged from 849 to 945 bp. Of the 945 nucleotides sites, 264 were variable and 142 were parsimony informative. Sequence divergenc e (Tamura Neicorrected distances) ranged from 0.0% to 5.7% within Amphilius 0.0% to 2.9% within the High African group, 0.0% to 5.6% within the High African group, and 0.0% to 1.4% within the Amphilius jacksonii complex. The best fit model of sequence ev olution selected for Rag 2 was GTR+I+G. The length of the nucleotide sequences generated for the S7 locus ranged from 522 to 591 bp. Of the 649 nucleotides sites, 388 were variable and 199 were parsimony informative. Sequence divergence (TamuraNeicorrec ted distances) ranged from 0.0% to 20.9% within Amphilius 0.0% to 7.9% within the High African group, 0.3% to 20.9% within the High African group, and 0.3% to 6.7% within the Amphilius jacksonii complex. The best fit model of sequence evolution selected f or Rag 2 was GTR+G.

PAGE 49

49 Phylogenetic R elationships: cyt b The results of the Bayesian and likelihood analyses are shown in Figures 3 1 and 3 2. The likelihood analysis did not provide support for the monophyly of Amphilius but the genus was recovered as a monophyletic group in the Bayesian analysis with a posterior probability of 1.0. The H igh African morphological group of Amphilius was recovered as a well supported clade in both analyses (1.0/94), and a clade of all the H igh Africa n species except Amphilius lampei and A. kivuensis was recovered in both analyses. The clade was composed of two clades that were well supported in both analyses. One clade included mostly species found in the Congo and Zambezi River basins and basins south the Zambezi River. The ot her clade included mostly species found in Tanzania. The Low African group was not recovered as a monophyletic group in either analysis. In both analyses, the Amphilius jacksonii complex was recovered as a well supported clade (1.0/80), and was recovered as the sister group of A. cf. lentiginosus In the Bayesian analysis, the Low African species were recovered in three strongly supported clades: a clade that included the A. jacksonii complex + A. cf. lentiginosus a clade that included the upper and lower Guinea species, and a clade that included two species from the Congo River basin. In the likelihood analysis, the A. jacksonii complex + A. cf. lentiginosus clade and the Congo River basin species clade were recovered with strong support but the upper and lower Guinea species were not recovered as a clade. Phylogenetic R elationships: Rag 2 The results of the Bayesian and likelihood analyses for Rag 2 are shown in Figures 3 3 to 38. Both the Bayesian and likelihood analyses recovered Amphilius as a

PAGE 50

50 clade w i th weak support (0.77/59). The H igh African morphological group of Amphilius was recovered in a well supported clade in both analyses (1.0/98), and a w ell supported clade of all the H igh Africa n species except A. lampei was recovered in both analyses. In both analyses three well supported clades within this clade were recovered. One clade included A. natalensis and A. laticaudatus and the other two clades were similar in composition to the clades recovered in the cyt b analyses. In b oth the Bayesian and likelihood analyses, the Low African species were recovered as a well supported clade (1.0/91) and the A. jacksonii complex was recovered as a well supported clade (1.0/98) sister to A. cf. lentiginosus The remaining species were recovered in two well supported clades in both analyses: a clade that included the upper Guinea species + A. longirostris and a clade that included two species from lower Guinea + A. maesii Phylogenetic R elationships: S7 Intron 1 The results of the Bayesian and likelihood analy ses are shown in Figures 3 9 and 3 10. Amphilius was not recovered as a clade in either analysis. The H igh African morphological group of Amphilius was recovered in a well supported clade in both analyses (1.0/97) and was recovered as sister to Dolichamphi lius In the H igh African group, two well supported clades we recovered that were similar in composition to the clades recovered in the cyt b and Rag 2 analyses. The Low African species were recovered as a clade in both analyses (1.0/59) and the A. jackso nii complex was recovered as a well supported clade (1.0/99) sister to A. cf. lentiginosus The upper Guinea species were recovered in a well supported clade (1.0/80) and A. maesii and A. nigrocaudatus were recovered as sister species in both analyses.

PAGE 51

51 Phy logenetic R elationships: cyt b + Rag 2 + S7 Intron 1 The results of the concatenated Bayesian and likelihood analysis of the cyt b, Rag 2, and the S7 intron are shown in Figure 3 11 and 312. Amphilius was recovered as a well supported clade (1.0/93). The High and Low African groups were also recovered as well supported clades (High African group: 1.0/100; Low Africa n group: 1.0/98). In the H igh African clade, a well supported clade (1.0/95) was recovered that included all species except A. lampei Within t his clade, a well supported clade (1.0/100) was recovered that included all species from the Congo, Zambezi, and Rovuma River basins and basins south of the Zambezi River except A. natalensis and A. laticaudatus A weakly supported clade (0.53/50) that inc luded most of the species from Tanzania was also recovered. In the L ow African clade, two well supported clades were recovered. A clade including the upper Guinea species + A. longirostis and one specimen of A. nigricaudatus (1.0/95), and a clade including the A. jacksonii complex, A. cf. lentiginosus and all other lower Guinea species including a second specimen of A. nigricaudatus (0.99/97). The lower Guinea species were recovered as a well supported clade (1.0/100) and the A. jacksonii complex was recovered as a well supported clade (1.0/100) sister to A. cf. lentiginosus Discussion Many of the specimens used in this analysis could not be placed in any known species of Amphilius and likely represent undescribed species. It was therefore necessary to li st these specimens with various terms indicating their apparent taxonomic position. What follows is an explanation of all the specimens that could not be assigned to species. Amphilius cf. lentiginosus is a species very similar to A. cf.

PAGE 52

52 lentiginosus that is found in the Cuanza River basin and the Kwango River system, Congo River basin in Angola. A mphilius cf. jacksonii is a species of the A. jacksonii complex from the Ive River in Uganda that is being described by another researcher. The sequences identifi ed as A. cf. rheophilius are from a specimen that was identified as such by Schmidt and Pezold (2011). Amphilius n. sp. Congo is an undescribed species known from three specimens from the Republic of the Congo. Amphilius n. sp. A is an undescribed species from the lower Congo River. Amphilius sp. B and C are also specimens from the lower Congo River, both of which are very similar to A. zairensis The specimens identified as A. cf. zairensis are also very similar to A. zairensis but are from Wagenia Falls, middle Congo River. Amphilius zairensis was described from the lower Congo River so Amphilius sp. B or C may actually be A. zairensis but direct comparison of the specimens to the types is needed to determine which one is A. zairensis Initially, the speci men of Amphilius sp. C and the specimens of Amphilius cf. zairensis were identified as the same species but the molecular data gave conflicting results as to whether A. cf. zairensis is the same species as either Amphilius sp. B or C. The sequences ident ified as Amphilius n. sp. Rovuma are from a small specimen from a tributary of the lower Rovuma River. The specimen and other specimens collected with it are very similar to several small specimens collected from a southeast ern tributary of Lake Malawi in Mozambique. Sequences from one of these specimens are identified as Amphilius n. sp. Malawi trib. The sequences identified as Amphilius n. sp. Malawi are from a specimen from the North Rukuru River, a northwestern tributary of Lake Malawi. Sequences identi fied as Amphilius n. sp. S.

PAGE 53

53 Africa are from specimens from the Limpopo, Incomatti, and Pongola River basins in South Africa and Swaziland. Sequences identified as Amphilius n. sp. mottled and Amphilius n. sp. Rufiji are from three specimens collected from three sites in the Rufiji River basin. All three specimens are heavily mottled unlike all other specimens collected in the basin except one other specimen collected with one of the specimens above. This other mottled specimen is morphologically distinct fr om the above specimens but could not be included in the analyses because a tissue sample was not taken from it. Amphilius cf. natalensis are from the Buzi River basin. They are similar to A. natalensis but differ from A. natalensis from South Africa in that they have an adipose fin confluent with the caudal fin. Amphilius n. sp. Ruvu and Amphilius n. sp. Mangula are species being described in a taxonomic revision of Amphilius uranoscopus (Chapter 3 ). Amphilius n. sp. Ruvu is known from one location in the Ruvu River basin and Amphilius n. sp. Mangula is known from several localities in the Rufiji River basin. All specimens of Amphilius n. sp. Malagarasi are from the Lower Malagarasi River in Tanzania that were formerly identified as A. uranoscopus The sequences identified as Amphilius sp. Great Ruaha are from a specimen collected from the Great Ruaha River at a bridge in Kidatu with one of the mottled specimens mentioned above. The cyt b analysis indicates it should be identified as A. chalei while the Rag 2 analysis indicates it should be identified as Amphilius n. sp. Ruvu; however it is morphologically distinct from both of these species. The A mphilius jacksonii complex was recovered as a well supported clade and sister to A. cf. lentiginosus in all anal yses. In addition to being the only species of the Low African morphological group distributed in H igh Africa, they differ from all other

PAGE 54

54 species of the genus by having a variably mottled body pattern that includes dark saddles (vs. body coloration not mot tled and without dark saddles). The relationships recovered within the A. jacksonii complex are consistent with a taxonomic revision of the complex that will describe several new species (Chapter 4). The concatenated analysis recovered a clade of A. cf. ja cksonii and the following geographic groupings: Lake Edwards drainage, Uganda; Lake Kyogo and Lake Victoria drainages, Kenya and Uganda; Rufiji and Wami basins, Tanzania; Upper Congo River; and Malagarasi River drainage, Tanzania. Most of the resolution of these clades is due to variation in the cyt b locus. The Rag 2 locus was not variable enough to resolve relationships within the complex and too few sequences were available for the S7 locus to show relationships. Although Rag 2 was not able to resolve most relationships within the complex, it did recover A. cf. jacksonii as sister to all the remaining species in the complex in contrast to the cyt b analyses which recovered it as sister to specimens of A. jacksonii from the Lake Edwards drainage, Uganda. T he remaining species of the Low African Amphilius were generally recovered in two clades: a clade of mostly upper Guinea species and a clade of mostly lower Guinea species plus Congo River basin species. The two specimens of A. nigricaudatus were recovered in different clades. One of the specimens only had nuclear genes amplified and was recovered in the Guinea species and Congo River basin clade. The other specimen had only the mitochondrial gene amplified and was recovered in the upper Guinea clade sister to an A. longirostris specimen with which it was collected. The cyt b sequence from the A. nigricaudatus specimen was nearly identical to the cyt b

PAGE 55

55 sequence from the A. longirostris specimen indicating introgression between these species. Amphilius lampei was recovered as sister to all other species of H igh African Amphilius in all analyses except the cyt b likelihood analysis. The distribution of A. lampei is disjunct from all other species in the genus. It is found only in the Shebeli River basin in Ethiopia. It is also one of only three species of African Amphilius with a very posteriorly placed dorsal fin ( first dorsal pterygiophore intercept count usually four, five or six vs. usually two or less). The phylogenetic placement of the two other species w ith a very posteriorly placed dorsal fin, A. kivulensis and A. natalensis varied considerably between analyses. Amphilius kivulensis was recovered as sister to all other species of H igh African Amphilius in the cyt b likelihood analysis, sister to all oth er species of H igh African Amphilius except A. lampei in the cyt b Bayesian analysis, and nested within a clade of mostly Tanzanian species in both Rag 2 analyses and the concatenated Bayesian analysis. It was not included in the S7 analyses and was recovered in a polytomy with several other clades in the concatenated likelihood analysis. Amphilius natalensis was recovered as sister to A. cf. natalensis in the cyt b analyses and the concatenated likelihood analysis but as sister to a clade that included A laticaudatus and A cf. natalensis in the Rag 2 analyses and the concatenated Bayesian analysis. Amphilius natalensis was not included in the S7 analyses. The A mphilius uranoscopus species group recognized by Thomson and Page (2010) for H igh African Amp hilius species with a relatively elongate body ( usually 36 42 vs. 32 35 total vertebrae) and an anteriorly placed dorsal fin, was not recovered as a

PAGE 56

56 monophyletic group in any of the analyses. In addition to the various phylogenetic placements of A. kivulen sis and A. natalensis rendering the group paraphyletic relatively short bodied species were consistently found to be closely related to relatively elongate species. There did not appear to be any correlation between body length and phylogenetic position. Within the H igh African Amphilius two clades (identified as clades A & B in F igures 3 3 and 34) were consistently recovered. No morphological characters could be found to distinguish these two clades but clade A mostly contains species that are mottled or have darkly blotched bodies while species in clade B are never mottled and generally dont have dark blotches on their bodies. There is also a strong correlation with geography. Most of the species in clade A are from the Congo basin and river basins on the east coast of Africa from the Rovuma River and south. Most of the species in clade B are river basins on the east coast of Africa north of the Rovuma River. The H igh African Amphilius species were recovered as a well supported clade in all analyses. Th e L ow African Amphilius species were recovered as a clade in all the analyses that included nuclear genes and was well supported in all except for the S7 likelihood analyses which had 59% bootstrap support. Neither of the two analyses that included only th e cyt b locus recovered the Low African Amphilius species as a monophyletic group. Only the cyt b Bayesian analysis and the concatenated likelihood analyses recovered Amphilius as a well supported clade. These results are congruent with the results of Chapter 2 which also found strong support for the monophyly of the H igh

PAGE 57

57 African species and more support for the monophyly of the Low African species than for the entire genus. Given the stronger support for the individual High and Low African clades, the weak support for a clade including all the H ig h and L ow African species, and the morphol ogical differences between the High and Low African clades, the High and L ow African species are placed in seperate genera. The name Anoplopterus Pfeffer, 1889 is availab le for the H igh African species and the L ow African species are retained in Amphilius Taxonomic Descriptions Amphilius Gnther, 1864 Amphilius Gnther. 1864: 115, 134. (Type species: Pimelodus platychir Gnther, 1864. Type by monotypy. Diagnosis (based, in part, on Roberts, 2003 and Skelton, 2007b ) Distinguished from all genera of Amphiliidae except Anoplopterus by the follow ing combination of characters: s oft bodied, outermost pectoral and pelvic fin rays with elaborate lepidotrichia and unculiferous pads. Caudal peduncle short and deep, length more than five times in SL and depth less than three times in caudal peduncle length. Barbels slender and filamentous. Branchiostegal membrane divided midventrally, mouth terminal with broad gape. Amphilius is d istinguished from Anoplopterus by the presence (vs. absence) of an epidermal fold at the base of the caudal fin and 6 + 7 or 7 + 8 principal caudal fin rays (vs. 8+9). Species i ncluded Amphilius atesuensis, A. brevis, A. caudosignatus, A. dimonikensis, A. grammatophorus A. jacksonii, A. kakrimensis, A. korupi, A. lamani, A. lentiginosus, A.

PAGE 58

58 longirostris, A. maesii, A. mamonekenensis, A. nigricaudatus, A. opisthophthalmus, A. platychir, A. pulcher, and A. rheophilus. Distribution Species of Amphilius are distributed in SubSaharan Africa from the Senegal basin in West Africa west and south to the Cuvo River basin in Angola. One species, A. jacksonii is found in the upper Nile River basin in Uganda, Burundi, Rwanda, and Kenya, in the Wami and Rufiji basins in eastern Tanzania, the Lake Rukwa basin of southwestern Tanzania and northeastern Zambia, and the Lake Manyara basin in northeast Tanzania. Several species of the genus are found in the lower and middle Congo River basin, but only A. jacksonii is found in the upper Congo River (including tributaries of Lake Tanganyika). Notes Recent authors have considered A. grammatophorus a synonym of A. platychir but Schmidt and Pezold (2011) presented data that indicates that it is a valid species. The sequences used in this study are from a specimen that they identified as A. grammatophorus Anoplopterus Pfeffer 1889 Anoplopterus Pfeffer, 1889: 15. (Type species: Anoplopterus uranoscopus Pfeffer, 1889. Type by monotypy. Chimarrhoglanis Vaillant, 1897: 81. (Type spec ies: Chimarrhoglanis leroyi Vaillant 1897. Type by monotypy. Diagnosis (based, in part, on Roberts, 2003 and Skelton, 2007b ) Distinguished from all genera of Amphiliidae except Amphilius by the following combination of characters: Soft bodied, outermos t pectoral and pelvic fin rays with elaborate lepidotrichia and unculiferous pads. Caudal peduncle short and deep, length

PAGE 59

59 more than five times in SL and depth less than three times in caudal peduncle length. Barbels slender and filamentous. Branchiostegal membrane divided midventrally, mouth terminal with broad gape. Anoplopterus is distinguished from Amphilius by the absence (vs. presence) of an epidermal fold at the base of the caudal fin and 8+9 princip al caudal fin rays (vs. 6 + 7 or 7 + 8). Species i n cluded Anoplopterus athiensis A. brevidorsalis A. chalei, A. cryptobullatus, A. cubangoensis, A. grandis, A. hargeri A. kivuensis, A. krefftii, A lampei, A. laticaudatus, A. leroyi, A. natalensis, A. uranoscopus and A. zairensis. Distribution Species of Anoplopterus are distributed in SubSaharan Africa from the Ewaso Ngiro River basin in Kenya south to the Umzimkulu ( Mzimkulu ) River b asin in South Africa. One species, A. cubangoensis is found in the upper Zambezi River basin and the Okavango basin. Several species are found in the Congo River including the lower, middle, and upper parts of the basin. One species, A. kivuensis, is found in the upper Nile River basin in Uganda, Burundi, Rwanda, and Kenya and A. lampei is found in the Shebele River bas in in Ethiopia. Notes Anoplopterus brevidorsalis A. cubangoensis, A. hargeri and A. leroyi have been considered synonyms of A. uranoscopus by recent authors but are being redescribed and recognized as valid species in a taxonomic revision of A. uranoscopus (Chapter 3).

PAGE 60

60 Table 3 1 Locality, museum catalogue numbers, tissue numbers, and GenBank accession numbers for specimens used in the molecular analyses. DRC = Democratic Republic of the Congo. Taxon Collection location Voucher catalog # Tissue # GenB an k nos. cyt b Rag 2 S7 Amphilius brevidorsalis Muoohwa River, Mozambique SAIAB 67582 KC331845 JX573386 JX944606 Amphilius brevidorsalis Muzuma River, Mozambique SAIAB 67618 KC620191 Amphilius brevidorsalis Muchanga Stream, Mozambique SAIAB 6 7621 SAIAB 67621 1 KC620192 Amphilius cf. jacksonii Ivi River trib., Uganda UF 169237 FLMNH 2008 0376 KC620284 KC667103 Amphilius cf. jacksonii Ivi River trib., Uganda UF 169237 FLMNH 2008 0377 KC620285 KC667104 Amphilius cf. jacksonii Ivi River trib., Uganda UF 169237 FLMNH 2008 0378 KC620286 Amphilius cf. jacksonii Ivi River trib., Uganda UF 169237 FLMNH 2008 0379 KC620287 Amphilius cf. lentiginosus Kuquema River, Angola SAIAB 85058 SAIAB 85058 1s KC620288 KC667076 Amphilius cf. na talensis Muchinene Stream, Mozambique SAIAB 67413 SAIAB 67413a KC620226 KC667077 Amphilius cf. natalensis Muchinene Stream, Mozambique SAIAB 67413 SAIAB 67413b KC620227 KC667078 Amphilius cf. rheophilus Bafing River, Guinea AMNH 248685 11803F KC62029 1 KC667121 Amphilius cf. zairensis Wagenia Falls, DRC CU 95907 CU 1862 KC620205 KC667080 Amphilius cf. zairensis Wagenia Falls, DRC CU 95907 CU 1863 KC620206 JX573387 Amphilius chalei Little Ruaha River, Tanzania UF 170728 FLMNH 2007 0900 KC620207 KC667075 Amphilius cryptobullatus Mulembo River, Zambia CU 91058 KC620208 KC667081

PAGE 61

61 Table 31 Continued Taxon Collection location Voucher catalog # Tissue # GenB ank nos. cyt b Rag 2 S7 Amphilius cubangoensis Kaluebo, DRC AMNH 252831 AMNH t 0 81 8077 KC331863 JX573403 JX944620 Amphilius cubangoensis SAIAB 95820 SAIAB 95820 1 KC620209 KC667082 KC620193 Amphilius grammatophorus Ftor River, Guinea AMNH 250593 11303F KC331869 JX573418 JX944624 Amphilius grandis Sagana River, Kenya NMK Uncat. Tissue # 32 KC331850 JX573388 JX944611 Amphilius grandis Ewaso Nyiro River, Kenya NMK uncat. Tissue # 98 KC620210 Amphilius hargeri Ruo River, Malawi SAIAB 87161 (tissue only) SAIAB QQ0619 KC331851 JX573359 JX944612 Amphilius jacksonii Malagarasi R iver, Tanzania CU 90414 KC620257 Amphilius jacksonii Malagarasi River, Tanzania CU 95207 CU 1235 KC620258 KC667089 KC620200 Amphilius jacksonii Malagarasi River, Tanzania CU 95207 CU 1236 KC620259 KC667090 Amphilius jacksonii Malagarasi River, Ta nzania CU 95207 CU 1237 KC620260 KC667091 Amphilius jacksonii Malagarasi River, Tanzania CU 95211 CU 1562 KC620261 KC667092 KC620201 Amphilius jacksonii Malagarasi River, Tanzania CU 95211 CU 1563 KC620262 KC667093 Amphilius jacksonii Malagarasi Riv er, Tanzania CU 95211 CU 1564 KC620263 KC667094 Amphilius jacksonii Malagarasi River, Tanzania CU 90426 FLMNH 2008 0167 KC620264 Amphilius jacksonii Malagarasi River, Tanzania CU 90426 FLMNH 2008 0168 KC620265

PAGE 62

62 Table 31 Continued Taxon Collec tion location Voucher catalog # Tissue # GenB ank nos. cyt b Rag 2 S7 Amphilius jacksonii Malagarasi River, Tanzania CU 90426 FLMNH 2008 0169 KC620266 KC667088 Amphilius jacksonii Wami River, Tanzania UF 170707 FLMNH 2007 0835 KC620268 KC667101 Amphilius jacksonii Divue River, Tanzania UF 170708 FLMNH 2007 0886 KC620267 KC667102 Amphilius jacksonii Sonjo River, Tanzania UF 170702 FLMNH 2007 0928 KC620269 KC667099 Amphilius jacksonii 12 Bridges River, Tanzania UF 170706 FLMNH 2007 0994 KC62 0270 KC667100 Amphilius jacksonii Manafwa River, Uganda UF 169233 FLMNH 2008 0286 KC620271 KC667098 Amphilius jacksonii Manafwa River, Uganda UF 169233 FLMNH 2008 0287 KC620272 KC667097 Amphilius jacksonii Manafwa River, Uganda Voucher lost FLMNH 2008 0288 KC620273 Amphilius jacksonii Musanyi River, Kenya NMK uncat. Tissue # 1596 KC620274 Amphilius jacksonii Ishasha River, Uganda UF 169242 FLMNH 2008 0353 KC620275 Amphilius jacksonii Ishasha River, Uganda UF 169242 FLMNH 2008 0354 KC620276 KC667095 Amphilius jacksonii Munyage River, Uganda UF 169243 FLMNH 2008 0389 KC620277 KC667096 Amphilius jacksonii Munyage River, Uganda UF 169243 FLMNH 2008 0390 KC620278 Amphilius jacksonii Luongo River, Zambia CU 91053 KC620279 Amphilius jacksonii Luongo River, Zambia CU 91055 KC620280 KC667084

PAGE 63

63 Table 31 Continued Taxon Collection location Voucher catalog # Tissue # GenB ank nos. cyt b Rag 2 S7 Amphilius jacksonii Kanchibiya Stream, Zambia CU 91056 KC620281 KC 667086 Amphilius jacksonii Diptera River, DRC SAIAB 77478 SAIAB 77478a KC620282 KC667085 Amphilius jacksonii Diptera River, DRC SAIAB 77478 SAIAB 77478b KC620283 KC667087 Amphilius kivuensis Ihihizo River, Uganda Voucher lost FLMNH 2008 0339 KC620 211 Amphilius krefftii Kihuhwi River, Tanzania UF 170713 FLMNH 2007 0748 KC620212 KC667105 Amphilius krefftii Pangani River, Tanzania UF 170712 FLMNH 2007 0765 KC620213 KC667106 Amphilius krefftii Soni Falls, Tanzania UF 170744 FLMNH 2007 0773 K C620214 KC667107 Amphilius krefftii Soni Falls, Tanzania UF 170744 FLMNH 2007 0774 KC620215 Amphilius krefftii Kikuletwa River, Tanzania UF 170721 FLMNH 2007 0784 KC620216 Amphilius krefftii Mokoyetti River, Kenya NMK uncat. Tissue # 1758 KC62 0217 Amphilius lampei Ukuma River, Ethiopia KU 29940 KU T1387 KC620218 Amphilius lampei Ukuma River, Ethiopia KU 29940 KU T1388 KC620219 Amphilius laticaudatus Muchinene Stream, Mozambique SAIAB 67411 SAIAB 67411 1 KC620220 Amphilius la ticaudatus Mussapa Grande R., Mozambique SAIAB 67584 KC667108 Amphilius leroyi Ruvu River, Tanzania UF 170711 FLMNH 2007 0801 KC620221 KC667109 Amphilius leroyi Ruvu River, Tanzania UF 170711 FLMNH 2007 0802 KC331856 JX573394 JX944615

PAGE 64

64 Tab le 3 1 Continued Taxon Collection location Voucher catalog # Tissue # GenB ank nos. cyt b Rag 2 S7 Amphilius leroyi Ruvu River, Tanzania CU 93725 FLMNH 2007 0812 KC620222 KC667110 Amphilius leroyi Ruvu River, Tanzania UF 170717 FLMNH 2007 0813 KC6 20223 KC667111 Amphilius leroyi Ruvu River, Tanzania UF 170726 FLMNH 2007 0817 KC620224 KC667112 Amphilius leroyi Ruvu River, Tanzania UF 170726 FLMNH 2007 0818 KC620225 KC667113 Amphilius longirostris Ngonebok Creek, Cameroon CU 89988 KC620289 K C667114 KC620202 Amphilius maesii Kwango River, DRC ZSM 38161 DRC 2008/633 KC620290 KC667115 Amphilius n. sp. A River Louvoubi, Rep. Congo AMNH 240436 AMNH t 022 2158 KC620228 KC667117 Amphilius n. sp. A River Louvoubi, Rep. Congo AMNH 240240 KC620 203 KC667118 KC620194 Amphilius n. sp. Buzi Muchanga Stream, Mozambique SAIAB 67621 SAIAB 67621 1 KC620230 KC667120 Amphilius n. sp. Congo Lekenie River, Rep. Congo CU 92435 KC667122 Amphilius n. sp. Rufiji Sonjo River, Tanzania UF 184935 FLMNH 2 007 0926 KC620243 KC667123 KC620195 Amphilius n. sp. Rufiji Great Ruaha River, Tanzania UF 170715 FLMNH 2007 0977 KC620231 KC667124 Amphilius n. sp. Malawi trib Lake Malawi trib., Mozambique SAIAB 87161 SAIAB 87161 KC620232 KC667125 Amphilius n. sp. Lake Rukwa Saisi River Zambia MMo7 G KC620233

PAGE 65

65 Table 31 Continued Taxon Collection location Voucher catalog # Tissue # GenB ank nos. cyt b Rag 2 S7 Amphilius n. sp. Lake Rukwa Saisi River Zambia MMo7 H KC620234 KC667126 Amphilius n. sp. Malagarasi Malagarasi River, Tanzania CU 90401 KC331859 JX573398 JX944616 Amphilius n. sp. Malagarasi Malagarasi River, Tanzania CU 90423 KC620236 KC667131 KC620199 Amphilius n. sp. Malagarasi Malagarasi River, Tanzania CU 95213 CU 1242 KC620237 KC667 127 Amphilius n. sp. Malagarasi Malagarasi River, Tanzania CU 95216 CU 1558 KC620240 KC667128 KC620196 Amphilius n. sp. Malagarasi Malagarasi River, Tanzania CU 95216 CU 1559 KC620238 KC667129 KC620197 Amphilius n. sp. Malagarasi Malagarasi River, Tan zania CU 95218 CU 1582 KC620239 KC667130 KC620198 Amphilius n. sp. Malawi North Rukuru R., Malawi SAIAB 78387 KC620241 KC667132 Amphilius n. sp. Mangula Lukosi River, Tanzania CU 93743 FLMNH 2007 0910 KC620235 KC667133 Amphilius n. sp. Mangula Sonj o River, Tanzania UF 184935 FLMNH 2007 0925 KC620242 KC667134 Amphilius n. sp. Mottled Lukosi River, Tanzania UF 184937 FLMNH 2007 0911 KC620244 KC667135 Amphilius n. sp. Rovuma Chimbonila River, Mozambique SAIAB 87188 KC620245 KC667083 Amphilius n. sp. Ruvu Ruvu River, Tanzania CU 9 7551 FLMNH 2007 0824 KC331860 JX573370 JX944617 Amphilius n. sp. Ruvu Ruvu River, Tanzania UF 184938 FLMNH 2007 0825 KC620247 KC667136

PAGE 66

66 Table 31 Continued Taxon Collection location Voucher catalog # Tissue # GenB ank nos. cyt b Rag 2 S7 Amphilius n. sp. S. Africa Komati River, Swaziland SAIAB 66126 KC620248 KC667137 Amphilius n. sp. S. Africa Sakane River, Swaziland SAIAB 66166 SAIAB 66166 1 KC331861 JX573399 JX944618 Amphilius n. sp. S. Africa Ngwengwa na River, Swaziland SAIAB 70713 SAIAB 70713 1 KC620251 KC667139 Amphilius n. sp. S. Africa Groot Marico R., South Africa SAIAB 78430 KC620250 Amphilius n. sp. S. Africa Magalies River, South Africa SAIAB 78453 KC620249 KC667138 Amphilius sp. B Lower Congo River, DRC AMNH 246501 KC620229 KC667119 Amphilius sp. C Lower Congo River, DRC AMNH 242332 KC620204 Amphilius sp. Rufiji Great Ruaha River, Tanzania UF 170714 FLMNH 2007 0978 KC620246 KC667140 Amphilius uranoscopus Mbulumi River, Tanzania UF 170716 FLMNH 2007 0873 KC620252 Amphilius uranoscopus Divue River, Tanzania UF 170723 FLMNH 2007 0881 KC620253 KC667141 Amphilius uranoscopus Divue River, Tanzania UF 170723 FLMNH 2007 0882 KC620254 KC667142 Amphilius uranoscopus 12 Bridges River, Tanzania UF 170720 FLMNH 2007 0993 KC620255 KC667116 Amphilius uranoscopus Wami River, Tanzania UF 170718 FLMNH 2007 1004 KC620256 KC667143

PAGE 67

67 F igure 31 Majority rule consensus tree from the Bayesian analysis of the mitochondrial cyt b locus. Numbers on nodes are posterior probabilities. A sterisks indicate posterior probability values of 1.0. Catalog numbers of voucher specimens or tissue numbers follow the species name and correspond to data in Table 3 1.

PAGE 68

68 Figure 32 Majority rule consensus tree from the likelihood analysis of the mitochondrial cyt b locus. Numbers on nodes are bootstrap support values. A sterisks indicate bootstrap support values of 100. Catalog numbers of voucher specimens or tissue numbers follow the species name and correspond to data in Table 3 1.

PAGE 69

69 Figure 33 Majority rule consensus tree from the Bayesian analysis of the nuclear Rag 2 gene. Numbers on nodes are posterior probability values. A sterisks indicate posterior probability values of 1.0. Catalog numbers of voucher specimens or tissue numbers follow the species name and correspond to data in Table 3 1.

PAGE 70

70 Figure 34 Majority rule consensus tree of the from the Bayesian analysis of the nuclear Rag 2 gene showing only the Amphilius jacksonii complex. Numbers on nodes are posterior probability values. A sterisks indicate posterior probability values of 1.0. Catalog numbers of voucher specimens or tissue numbers follow the species name and correspond to data in Table 3 1.

PAGE 71

71 Figure 35 Majority rule consensus tree of the from the Bayesian analysis of the nuclear Rag 2 gene showing only the H igh African Amphilius Numbers on nodes are posterior probability values. A sterisks indicate posterior probability values of 1.0. Catalog numbers of voucher specimens or tissue numbers follow the species name and correspond to data in Table 3 1.

PAGE 72

72 Figure 36 Majority rule consensus tree from the likelihood analyses of the nuclear Rag 2 gene. Numbers on nodes are bootstrap support values. A sterisks indicate bootstrap support values of 100. Catalog number of voucher specimen or tissue number of tissue samples follows species name and correspond to data in Table 3 1.

PAGE 73

73 Figure 37 Majority rule consensus tree from the likelihood analyses of the nuclear Rag 2 gene showing only the Amphilius jacksonii complex Numbers on nodes are bootstrap support values. A sterisks indicate bootstrap support values of 100. Catalog number of voucher specimen or tissue number of tissue samples follows species name and correspond to data in Table 3 1

PAGE 74

74 Figure 38 Majority rule consensus tree from the likelihood analyses of the nuclear Rag 2 gene showing only the H igh African Amphilius Numbers on nodes are bootstrap support values. A sterisks indicate bootstrap support values of 100. Catalog number of voucher specimen or tissue number of tissue samples follows species name and correspond to data in Table 3 1.

PAGE 75

75 Figure 39 Majority rule consensus tree from the Bayesian analysis of the nuclear S7 intron. Numbers on nodes are posterior probability val ues. A sterisks indicate posterior probability values of 1.0. Catalog numbers of voucher specimens or tissue numbers follow the species name and correspond to data in Table 3 1.

PAGE 76

76 Figure 310 Majority rule consensus tree from the likelihood analys i s of th e nuclear S7 intron. Numbers on nodes are bootstrap support values. A sterisks indicate bootstrap support values of 100. Catalog numbers of voucher specimens or tissue numbers follow the species name and correspond to data in Table 3 1.

PAGE 77

77 Figure 311 Maj ority rule consensus tree from the Bayesian analysis of the concatenated cyt b, Rag 2, and S7 data. Numbers on nodes are posterior probability values. A sterisks indicate posterior probability values of 1.0. Catalog numbers of voucher specimens or tissue numbers follow the species name and correspond to data in Table 3 1.

PAGE 78

78 Figure 312 Majority rule consensus tree from the likelihood analysis of the concatenated cyt b, Rag 2, and S7 data. Numbers on nodes are posterior probability values. A sterisks indicate posterior bootstrap support values of 100. Catalog numbers of voucher specimens or tissue numbers follow the species name and correspond to data in Table 3 1.

PAGE 79

79 CHAPTER 4 TAXONOMY OF THE AFRICAN CATFISH GENUS ANOPLOPTERUS (SILURIFORMES: AMPHILIIDAE) WITH DESCRIPTIONS OF TWO NEW SPECIES AND REDESCRIPTIONS OF ANOPLOPTERUS URANOSCOPUS CHIMARRHOGLANIS LEROYI AMPHILIUS BREVIDORSALIS, AMPHILIUS HARGERI AND AMPHILIUS CUBANGOENSIS Background Thomson and Page (2010) recognized the Amphilius uranoscopus species group for species of Amphilius that lack an epidermal fold at the base of the caudal fin, have 8 + 9 principal caudal fin rays a relatively elongate body ( usually 36 42 vs. 32 35 total vertebrae), and an anteriorly placed dorsal fin ( leading pterygiophore of the dorsal fin intercepts the vertebral column at the first, second or third post Weberian vertebra vs. at the fourth, fifth or sixth post Weberian vertebra). They recognized six species in the group: Amphilius uranoscopus (Pfeffer 1889) A. grandis Bo ulenger 1905, A. krefftii Boulenger 1911, A. cryptobullatus Skelton 1986, A. chalei Seegers 2008, and A. athiensis Thomson & Page 2010 from Kenya, Tanzania, and the upper Congo River basin (Figure 41). Two of the species, A. grandis and A. krefftii had p reviously been placed in the synonymy of A. uranoscopus but the study did not address the status of four other nominal species ( Chimarrhoglanis leroyi Vaillant 1897, A. hargeri Boulenger 1907, A. brevidorsalis Pellegrin 1919, and A. platychir var. cubango ensis Pellegrin 1936) that have long been placed in the synonymy of A. uranoscopus Following a phylogenetic analysis of Amphilius (Chapter 3 ), Amphilius was restricted to species that have an epidermal fold at the base of the caudal fin and 6+7 or 7+8 pr incipal caudal fin rays The name Anoplopterus Pfeffer 1889 was resurrected for species that lack an epidermal fold at the base of the caudal fin, have 8 + 9 principal caudal fin rays Additionally, the A. uranoscopus species group was not recovered as a

PAGE 80

80 monophyletic group. The two short bodied species of Anoplopterus ( A. laticaudatus and A. zairensis ) and one of the species with a posteriorly placed dorsal fin ( A. natalensis ) were consistently recovered as part of a clade that included all of the elongate bodied species. In this study, the status of the nominal species not dealt with by Thomson & Page ( 2010) are addressed by examining type specimens and more recently collected material from the river drainages from which they were described. A ll four speci es are valid, and they are redescribed along with A. uranoscopus Additionally, two new species from Tanzania, Anoplopterus n. sp. Ruvu from the Ruvu basin, and A. n. sp. Mangula from the Rufiji basin, are described. Material and Methods Measurements were made point to point with digital calipers, and data were recorded to tenths of a millimeter. Subunits of the head are presented as proportions of head length (HL). Head length and measurements of other body parts are given as proportions of standard length (SL). Use of the terms origin and insertion to designate, respectively, the most anterior and posterior points on the bases of all fins follows Cailliet et al. (1986). Counts and measurements were made on the left side of a specimen when possible and foll ow Thomson & Page (2010). Only specimens 40.0 mm SL or larger were measured, but counts were made on all specimens. For fin ray counts, numbers of unbranched soft rays are indicated by lower case Roman numerals, and branched soft rays by Arabic numerals. T he number of anterior unbranched rays in the anal fin is difficult to determine, and the counts were checked with radiographs whenever possible. Amphiliids typically have a small spinelet in front of the first unbranched dorsal fin ray. The spinelet is not included in the counts.

PAGE 81

81 Branchiostegal ray counts only include rays that articulate with the ceratohyal (anterohyal). Vertebrae were counted by means of radiographs; counts exclude the five Weberian vertebrae. The last abdominal vertebra is the last vertebra in which the distal end of the hemal spine lies anteriorly to the first anal fin pterygiophore. Posta bdominal counts include all vertebrae posterior to the abdominal vertebrae, with the ural centrum counted as one vertebra. The first dorsal pterygiophore intercept count is the number of vertebrae up to and including the vertebra opposite, or anterior to, the spine of the leading dorsal pterygiophore. Material examined is given under each species account and is listed by drainage followed by catalog number, country, locality, geographic coordinates and, in parentheses, the number of specimens and the size range in mm SL. Materials examined in this study are deposited in the following institutions: the American Museum of Natural History, New York, New York (AMNH), the Natural History Museum, London (BMNH), the California Academy of Sciences San Francisco, California (CAS), the Cornell University Vertebrate Collections, Ithaca, New York (CU), the Field Museum of Natural History Chicago, Illinois (FMNH), th e Harvard Museum of Comparative Zoology Cambridge, Massachusetts (MCZ), the Musum National d'Histoire naturelle, Paris, France (MNHN); the Royal Museum of Central Africa, Tervuren, Belgium (MRAC), the South African Institute for Aquatic Biodiversity Gra hamstown, South Africa (SAIAB), the Florida Museum of Natural History, Gainesville, Florida (UF), the Smithsonian Institution National Museum of Natural History Washington, DC (USNM), and the Universitt Hamburg, Biozentrum Grindel und Zoologisches Museum Ichthyology, Hamburg, Germany (ZMH).

PAGE 82

82 Synonymies include all references to the species from eastern Africa. References of species outside eastern Africa that give only a general distribution in eastern Africa are excluded. T he first page of the reference to the species and all figures are listed. If the species is also listed in a key on a separate page from the account, that page is also listed. The type of information in the reference is given followed by the locality for the species as given in the account. Additionally, any specimens on which that account is known to be based are listed. If the account is only based in part on the species, only the information that is applicable to that species is listed. Taxonomic Descriptions Anoplopterus uranoscopus Pfeffer, 1889 (Figure 4 2; Table 4 1 ) Anoplopterus uranoscopus Pfeffer 1889 : 16, Original description, Type locality: Ushonda and Monda, upper ranges of Wami River, Tanzania [ ZMH 11944 and ZMH 8401] ; Pfeffer, 1893: 161. Pl. 2, ( figs. 1 2 ), description, Bach bei Ushonda (Ungu) and Bder bei Mhonda; Pfeffer, 1896: 33, ( fig. 14 ), Bach bei Ushonda (Ungu) and Bder bei Mhonda; Boulenger, 1898a: 255 reference to original description; Boulenger, 1901: 447 distinctive characters; no distribution information; P oche, 1902a: 121; Boulenger, 1902: 41 in key; Amphilius uranoscopus Boulenger, 1905a: 64 in key; Boulenger, 1905b: 48 Upper ranges of the Wami R iver; Boulenger 1911: 357, ( fig. 277 ), Ushonda and Mhonda, Upper ranges of Wami River ; Harry, 1953: 187, Sy nonymy ; Copley, 1958: 99 (in part), upper ranges of the Wami River ; Bailey, 1969: 192, Wami River; Bernacek, 1980: 36 (in part), listed as present in the Wami drainage; Skelton, 1984: 41 (in part) Wami River; Skelton, 1986: 263 (in part), comparisons base d on type specimens ; Skelton, 1994: 113 (in part), Wami River; Seegers, 1996 : 192 (in part), ( figs. 136 137), type specimens, Lectotype designated. Material Examined Wami River basin: BMNH 2010.2.16.1 2, ex. UF 170723, (2: 53.5 70.7); CU 93740, Tanzania, Divue River above and below falls along road from Dumila to Turiani, altitude 374 m, 610'27"S, 3735' 00"E (21: 31.7 97.8); CU 93741, Tanzania, Mbulumi

PAGE 83

83 River at bridge in Turiani on road from Dumila to Turiani, altitude 377 m, 608'38"S, 3735' 47"E (5: 61.0 109.4); CU 93742, Tanzania, Wami River at rapids above bridge on road from Chalinze to Segera, altitude 60 m, 614'42"S, 3823' 00"E (1: 81.4); MRAC 201008P 1 2, ex. UF 170723, (2: 58.6 68.3); SAIAB 87472 ex. UF 170723, (2: 49.0 62.7); UF 170716, same data as CU 93741, (7: 26.3 108.5); UF 170718, same data as CU 93742, (2: 43.3 46.8); UF 170723, same data as CU 93740, (19: 31.1 96.2) ; ZMH 8401, Tanzania, Bad bei Ushonda (Ungu) ca. 6 20S, 37 10E (paralectotype, photographs and x rays examined) ; ZMH 11944, same data as ZMH 8401, (lectotype, photographs and x rays examined) Rufiji River basin: CU 93748 Tanzania, 12 Bridges River at overhead bridge for train on road from Mikumi to Ifakara, altitude 419 m 727'53"S, 3700' 52"E (7: 51.8 82.0 ), U F 170720, same data as CU 93748, (7: 56.7 91.0). Diagnosis Diagnostic char acters are summarized in Tables 4 2 and 4 3 Anoplopterus uranoscopus is diagnosed from A. n. sp. Ruvu A. n. sp. Mangula A. krefftii and A. cryptobullatus by having a forked caudal fin (vs. emarginated caudal fin) and from A. n. sp. Ruvu A. athiensis, and A. cubangoensis by having a body color that is not finely spotted or mottled (vs. body finely spotted in A. n. sp. Ruvu and A. athiensis and mottled in A. cubangoensis ). It is d iagnosed from A. brevidorsalis A. cubangoensis A. n. sp. Ruvu and A. chalei by having a wider interorbital width (25.228.1% HL vs. 17.925.3% HL) and from A. leroyi by having a smaller head height (10.0 11.9% SL vs. 11.713.8% SL). Anoplopterus uranosc opus is diagnosed from A leroyi A. hargeri, A. n. sp. Ruvu A. n. sp. Mangu l a and A. krefftii by having a longer body with 38 to 40 vertebrae (vs. usually 37 or fewer vertebrae). It is diagnosed from A. hargeri and A. n. sp. Ruvu by

PAGE 84

84 its first dorsal pte rygiophore intercept count (1 vs. usually 2) and from A. athiensis and A. grandis by having distinct pale patches at origin and insertion of dorsal fin (vs. pale patches absent). Anoplopterus uranoscopus is further diagnosed from A. n. sp. Mangula by having a longer predorsal lengt h (35.038.6% SL vs. 31.634.9 % SL) and from A. cubangoensis and A. chalei by having a deeper body (body depth at anus 12.716.3 % SL vs. 8.9 12.2 % SL). It is further diagnosed from A. cubangoensis A. n. sp. Mangula and A. chalei by having a shorter caudal peduncle length (15.117.9% SL vs. 17.7 21.9% SL) and from A. cubangoensis and A. chalei by having a deeper caudal peduncle (10.914.3 SL vs. 7.110.9 % SL). Anoplopterus uranoscopus is further diagnosed from A. brevidorsalis and A. n. sp. Ruvu by usually having 9 to 11 gill rakers on the first arch (vs. usually having 6 to 8, rarely 9) and from A. n. sp. Ruvu by having 6 or 7 branched anal fin rays (vs. usually 5, rarely 6). It is further diagnosed from A. n. sp. Ruvu by having l onger maxillary barbels (69.599.1% HL vs. 37.349.6 % HL). Description Morphometric d ata in Table 4 1 Body elongate, ventral profile flattened ventrally to anal fin base, then tapered dorsally to end of caudal peduncle. Dorsal profile rising gently from tip of snout to origin of dorsal fin, then nearly horizontal to end of caudal peduncle. Greatest body depth at dorsal fin origin. Caudal peduncle laterally compressed. Anus and urogenital openings located just posterior to base of pelvic fin, closer to ins ertion of pelvic fin than to origin of anal fin. Skin smooth. Lateral line complete, extending from dorsal edge of opercular cavity to base of caudal fin. Total vertebrae 38 (9) or 39 (5). Abdominal vertebrae 21 (1), 22 (8), or 23 ( 5 ). Caudal vertebrae 15 (1), 16 11), or 17 (2). First dorsal pterygiophore intercept count 1 (14).

PAGE 85

85 Head and anterior part of body depressed and broad. Head wedgeshaped in lateral view. Snout broad, moderately pointed when viewed from above. Head becoming wider from tip of snout to pectoral fin base. Branchiostegal membranes moderately joined at isthmus forming a V shaped or deeply concave connection. Mouth broad, gently curved, subterminal. Lips moderately fleshy, papillate. Rictal lobe large and papillate. Anterior portion of premaxillary tooth band exposed with mouth closed. Premaxillary tooth patches joined, forming U shaped band, with short conical teeth. Dentary teeth short and conical, tooth patches forming U shaped band, separated medially. Three pairs of simple, tapered c ircumoral barbels. Maxillary barbel large, fleshy and flattened with pointed tip; barbel extending posterolaterally from corner of mouth to just short of pectoral fin base. Outer mandibular barbel thin with pointed tip, origin at posterior corner of lower jaw, extending to edge of branchiostegal membrane. Inner mandibular barbel originates anterolaterally of inner mandibular barbel, extending to edge of branchiostegal membrane. Branchiostegal membrane with 8 (43), or 9 (10) rays. Gill rakers on first epibra nchial 2 (6), 3 (48) or 4 (1); rakers on first ceratobranchial 5 (3), 6 (21), 7 (25) or 8 (6); total gill rakers on first arch 8 (5), 9 (23), 10 (20), or 11 (7). Eyes small, positioned dorsolaterally approximately midway between tip of snout and posterior margin of operculum. Horizontal diameter of eye slightly wider than vertical diameter. Eye without free orbit; covered with skin confluent with dorsal surface of head. Anterior and posterior nares with prominent tubular rims; nares separate but relatively close to each other. Posterior nare located about midway between eye and tip of snout.

PAGE 86

86 Dorsal fin origin at point over tip of pectoral fin. Dorsal fin with i,6 (57) rays, and fin margin straight. Pectoral fin with i,8 (1), i,9 (5) or i,10 (51) rays ; unbranched ray greatly thickened. Pectoral fin with four or five innermost rays progressively shorter making posterior fin margin rounded. Pelvic fin inserted posteriorly to dorsal f in base. Pelvic fin with i,5 (57) rays with first ray unbranched and greatly thickened. Pelvic fin with straight posterior margin. Adiposefin base longer than anal fin base; origin anterior to origin of anal fin base; fin extending past anal fin insertion. Margin strongly convex with sharply rounded edge, not deeply incised posteriorly. Caudal fin emarginate, with i,7,7,i (1), or i,7,8,i (55) principal rays. Anal fin with short base; origin posterior to origin of adiposefin base; fin with ii,6 (27), ii,7 (4), iii,5 (4), iii,6 (20), or iii,7 (2) rays. Anal fin margin almost straight. Coloration Dorsal and lateral surfaces of head and body dark grey with diffuse black stripe along side. Ventral region light brown. Light dorsal saddles at origin and insertion of dorsal fin and adipose fin. Dorsal, adipose, caudal, and anal fins dark grey. Pectoral and pelvic fins positioned horizontally with upper surfaces dark grey and l ower surfaces light yellow. Maxillary and mandibular barbels grey. Caudal fin with dark crescent shaped band at base and broad diffuse band on rays. Juvenile coloration similar to that of adult Distribution Known only from the Wami River and Rufiji River basins Tanzania (Figure 4 3 ). Anoplopterus leroyi (Vaillant 1897) (Figure 4 4; Table 4 4 ) Chimarrhoglanis leroyi Vaillant 1897: 82, Original description, Type locality : Torrent of Mrogoro at Zanguebar, Oukami, East Africa [ MNHN 18970003] ; Boulenger, 1898a: 254 as synonym of Amphilius platychir type locality information; Boulenger, 1898b : 4 as synonym of Amphilius platychir ; Poche, 1902a: 121 as synonym of Anoplopterus platychir ; Poche, 1902b: 211, as synonym of

PAGE 87

87 Anoplopterus platychir ; Pellegrin, 1905: 177 type information; Bertin & Estve, 1950: 35 type information; Harry, 1953: 186, as synonym of Amphilius platychir Amphilius grandis (non Boulenger): Bailey, 19 69: 192 Ruvu River Amphilius platychir (non Gnther): Boulenger, 1905d: 48 (in part), Headwaters of the Ruvu River. Amphilius uranoscopus (non Pfeffer): Bernacek, 1980: 36 (in part), listed as present in the Ruvu drainage; Skelton, 1984: 45 (in part), in synonymy, type information, Ruvu River; Skelton, 1986: 263 (in part), comparisons based on type specimens ; Skelton, 1994: 113 (in part), listed as present in Ruvu drainage. Material Examined Ruvu River basin: CU 93725, Tanzania, Kibundi River along road from Kipera to Mgeta, altitude 633m 659'08"S, 3733'31"E (11: 27.752.0); CU 93727, Tanzania, Ruvu River at bridge on road from Mikese to Mtambo, altitude 96 m 701'24"S, 3748'36 "E (5: 36.2 71.6 ); FMNH 111682 Tanzania, Uluguru Mts., Uluguru North F orest Reserve, 3 km W, 1.3 km N Tegetero, 1345 m 655'45"S, 3742'20 "E (1: 55.7 ); FMNH 111683, Tanzania, Uluguru Mts., Uluguru North Forest Reserve, 5.1 km W, 2.3 km N Tegetero, 1535 m 655'12"S, 3741'00"E (3: 26.4 52.4 ); MNHN 18970003 Torrent of Mrog oro at Zanguebar, Oukami, East Africa, ca. 646'S, 3744' E (1: 117.3, holotype); UF 84882 Tanzania, Southern side of Uluguro Mountains, about 1 km N of Ruvu River, S of Kibungo Village [on Rd from Mi kese to Matambo], Kimboza Forest, ca. 702S,3748' E (7: 21.9 40.4); UF 84885 same locality as UF 84882 (1: 9.2); UF 170711, same data as CU 93727, (5: 39.9 76.5 ); UF 170717, same data as CU 93725, (12: 31.2 116.1). Diagnosis Diagnostic characters are summarized in Tables 4 2 and 4 3 Anoplopterus leroyi is d iagnosed from A. n. sp. Ruvu A. n. sp. Mangula and A. krefftii by having a forked

PAGE 88

88 caudal fin (vs. emarginated caudal fin) and from A. n. sp. Ruvu A. athiensis, and A. cubangoensis by having a body color that is not finely spotted or mottled (vs. body fi nely spotted in A. n. sp. Ruvu and A. athiensis and mottled in A. cubangoensis ). It is diagnosed from A. brevidorsalis A. cubangoensis A. n. sp. Ruvu and A. chalei by having a wider interorbital width (25.830.2 % HL vs. 17.925.3 % HL) and from A. uranoscopus by having a larger head height (11.713.8 % SL vs. 10.011.9 % SL). Anoplopterus leroyi is diagnosed from A. uranoscopus, A. n. sp. Ruvu A. krefftii, A. chalei and A. athiensis by its total vertebrae count (usually 37 vs. usually 38 or more in A. ur anoscopus and A. athiensis and usually 36 or fewer in A. n. sp. Ruvu and A. krefftii ). It is diagnosed from A. hargeri and A. n. sp. Ruvu by its first dorsal pterygiophore intercept count (1 vs. usually 2) and from A. athiensis and A. grandis by having distinct pale patches at origin and insertion of dorsal fin (vs. pale patches absent). Anoplopterus leroyi is further diagnosed from A. hargeri and A. cryptobullatus by having a shorter predorsal length (34.736.8% SL vs. 37.3 38.4 % SL) and from A. cubangoensis A. n. sp. Mangula, and A. chalei by having a deeper body (body depth at anus 13.016.0% SL vs. 8.9 13.1% SL). It is further diagnosed from A. cubangoensis A. n. sp. Mangula, and A. chalei by having a shorter caudal peduncle length (14.917.8% SL vs. 17.721.9% SL) and from A. chalei by having a deeper caudal peduncle (10.013.4 SL vs. 7.1 9.7 % SL). Anoplopterus leroyi is further diagnosed from A. n. sp. Ruvu by usually having 8 to 10 gill rakers on the first arch (vs. usually having 6 or 7, rarely 8) and by having 6 or 7 branched anal fin rays (vs. usually 5, rarely 6). It is further diagnosed from A. n. sp. Ruvu by having longer maxillary barbels (69.094.8 % HL vs. 37.349.6 % HL).

PAGE 89

89 Description Morphometric d ata in Table 4 4 Body elongate, ventral pr ofile flattened ventrally to anal fin base, then tapered dorsally to end of caudal peduncle. Dorsal profile rising gently from tip of snout to origin of dorsal fin, then nearly horizontal to end of caudal peduncle. Greatest body depth at dorsal fin origin. Caudal peduncle laterally compressed. Anus and urogenital openings located just posterior to base of pelvic fin, closer to insertion of pelvic fin than to origin of anal fin. Skin smooth. Lateral line complete, extending from dorsal edge of opercular cavi ty to base of caudal fin. Total vertebrae 36* (3) or 37* (13). Abdominal vertebrae 19 (1), 20* (4), 21 (10) or 22 (1). Caudal vertebrae 15 (1), 16* (13), or 17 (1). First dorsal pterygiophore intercept count 1 (15) or 2* (1). Head and anterior part of body depressed and broad. Head wedgeshaped in lateral view. Snout broad, moderately pointed when viewed from above. Head becoming wider from tip of snout to pectoral fin base. Branchiostegal membranes moderately joined at isthmus forming a V shaped or deeply concave connection. Mouth broad, gently curved, subterminal. Lips moderately fleshy, papillate. Rictal lobe large and papillate. Anterior portion of premaxillary tooth band exposed with mouth closed. Premaxillary tooth patches joined, forming U shaped ban d, with short conical teeth. Dentary teeth short and conical, tooth patches forming U shaped band, separated medially. Three pairs of simple, tapered circumoral barbels. Maxillary barbel large, fleshy and flattened with pointed tip; barbel extending poster olaterally from corner of mouth to just short of pectoral fin base. Outer mandibular barbel thin with pointed tip, origin at posterior corner of lower jaw, extending to edge of branchiostegal membrane. Inner

PAGE 90

90 mandibular barbel originates anterolaterally of inner mandibular barbel, extending to edge of branchiostegal membrane. Branchiostegal membrane with 8* (23), 9 (20) or 10 (2) rays. Gill rakers on first epibranchial 2 (16), 3 (23 ) or 4 (5 ); rakers on first ceratobranchial 4 (1), 5* (8), 6 ( 24) or 7 (11) ; total gill rakers on first arch 7 (1), 8 (17), 9 ( 17), or 10 (9). Eyes small, positioned dorsolaterally approximately midway between tip of snout and posterior margin of operculum. Horizontal diameter of eye slightly wider than vertical diameter. Eye w ithout free orbit; covered with skin confluent with dorsal surface of head. Anterior and posterior nares with prominent tubular rims; nares separate but relatively close to each other. Posterior nare located about midway between eye and tip of snout. Dors al fin origin at point over tip of pectoral fin. Dorsal fin with i,6 (46 ) rays, and fin margin straight. Pectoral fin with i,9* (19) or i,10 (27 ) rays ; unbranched ray greatly thickened. Pectoral fin with four or five innermost rays progressively shorter making posterior fin margin rounded. Pelvic fin inserted posteriorly to dorsal f in base. Pelvic fin with i,5 (46) rays with first ray unbranched and greatly thickened. Pelvic fin with straight posterior margin. Adiposefin base longer than anal fin base; ori gin anterior to origin of anal fin base; fin extending past anal fin insertion. Margin strongly convex with sharply rounded edge, not deeply incised posteriorly. Caudal fin emarginate, with i,6,7,i ( 2), i,7,7,i (5 ), or i,7,8,i (39 ) principal rays. Anal fi n with short base; origin posterior to origin of adiposefin base; fin with ii,6 (1), ii,7 (1), iii,6 (34 ) or iii,7 (10 ) rays. Anal fin margin almost straight.

PAGE 91

91 Coloration Dorsal and lateral surfaces of head and body brown with diffuse dark brown stripe along side Ventral region light brown. Light dorsal saddles at origin and insertion of dorsal fin and adipose fin. Dorsal, adipose, caudal, and anal fins brown. Pectoral and pelvic fins positioned horizontally with upper surfaces brown and lower surfaces l ight yellow. Maxillary and mandibular barbels brown. Caudal fin with dark crescent shaped band at base and broad diffuse band on rays Juvenile coloration similar to that of adult Distribution Known only from the Ruvu River basin, Tanzania (Figure 4 3 ). A noplopterus hargeri ( Boulenger 1907) (Figure 4 5; Table 4 5 ) Amphilius hargeri Boulenger 1907: 488, Original description, Type locality: Ruo River, Mlanji, British Central Africa [Malawi] [ BMNH 1907.10.14.8] ; Boulenger 1911: 358, ( fig. 278 ), d escription; in key sources of Ruo River, Mlanji; Boulenger, 1917: 432, diagnosis from A. natalensis ; Pellegrin, 1919: 401, diagnosis from A brevidorsalis ; Pellegrin 1933: 172 A. kivuensis ; Harry, 1953: 189 s ynonymy Amphilius platychir (non Gnther): Boulenger 1911 : 357, ( fig. 277 ), Ruo River; Tweddle & Willoughby, 1976: 12, Listed from Shire Valley Game Reserves ; Tweddle & Willoughby, 1979: 18, Shire River system; Amphilius uranoscopus (non Pfeffer): Skelton, 1984: 67 (in part), type specimens, Ruo River; Skelt on, 1994 : 129 (in part), Lower Zambezi drainage. Holotype. BMNH 1907.10.14.8 Ruo River, Mlanji, British Central Africa [Malawi] ca. 1602'S, 3530'E (1: 48.8). Material examined Lower Zambezi River basin: BMNH 1893.11.15.8890 Malawi, Mlanji, River Ruo ca. 1 602'S, 353 0' E (3: 22.432.1); BMNH 1907.10.14.8, British Central Africa [Malawi] Mlanji, Ruo River, ca. 1602'S, 3530'E (1: 48.8, holotype); BMNH 1921.9.6.5463, Malawi, Mlanji, Nswadzi River, Cholo, ca. 1 604 'S, 35 08 'E (9: 56.7106.1); BMNH 1960.12.30.12, Malawi, Mlanji, ca. 1 602'S, 3530' E (1: 49.8); BMNH

PAGE 92

92 1978.8.3.2041 Malawi ca. 1559N, 3527E (1: 72.8); SAIAB 34319 Malawi Maperera River, Near Chikwawa ca. 1 605'40S, 34 54' 30 E (1: 61.5); SAIAB 34327, Malawi Ntwadzi River, Thyolo road, Satemba Estate, ca. 1 604'10S, 35 05' 30 E (13: 38.1 78.1); SAIAB 34332 Malawi Nachidwa River, Mujiwa, ca. 1 603'S, 35 47'E (22: 43.0128.9); SAIAB 34344 Malawi Lichenya River, Mini Mini Estate ca. 1 603'30S, 3535 00 E (19: 26.674.8 ); SAIAB 34349 Malawi Lichenya River, 5km above confluence Ruo River ca. 1 605'40S, 35 29' 43 E (11: 26.71 77.28); SAIAB 34354 Malawi, Mlanji, Ruo River at bridge, ca. 1 604'36S, 3540' 17E (1: 59.4); SAIAB 34356 Malawi, Likabula River, tributary of Ruo River at bridge on Mulange Phalumbe road ca. 1 555'S, 353 0' E (4: 40.3 62.8 ); SAIAB 51955 Malawi, Ruo River at Sankhulani ca. 1 627'S, 3516 'E (1: 76.9 ); USNM 86637 Malawi Nswadgi River, Cholo ca. 1 604'S, 3508' E (1: 66.3). Rovuma Basin: SAIAB 35795 Ma lawi Mulanje, Mandimba River at Rte 131 crossing ca. 1 424'S, 3536' E (5: 57126). Diagnosis Diagnostic characters are summarized in Tables 4 2 and 4 3 Anoplopterus hargeri is diagnosed from A. uranoscopus, A. leroyi, A. brevidorsalis, A. cubangoensis, A. n. sp. Mangula A. krefftii, A. chalei, and A. athiensis by its first dorsal pterygiophore intercept count of usually 2 (vs. usually 1). It is diagnosed from A. n. sp. Ruvu, A. n. sp. Mangula, and A. krefftii by having a forked caudal fin (vs. emarginat ed caudal fin) and from A. n. sp. Ruvu by having a body coloration that is not heavily spotted (vs. body h eavily spotted, with spots usually coalesced to form vermiculations ). Anoplopterus hargeri is diagnosed from A. brevidorsalis, A. cubangoensis, A. n. sp. Ruvu and A. chalei by having a wider interorbital width (25.721.7% HL vs. 17.9 25.3 % HL) and from A. uranoscopus and A. chalei by having a greater head height (12.313.2% SL vs. 9.7 -

PAGE 93

93 12.4% SL). Anoplopterus hargeri is diagnosed from A. uranoscopus, A. chalei, and A. athiensis by having a shorter body with 36 to 37 vertebrae (vs. usually 38 or more vertebrae) and from A. athiensis and A. grandis by having distinct pale patches at origin and insertion of dorsal fin (vs. pale patches absent). Anoplopterus hargeri is further diagnosed from A. leroyi, A, n. sp. Mangula, and A. chalei by having a longer predorsal length (37.338.4% SL vs. 31.636.8% SL) and from A. cubangoensis and A. chalei by having a deeper body (body depth at anus 13.014.4 % SL vs. 8.9 12 .2 % SL). It is further diagnosed from A. cubangoensis, A. n. sp. Mangula and A. chalei by having a shorter caudal peduncle length (14.416.8% SL vs. 17.721.9% SL) and from A. cubangoensis and A. chalei by having a deeper caudal peduncle (11.813.5 vs. 7. 1 10.9% SL). Amphilius hargeri is further diagnosed from A. n. sp. Ruvu by usually having 8 to 10 gill rakers on the first arch (vs. usually having 6 or 7, rarely 8) and by usually having 6 branched anal fin rays (vs. usually 5, rarely 6). It is further di agnosed from A. n. sp. Ruvu by having longer maxillary barbels (62.295.9 % HL vs. 37.3 49.6 % HL). Description Morphometric d ata in Table 4 5 Body elongate, ventral profile flattened ventrally to anal fin base, then tapered dorsally to end of caudal pedunc le. Dorsal profile rising gently from tip of snout to origin of dorsal fin, then nearly horizontal to end of caudal peduncle. Greatest body depth at dorsal fin origin. Caudal peduncle laterally compressed. Anus and urogenital openings located just posterior to base of pelvic fin, closer to insertion of pelvic fin than to origin of anal fin. Skin smooth. Lateral line complete, extending from dorsal edge of opercular cavity to base of caudal fin. Total vertebrae 36 (2) or 37* (5). Abdominal vertebrae 20 (3) o r 21* (4). Caudal vertebrae 15 (1), 16* (4), or 17 (2). First dorsal pterygiophore intercept count 2 (6) or 3* (1).

PAGE 94

94 Head and anterior part of body depressed and broad. Head wedgeshaped in lateral view. Snout broad, moderately pointed when viewed from abov e. Head becoming wider from tip of snout to pectoral fin base. Branchiostegal membranes moderately joined at isthmus forming a V shaped or deeply concave connection. Mouth broad, gently curved, subterminal. Lips moderately fleshy, papillate. Rictal lobe l arge and papillate. Anterior portion of premaxillary tooth band exposed with mouth closed. Premaxillary tooth patches joined, forming U shaped band, with short conical teeth. Dentary teeth short and conical, tooth patches forming U shaped band, separated m edially. Three pairs of simple, tapered circumoral barbels. Maxillary barbel large, fleshy and flattened with pointed tip; barbel extending posterolaterally from corner of mouth to just beyond pectoral fin base. Outer mandibular barbel thin with pointed ti p, origin at posterior corner of lower jaw, extending to pectoral fin base. Inner mandibular barbel originates anterolaterally of inner mandibular barbel, extending to edge of branchiostegal membrane. Branchiostegal membrane with 7 (16), 8* (25), or 9 (1) rays. Gill rakers on first epibranchial 2 (5) or 3* (36 ); rakers on first ceratobranchial 5 (1), 6 (19 ), 7 (10) or 8 (1); total gill rakers on first arch 7 (1), 8* (2 ), 9 (29), 10 (8), or 11 (1). Eyes small, positioned dorsolaterally approximately midway between tip of snout and posterior margin of operculum. Horizontal diameter of eye slightly wider than vertical diameter. Eye without free orbit; covered with skin confluent with dorsal surface of head. Anterior and posterior nares with prominent tubular r ims; nares separate but relatively close to each other. Posterior nare located closer to tip of snout than eye.

PAGE 95

95 Dorsal fin origin at point over tip of pectoral fin. Dorsal fin with i,5 (1) or i,6 ( 41 ) rays, and fin margin straight. Pectoral fin with i,8 ( 1), i,9* (20), i,10* (19) or i,11 (1) rays ; unbranched ray greatly thickened. Pectoral fin with four or five innermost rays progressively shorter making posterior fin margin rounded. Pelvic fin inserted posteriorly to dorsal f in base. Pelvic fin with i,5 ( 42) rays with first ray unbranched and greatly thickened. Pelvic fin with straight posterior margin. Adiposefin base longer than anal fin base; origin anterior to origin of anal fin base; fin extending past anal fin insertion. Margin strongly convex with sharply rounded edge, deeply incised posteriorly. Caudal fin forked, with i,7,7,i (2), or i,7,8,i* (40 ) principal rays. Anal fin with short base; origin posterior to origin of adiposefin base; fin with ii,5 (2 ), ii,6 (31), ii,7 (2), iii,5 (1 )*, iii,6 (2), or iv,6 (1) rays. Anal fin margin almost rounded. Coloration Dorsal and lateral surfaces of head and body dark brown, usually with black blotches or spots. Ventral region light brown. Light dorsal saddles at origin and insertion of dorsal fin and adipose fin. Dorsal, adipose, caudal, and anal fins brown. Pectoral and pelvic fins positioned horizontally with upper surfaces dark brown and may have black spots. L ower surfaces light yellow. Maxillary and mandibular barbels dark brown. Caudal fin with dark cres cent shaped band at base and broad diffuse band on rays. Juvenile coloration similar to that of adult but usually with larger black blotches on body and may appear mottled. Distribution Known from the Shire River system of the Lower Zambezi River basin and Rouvuma River basin in Malawi and Mozambique (Figure 4 6 ).

PAGE 96

96 Anoplopterus brevidorsalis ( Pellegrin 1919) (Figure 4 7; Table 4 6 ) Amphilius brevidorsalis, Pellegrin 1919: 399, Original description, Type locality: Revue River at Manica, Mozambique [ MNHN 19190488] ; Bertin & Estve, 1950: 36, type information. Amphilius grandis (non Boulenger): Crass, 1960: 443 (in part), Mozambique; Jubb, 1961: 120 (in part), Buzi River basin; Amphilius natalensis (non Boulenger); Barnard, 1942: 257. Amphilius platychir ( non Gnther): ?BellCross, 1973: 9, 14, Buzi/Revue system ( could be A. brevidorsalis or A. natalensis ) ; BellCross & Jubb, 1973: 2. Amphilius uranoscopus (non Pfeffer): Skelton, 1984: 67, type specimen, Revue River. Material examined Buzi River basin: MN HN 19190488 Mozambique, Revue River at Manica, ca. 1858'S, 3303' E ( 1: 42.8 holotype); SAIAB 53691 Mozambique, Bonde River at causeway ca. 1910'S, 3305' E ( 13: 26.888.2); SAIAB 53695 same locality as SAIAB 53691, (3: 37.4 80.3 ); SAIAB 53698 Mozam bique, Bonde River, Inflow to Lake Chicamba ca. 1910'S, 3303' E ( 9: 36.0 48.4); SAIAB 53712 Mozambique, Mupandeia River at bridge, ca. 1910'S, 3258' E ( 58: 27.4 92.7); SAIAB 60023 Mozambique, Small River Below Lake Chicamba Wall ca. 1920'S, 3256' E ( 36: 23.4 67.5 ); SAIAB 60024 Mozambique, Pool Below Chicamba Wall, near Chaine ca. 1911'S, 3254' E ( 1: 37.5 ); SAIAB 60025 Mozambique, Nyamanguena River at bridge ca. 1917'S, 3302' E ( 10: 29.545.4 ); SAIAB 67527 Mozambique, v ery small tributary of Mussapa Grande, ca. 1939'S, 3309' E ( 4: 80.2 111.3); SAIAB 67550 Mozambique, Mussapa Pequena, near park 'gate' ca. 1935'S, 3305' E ( 1: 42.0 ); SAIAB 67576 Mozambique, Mukuru Stream on Chikukwe Road, 1903'06S, 3304'09 E ( 3: 40.5 69.2 ); SAIAB 67582 Moz ambique 1942'18S, 3258'35 E ( 3: 71.3 89.5 ); SAIAB 67586 Mozambique Nyabawa Stream

PAGE 97

97 1942'19S, 3301'29 E ( 2: 64.2 66.0 ); SAIAB 67618 Mozambique Muzuma River 1947'19S, 3320'55 E ( 1: 54.5 ); SAIAB 85585 Mozambique, Bonde River causeway above Lak e Chicamba 1919'23S, 3254'15 E ( 15: 24.174.4); SAIAB 85593 Mozambique, 1910'12S, 3253'06 E ( 2: 53.7 68.5); SAIAB 85659 Mozambique, Msika River at Bridge, about 10 km WSW of Manica 1859'14S, 3302'57 E ( 4: 45.6 62.5 ); SAIAB 85740 same locality as SAIAB 85585 (8: 29.8 58.1 ); SAIAB 85828 Mozambique, Zonue River Inflow ca. 1909'S, 3254' E ( 1: 55.2 ); SAIAB 85862 same locality as SAIAB 85585 (20: 28.6 67.0 ). Pungue River basin: BMNH 1973.9.27.23 Zimbabwe, Nyazengu stream, above Pungwe River ca. 1820'S, 3250' E ( 2: 77.8 89.8); USNM 296983, Mozambique, Beina Di strict, ca. 1942'S, 3454' E ( 1 ). Diagnosis Diagnostic characters are summarized in Tables 4 2 and 4 3 Anoplopterus brevidorsalis is diagnosed from A. n. sp. Ruvu A. n. sp. Mangula and A. krefftii by having a forked caudal fin (vs. emarginated caudal fin) and from A. n. sp. Ruvu by having a body coloration that is not heavily spotted (vs. body heavily spotted, with spots usually coalesced to form vermiculations ). It is diagnosed from A. uranoscopus A. leroyi and A. hargeri by having a narrower interorbital width (22.225.3 % HL vs. 25.2 31.7% HL). Anoplopterus brevidorsalis is diagnosed from A. n sp Ruvu and A. krefftii by having a longer body with 37 to 38 total vertebrae (vs. usu ally 36 or less). It is diagnosed from A. hargeri and A. n. sp. Ruvu by its first dorsal pterygiophore intercept count (1 vs. usually 2) and from A. athiensis and A. grandis by having distinct pale patches at origin and insertion of dorsal fin (vs. pale pa tches absent). Anoplopterus brevidorsalis is further diagnosed from A. n. sp. Mangula by having a longer predorsal length (35.938.9% SL vs. 31.634.9% SL) and from A. chalei by having a deeper body

PAGE 98

98 (body depth at anus 12.514.0% SL vs. 8.9 10.8 % SL). It i s further diagnosed from A. n. sp. Mangula and A. chalei by having a shorter caudal peduncle length (16.417.3% SL vs. 18.2 21.9% SL) and from A. chalei and A. cubangoensis by having a deeper caudal peduncle (10.713.7 vs. 7.1 10.9% SL). Anoplopterus brevi dorsalis is further diagnosed from A. uranoscopus, A. hargeri, A. cubangoensis, and A. athiensis by usually having 7 or 8 gill rakers on the first arch (vs. usually 9 to 10, rarely 7 or 8) and from A. n. sp. Ruvu by usually having 6 or 7 branched anal fin rays (vs. usually 5, rarely 6). It is further diagnosed from A. n. sp. Ruvu by having longer maxillary barbels (64.986.0 % HL vs. 37.349.6 % HL). Description Morphometric d ata in Table 4 6 Body elongate, ventral profile flattened ventrally to anal fin ba se, then tapered dorsally to end of caudal peduncle. Dorsal profile rising gently from tip of snout to origin of dorsal fin, then nearly horizontal to end of caudal peduncle. Greatest body depth at dorsal fin origin. Caudal peduncle laterally compressed. A nus and urogenital openings located just posterior to base of pelvic fin, closer to insertion of pelvic fin than to origin of anal fin. Skin smooth. Lateral line complete, extending from dorsal edge of opercular cavity to base of caudal fin. Total vertebrae 37* (1) or 38 (6). Abdominal vertebrae 21* (1) or 22 (6). Caudal vertebrae 16 (7). First dorsal pterygiophore intercept count 0* (1) or 1 (6). Head and anterior part of body depressed and broad. Head wedgeshaped in lateral view. Snout broad, moderately pointed when viewed from above. Head becoming wider from tip of snout to pectoral fin base. Branchiostegal membranes moderately joined at isthmus forming a V shaped or deeply concave connection.

PAGE 99

99 Mouth broad, gently curved, subterminal. Lips moderately fleshy, papillate. Rictal lobe large and papillate. Anterior portion of premaxillary tooth band exposed with mouth closed. Premaxillary tooth patches joined, forming U shaped band, with short conical teeth. Dentary teeth short and conical, tooth patches formi ng U shaped band, separated medially. Three pairs of simple, tapered circumoral barbels. Maxillary barbel large, fleshy and flattened with pointed tip; barbel extending posterolaterally from corner of mouth to just short of pectoral fin base. Outer mandibu lar barbel thin with pointed tip, origin at posterior corner of lower jaw, extending to edge of branchiostegal membrane. Inner mandibular barbel originates anterolaterally of inner mandibular barbel, extending to edge of branchiostegal membrane. Branchiost egal membrane with 7 (2), 8 (28), or 9 (1) rays. Gill rakers on first epibranchial 2 (14) or 3* (18); rakers on first ceratobranchial 3 (1), 4 (1), 5* (20), or 6 (10); total gill rakers on first arch 6 (1), 7 (8), 8* (20), or 9 (3). Eyes small, positioned dorsolaterally approximately midway between tip of snout and posterior margin of operculum. Horizontal diameter of eye slightly wider than vertical diameter. Eye without free orbit; covered with skin confluent with dorsal surface of head. Anterior and post erior nares with prominent tubular rims; nares separate but relatively close to each other. Posterior nare located about midway between eye and tip of snout. Dorsal fin origin at point over tip of pectoral fin. Dorsal fin with i,6 (32) rays, and fin margi n straight. Pectoral fin with i,8 (1), i,9 (7), or i,10* (18) rays ; unbranched ray greatly thickened. Pectoral fin with four or five innermost rays progressively shorter making posterior fin margin rounded. Pelvic fin inserted posteriorly to dorsal f in bas e.

PAGE 100

100 Pelvic fin with i,5 (32) rays with first ray unbranched and greatly thickened. Pelvic fin with straight posterior margin. Adiposefin base longer than anal fin base; origin anterior to origin of anal fin base; fin extending past anal fin insertion. Marg in strongly convex with sharply rounded edge, not deeply incised posteriorly. Caudal fin emarginate, with i,6,7,i (1) or i,7,8,i* (31) princ ipal rays. Anal fin with short base; origin posterior to origin of adiposefin base; fin with ii,6 (15), ii,7 (7), i ii,5* (3), or iii,6 (7) rays. Anal fin margin almost straight. Coloration Dorsal and lateral surfaces of head and body light brown/yellow, usually with black blotches or spots. Ventral region light brown. Light dorsal saddles at origin and insertion of dor sal fin and adipose fi n. Dorsal, adipose, caudal, and anal fins light brown/yellow. Pectoral and pelvic fins positioned horizontally with upper surfaces light brown/yellow and lower surfaces light yellow. Maxillary and mandibular barbels brown. Caudal fin with dark crescent shaped band at base and broad diffuse band on rays. Juvenile coloration similar to that of adult but usually with larger black blotches on body and may appear mottled. Distribution Known only from the Buzi River and Pungue River basins Mozambique (Figure 4 6 ). Anoplopterus cubangoensis ( Pellegrin 1936) (Figure 4 8; Table 4 7 ) Amphilius platychir var. cubangoensis Pellegrin 1936: 56, Original description, Type locality: Cubango [Okavango] River, near Vila da Ponte, Angola [ MHNG 856.86, MNHN 1936 0101 to 0103, MRAC 138769, NMBA 5213] ; Harry, 1953 (in part) : 186, in synonymy of A. platychir ; Weber, 1998: 9, type information.

PAGE 101

101 Amphilius platychir (non Gnther): Jubb & Gaigher, 1973: 10, 18, 20 l isted as occurring in the Okavango drainage complex, Botswana; Balon, 1974: 653, listed from edge of Victoria Falls [ROM 31814, ROM 31873] Amphilius platychir cubangoensis BellCross & Jubb, 1973: 1 in synonymy of A. platychir ; Mahnert, 1976: 472, type information. Amphilius uranoscopus (non Pfe ffer): Skelton, 1984: 67 (in part), type information, Okavango basin and upper Zambezi River basin; Van der Waal & Skelton, 1984: 316, Liste d as present in Zambezi River at Caprivi, Namibia ; Skelton, 1985: 9 (in part), Okavango basin; Skelton, 1994: 129 (i n part); Okavango, upper Zambezi, and middle basins. Material examined Okavango River basin: MNHN 1936 0101, Angola, Cubango [Okavango] River, near Vila da Ponte [Vila arthur de Pavia], Angola, ca. 1427S, 1618E (1: 74.1, syntype); MNHN 19360103, sam e data as MNHN 1936 0101 (3: 60.6 90.9, syntypes); MRAC 138769, same data as MNHN 1936 0101, (1: 75.5, syntype); SAIAB 22055, Namibia Okavango River at Popa Rapids (or Falls), near Bagani ca. 1806S, 2136E (8: 38.782.7 ); SAIAB 22090, same locality as SAIAB 22055, (3: 39.0 76.5); SAIAB 27387, Namibia Okavango River at Kahenge ca. 1740S, 1840E (3: 31. 63.3); SAIAB 36842 same locality as SAIAB 22055, (1: 35.0 ); SAIAB 39055 same locality as SAIAB 22055, (1: 68.5 ); SAIAB 40755 Namibia Okavango R iver at Mabushe, ca. 1755S, 2025E (1: 15.7 ); SAIAB 42970 Namibia Okavango River at Popa ca. 1807S, 2136E (2: 19.8 22.9 ); SAIAB 43421 same locality as SAIAB 22055, (6: 38.9 82.5mm SL); SAIAB 55056 same locality as SAIAB 22055, (7: 43.4 107.3); SAIAB 56944, Namibia Okavango River at Mashare ca. 1753S, 2013E (2: 63.567.0 ); SAIAB 70395 same locality as SAIAB 22055, (1: 73.5 ); SAIAB 96484 Namibia Okavango River at Popa Rapids (or Falls), west side near camp site, 180703S, 213457E (9: 45.9 102.1 ); SAIAB 96500 Namibia Okavango River at Popa Rapids (or

PAGE 102

102 Falls), side channel in campsite, 180719S, 213459E (8: 42.9 71.1 ). Lower Zambezi River basin: SAIAB 40163 Zimbabwe Sohwe Falls, near Muzarabani ca. 1 629'S, 3101 'E (4: 38.1 88. 4 ); SAIAB 83586 Zambia, Kundalila River, a bove falls ca. 1 309'12S, 30 42' 36 E (19: 26.284.7). Middle Zambezi River basin: ROM 31617, Zambia, Sichikwenkwe Stream at Kalomo, ca. 1703S, 2630E (2: 32.5 92.2 ); ROM 31651, Zambia, Kalomo River (1: 86.6) Upper Zambezi River basin: ROM 31814 Zambia, Victoria Falls Edge ca. 175525S, 255521E (22: 52.3104.8); ROM 31873, same locality as ROM 31814, (19: 30.34138.7); SAIAB 24214 Namibia Kwando (Cuando) River at Safari Camp, ca. 1812S, 2323E (1: 64.4); SAIAB 27723 Zambia Near Kalengwa, Zambia, Musondweji River system ca. 1328S, 2500E (1: 67.5 ); SAIAB 27725 same locality as SAIAB 27723 (3: 49.4 54.6 ); SAIAB 27729 Zambia, Musondweji River ca. 1330S, 2510E (2: 41.9 65.0 ); SAIAB 40099 same locality as SAIAB 27729 (6: 69.4.7); SAIAB 41046 Zambia, Lumwana River, off Solwezi Mwinilunga Road, ca. 1216S, 2540E (1: 66.0); SAIAB 41207 Zambia Lumwana River at East Lumwana confluence between MwinilungaSolwezi ca. 1215S, 2540E (1: 70.2); SAIAB 42892 Zambia, Sakeji River at Sakeji School, Ikelengi ca. 1114S, 2416E (9: 29.857.7); SAIAB 42900 Zambia, 10kms from Zambezi River source, ca. 1114S, 2421E (2: 37.1 49.0 ); SAIAB 42901 same locality as SAIAB 42892 (11: 35.971.6 ); SAIAB 42907 Zambia Kabompo River, Solwezi Mwinilinge road bridge, 115317S, 251502E (6: 33.8 69.0 ); SAIAB 71035 Zambia, Kabompo River near Kabompo, 133649S, 241220E (6: 31.8 120.2); SAIAB 71080 Zambia, Upper Zambezi River near town of Sioma, near junction of Route M10 and Route RD323, 163929S, 233419E (22: 23.2136.0); SAIAB 71116 Zimbabwe Upper Zambezi

PAGE 103

103 River near Sankaia, just west of Livingstone, about 15 km upstream of Victoria Falls 171521S, 254620E (19: 29.689.8); SAIAB 71131 Zimbabwe Upper Zambezi River near Sankaia, just west of Livingstone, about 14 km upstream of Victoria Falls 175151S, 254650E (3: 40.2 54.6 ); SAIAB 72566 Zambia, Upper Zambezi River just north of junction of route RD324 with route M10, near town of Ilwendo [=Iluando] 171200S, 240400E (3: 42.2 76.8 ); SAIAB 72661 Zambia, Sioma Falls on the Upper Zambezi River, near town of Sioma, near junction of Route M10 and Route RD323 163927S, 233420E (32: 39.0 100.0 ); SAIAB 73317 Zambia Ma heba River on Solwezi Mwinilinge road bridge (route T5) 122036S, 254428E (4: 44.3 103.2); SAIAB 73342 Zambia, East Lumwana River 121637S, 253957E (2: 59.3 ); SAIAB 73365, Zambia, Upper Zambezi River, north of Kalene Hill 110732S, 241129E (18: 44.8107.8); SAIAB 73374, Zambia Upper Zambezi River b elow Upper Zambezi bridge, 110809S, 240745E (13: 43.087.4); SAIAB 73386 Zambia, Luakela River at Rte T5 crossing near Ikatu north of Mwinilunga, 113137S, 242433E (45: 38.2 93.0 ); SAIAB 73402 Zambia, East Lumwana near at Rte T5 crossing 121639S, 253959E (21: 37.7 95.0 ); SAIAB 73424 Zambia Mwombezhi River 121435S, 253426E (37: 39.7114.3); SAIAB 73476, same locality as SAIAB 42907 (71: 26.4 88.3); SAIAB 73501, Zamb ia Kaseka River at MwinilungaIkelenge road (route T5) crossing about 40 km north of Mwinilunga, 112548S, 242009E (2: 71.6 71.8); SAIAB 73511 Zambia Lwinga River, near Ikelenge, 111401S, 241500E (66: 37.296.2 ); SAIAB 73521 Zambia, Upper Za mbezi River, above upper Zambezi bridge, 110809S, 240745E (19: 39.0 105.0); SAIAB 73554 Zambia, Madamyana River near junction with West Lunga River, just south of Mwinilunga, 114517S, 242613E (16: 42.9117.8); SAIAB

PAGE 104

104 73574, Zambia, Kanyanda Riv er, on road from Mwinilunga to Kabompo (route D286), about 33 km south of Mwinilunga, 115943S, 242104E (10: 27.9 120.1); SAIAB 73585, Zambia, Ramukonye River, on road from Mwinilunga to Kabompo (route D286), about 35 km south of Mwinilunga, 120026 S, 242033E (6: 67.596.8); SAIAB 73645 Zambia, Makondu River on road from Zambezi to Chavuma (route D293), about 21 km north of Zambezi 132222S, 230432E (3: 54.183.1 ); SAIAB 73674 Zambia Zambezi River near Katombora, at Zambezi Zimbabwe border 175059S, 252224E (24: 36.0 107.0); SAIAB 80446 Zambia, East Lumwana River downstream of Malundwe, 121419S, 254741E (10: 40.7 68.0); SAIAB 80468 Zambia, Malundwe tributary, c auseway across river, East Lumwana River tributary 121117S, 254835E (12: 37.2 78.8 ); SAIAB 80487 Zambia Malundwe tributary, Causeway across Malundwe, East Lumwana River tributary 121039S, 254937E (8: 41.8 72.0 ); SAIAB 80511 Zambia, Kabompo River, Solwezi Mwinilinge road bridge, 115317S, 251502E (6: 35.2 68.3 ); SAIAB 80539 Zambia East Lumwana River at Rte T5 crossing between Chisasa and Chibombe 121625S, 253959E (11: 31.092.0); SAIAB 80555 Zambia, Malundwe tributary, above East Lumwana confluence, 121359S, 254802E (9: 44.1 96.2 ); SAI AB 80564 Zambia, Chimiwungo tributary, bridge on mine site, East Lumwana River system 121506S, 253959E (6: 68.5105.5); SAIAB 80577 Zambia, Maheba River on Solwezi Mwinilinge road bridge (route T5) 122036S, 255428E (3: 43.080.3); SAIAB 958 70, Zambia 121540S, 251735E (3: 42.9 136.4 ). Diagnosis Diagnostic characters are summarized in Tables 4 2 and 4 3 Anoplopterus cubangoensis is diagnosed from A. n. sp. Ruvu A. n. sp. Mangula and A. krefftii by

PAGE 105

105 having a forked caudal fin (vs. emar ginated caudal fin) and from A. uranoscopus, A. leroyi, A. hargeri, A. brevidorsailis, A. n. sp. Mangula, A. n. sp. Ruvu A. grandis, A. krefftii, A. chalei and A. athiensis by having a mottled coloration in adults (vs. body not mottled in adults). It is diagnosed from A. uranoscopus A. leroyi A. hargeri. A. n. sp. Mangula, and A. athiensis by having a narrower interorbital width (19.823.9 % HL vs. 23.531.7 % HL). Anoplopterus cubangoensis is diagnosed from A. n. sp. Ruvu and A. krefftii by having a longer body with 37 to 38 total vertebrae (vs. usually 36 or less). It is diagnosed from A. hargeri and A. n. sp. Ruvu by its first dorsal pterygiophore intercept count (1 vs. usually 2) and from A. athiensis and A. grandis by having distinct pale patches at or igin and insertion of dorsal fin (vs. pale patches absent). Anoplopterus cubangoensis is further diagnosed from A. n. sp. Mangula and A. chalei by having a longer predorsal length (36.038.5 % SL vs. 31.6 34.9% SL) and from A. uranoscopus, A. leroyi, A. har geri, and A. n. sp. Ruvu by having a shallower body (body depth at anus 10.212.2 % SL vs. 12.516.3 % SL). It is further diagnosed from A. uranoscopus, A. leroyi, A. brevidorsalis and A. hargeri by having a longer caudal peduncle length (17.721.7% SL vs. 1 4.4 17.9 % SL) and from A. uranoscopus, A. brevidorsalis and A. hargeri by having a shallower caudal peduncle (8.110.9 vs. 10.7 14.3 % SL). Anoplopterus cubangoensis is further diagnosed from A. brevidorsalis and A. n. sp. Ruvu by usually having 9 or 10 gil l rakers on the first arch (vs. usually 6 to 8, rarely 9). It is further diagnosed from A. n. sp. Ruvu by having longer maxillary barbels (51.478.4 % HL vs. 37.349.6 % HL). Description Morphometric d ata in Table 4 7 Body elongate, ventral profile flattened ventrally to anal fin base, then tapered dorsally to end of caudal peduncle. Dorsal profile rising

PAGE 106

106 gently from tip of snout to origin of dorsal fin, then nearly horizontal to end of caudal peduncle. Greatest body depth at dorsal fin origin. Caudal pedun cle laterally compressed. Anus and urogenital openings located just posterior to base of pelvic fin, closer to insertion of pelvic fin than to or igin of anal fin. Skin smooth. Lateral line complete, extending from dorsal edge of opercular cavity to base of caudal fin. Total vertebrae 37 (4) or 38 (5). Abdominal vertebrae 20 (2), 21 (3), or 22 (3). Caudal verte brae 16 (2), 17 (4), or 18 (2). First dorsal pterygiophore intercept count 1 (9). Head and anterior par t of body depressed and broad. Head wedgeshaped in lateral view. Snout broad, moderately pointed when viewed from above. Head becoming wider from tip of snout to pectoral fin base. Branchiostegal membranes moderately joined at isthmus forming a V shaped or deeply concave connection. Mouth broad, gent ly curved, subterminal. Lips moderately fleshy, papillate. Rictal lobe large and papillate. Anterior portion of premaxillary tooth band exposed with mouth closed. Premaxillary tooth patches joined, forming U shaped band, with short conical teeth. Dentary t eeth short and conical, tooth patches forming U shaped band, separated medially. Three pairs of simple, tapered circumoral barbels. Maxillary barbel large, fleshy and flattened with pointed tip; barbel extending posterolaterally from corner of mouth to jus t short of pectoral fin base. Outer mandibular barbel thin with pointed tip, origin at posterior corner of lower jaw, extending to edge of branchiostegal membrane. Inner mandibular barbel originates anterolaterally of inner mandibular barbel, extending to edge of branchiostegal membrane. Branchiostegal membrane with 6 (2), 7 (37), 8 (28), or 9 (1) rays. Gill rakers on first epibranchial 2 (14), 3 (83) or 4 (2); rakers on first

PAGE 107

107 ceratobranchial 5 (7), 6 (49), 7 (37) or 8 (6); total gill rakers on first arch 7 (1), 8 (8), 9 (57), 10 (26), or 11 (7). Eyes small, positioned dorsolaterally approximately midway between tip of snout and posterior margin of operculum. Horizontal diameter of eye slightly wider than vertical diameter. Eye without free orbit; covered wi th skin confluent with dorsal surface of head. Anterior and posterior nares with prominent tubular rims; nares separate but relatively close to each other. Posterior nare located about midway between eye and tip of snout. Dorsal fin origin at point over t ip of pectoral fin. Dorsal fin with i,6 (108) rays, and fin margin straight. Pectoral fin with i,8 (1), i,9 (52), or i,10 (51) rays ; unbranched ray greatly thickened. Pectoral fin with four or five innermost rays progressively shorter making posterior fin margin rounded. Pelvic fin inserted posteriorly to dorsal fin base. Pelvic fin with i,5 (108) rays with first ray unbranched and greatly thickened. Pelvic fin with straight posterior margin. Adiposefin base longer than anal fin base; origin anterior to or igin of anal fin base; fin extending past anal fin insertion. Margin strongly convex with sharply rounded edge, not deeply incised posteriorly. Caudal fin emarginate, with i,7,8,i (99) principal rays. Anal fin with short base; origin posterior to origin of adiposefin base; fin with ii,5 (32), ii, 6 (61), iii,5 (11), or iii,6 (1) rays. Anal fin margin almost straight Coloration Dorsal and lateral surfaces of head and body m ottled black with light brown/cream patches Body head, and fins sometimes heavily spotted. Ventral region light brown. Light dorsal saddles at origin and insertion of dorsal fin and adipose fin. Dorsal, adipose, caudal, and anal fins brown. Pectoral and pelvic fins positioned

PAGE 108

108 horizontally with upper surfaces brown and lower surfaces light yellow. Maxillary and mandibular barbels brown. Caudal fin with dark crescent shaped band at base and broad diffuse band on rays. Juvenile color ation similar to that of adult. Distribution Known from the Okavango River and Zambezi River basins in Zambia, Zimbabwe, Namibia, and Angola (Figure 46 ). Anoplopterus n. sp. Mangula new species (Figure 4 9; Table 4 8 ) Holotype. UF 170702 Tanzania, Sonjo River at bridge in Man'gula on road from Mikumi to Ifakara, altitude 302 m Kilombero River drainage, Rufiji River basin, 7 48' 29.6"S, 36 53' 47.6 "E (1: 95.6). Paratypes CU 93746 same data as holotype (9 : 32.597.2 ); UF 1 8943, same data as holotype, (8: 32.5 124.2). Non types Rufiji River basin, Kilombero River drainage : AMNH 215941, Tanzania, Udz ungwa mountains national park, Njokamoni R iver system, ca. 7 51 S, 3653 'E (3: 15.141.7 ); same locality as AMNH 215941, (11: 40.064.6); Rufiji River basin, Great Ruaha River drainage : CU 93743 Tanzania, Lukosi River along road from Iringa to Morogoro (Rte A7), altitude 721 m 7 40' 08.5"S, 36 13' 50.3"E (3: 46.4 54.5); SAIAB 594 1 3 Tanzania, Iyovi River on road from Mikumi to Mbuyuni (Rte A7 crossing) above Kidatu Dam ca. 7 34 S, 36 47'E (2: 25.562.8); SAIAB 59431 Tanzania, Lukosi River along road from Iringa to Morogoro (Rte A7) just ENE of Matassi ca. 7 34 S, 36 47 'E (2: 69.970.4); UF 184936, same data as CU 93743, (2: 47.6 59.1 ). Diagnosis Diagnostic characters are summarized in Tables 4 2 and 4 3 Anoplopterus n. sp. Mangula is diagnosed from A. uranoscopus A. leroyi A. brevidorsalis A. hargeri, A.

PAGE 109

109 cubangoensis, A. grandis, A. chalei and A. athiensis by having a emarginate caudal fin (vs. forked caudal fin) and from A. n. sp. Ruvu A. athiensis, and A. cubangoensis by having a body color that is not finely spotted or mottled (vs. body finely spotted in A. n. sp. Ruvu and A. athiensis and mottled in A. cubangoensis ). It is diagnosed from A. cubangoensis and A. chalei by having a wider interorbital width (24.3 27.2% HL vs. 17.924.2 % HL). Anoplopterus n. sp. Ma ngula is diagnosed from Amphilius uranoscopus A. chalei, and A. athiensis by having a shorter body with 36 to 37 vertebrae (vs. usually 38 or more vertebrae). It is diagnosed from A. hargeri and A. n. sp. Ruvu by its first dorsal pterygiophore intercept c ount (1 vs. usually 2) and from A. athiensis and A. grandis by having distinct pale patches at origin and insertion of dorsal fin (vs. pale patches absent). Anoplopterus n. sp. Mangula is diagnosed from A. uranoscopus, A. leroyi, A. brevidorsalis, A. hargeri, A. cubangoensis, A. n. sp. Ruvu and A. athiensis by having a shorter predorsal length (31.634.9 % SL vs. 35.039.9% SL), from A. chalei by having a deeper body (body depth at anus 11.613.1 % SL vs. 8.9 10.8% SL), and A. leroyi and A. hargeri by having a shallower body (body depth at anus 11.613.1 % SL vs. 13.016.0 % SL). It is further diagnosed from A. uranoscopus, A. leroyi, A. brevidorsailis, A. hargeri, A. n. sp. Ruvu and A. athiensis by having a longer caudal peduncle length (18.220.3% SL vs. 14.4 18.3 % SL) and from and A. chalei by having a deeper caudal peduncle (9.912.4 vs. 7.1 9.7 % SL). Anoplopterus n. sp. Ruvu is further diagnosed from A. n. sp. Ruvu by usually having 8 or 9 gill rakers on the first arch (vs. usually 6 to 7, rarely 8) and by usually having 6 branched anal fin rays (vs. usually 5, rarely 6). It is further diagnosed from A. n. sp. Ruvu by having longer maxillary barbels (71.4 92.6% HL vs. 37.349.6% HL).

PAGE 110

110 Description Morphometric d ata in Table 4 8 Body elongate, ventral profile flattened ventrally to anal fin base, then tapered dorsally to end of caudal peduncle. Dorsal profile rising gently from tip of snout to origin of dorsal fin, then nearly horizontal to end of caudal peduncle. Greatest body depth at dorsal fin origin. Caudal ped uncle laterally compressed. Anus and urogenital openings located just posterior to base of pelvic fin, closer to insertion of pelvic fin than to origin of anal fin. Skin smooth. Lateral line complete, extending from dorsal edge of opercular cavity to base of caudal fin. Total vertebrae 36 (4) or 37* (7). Abdominal vertebrae 19* (2), 20 (7), or 21 (1). Caudal vertebrae 16 (4), 17 (5), or 18* (1). First dorsal pterygiophore intercept count 1* (11). Head and anterior part of body depressed and broad. Head wedgeshaped in lateral view. Snout broad, moderately pointed when viewed from above. Head becoming wider from tip of snout to pectoral fin base. Branchiostegal membranes moderately joined at isthmus forming a V shaped or deeply concave connection. Mouth broad, gently curved, subterminal. Lips moderately fleshy, papillate. Rictal lobe large and papillate. Anterior portion of premaxillary tooth band exposed with mouth closed. Premaxillary tooth patches joined, forming U shaped band, with short conical teet h. Dentary teeth short and conical, tooth patches forming U shaped band, separated medially. Three pairs of simple, tapered circumoral barbels. Maxillary barbel large, fleshy and flattened with pointed tip; barbel extending posterolaterally from corner of mouth to just short of pectoral fin base. Outer mandibular barbel thin with pointed tip, origin at posterior corner of lower jaw, extending to edge of branchiostegal membrane. Inner mandibular barbel originates anterolaterally of inner mandibular barbel, extending to

PAGE 111

111 edge of branchiostegal membrane. Branchiostegal membrane with 8* (17) or 9 (6) rays. Gill rakers on first epibranchial 2 (1) or 3 (22); rakers on first ceratobranchial 4 (3), 5 (12), or 6* (8); total gill rakers on first arch 7 (4), 8 (11), or 9* (8). Eyes small, positioned dorsolaterally approximately midway between tip of snout and posterior margin of operculum. Horizontal diameter of eye slightly wider than vertical diameter. Eye without free orbit; covered with skin confluent with dorsal sur face of head. Anterior and posterior nares with prominent tubular rims; nares separate but relatively close to each other. Posterior nare located about midway between eye and tip of snout. Dorsal fin origin at point over tip of pectoral fin. Dorsal fin with i,6* (22) or i,7 (1) rays, and fin margin straight. Pectoral fin with i,9 (6) or i,10* (17) rays ; unbranched ray greatly thickened. Pectoral fin with four or five innermost rays progressively shorter making posterior fin margin rounded. Pelvic fin inser ted posteriorly to dorsal fin base. Pelvic fin with i,5 (23) rays with first ray unbranched and greatly thickened. Pelvic fin with straight posterior margin. Adiposefin base longer than anal fin base; origin anterior to origin of anal fin base; fin extending past anal fin insertion. Margin strongly convex with sharply rounded edge, not deeply incised posteriorly. Caudal fin emarginate, with i,7,8,i (22) principal rays. Anal fin with short base; origin posterior to origin of adiposefin base; fin with ii,5 (2), ii,6* (20), or iii,6 (1) rays. Anal fin margin almost straight. Coloration Dorsal and lateral surfaces of head and body brown. Ventral region light brown. Light dorsal saddles at origin and insertion of dorsal fin and adipose fin. Dorsal, adipose, caudal, and anal fins brown. Pectoral and pelvic fins positioned horizontally

PAGE 112

112 with upper surfaces brown and lower surfaces light yellow. Maxillary and mandibular barbels brown. Caudal fin with dark crescent shaped band at base and broad diffuse band on rays. Juvenile color ation similar to that of adult. Distribution Known only from the Rufiji River basin Tanzania (Figure 4 3 ). Anoplopterus n. sp. Ruvu new species (Figure 4 10; Table 49 ) Holotype. UF 170727 Tanzania, Mgeta River at Mgeta Secondary School, al titude 1001 m, 7 02' 16.5"S, 37 34' 09.8"E (1: 75.6). Paratypes CU 9 7551 same data as holotype (1 5 : 47.0 80.5 ); UF 184938, same data as holotype (13: 44.180.6). Diagnosis Diagnostic characters are summarized in Tables 4 2 and 4 3 Anoplopterus n. sp. Ru vu is diagnosed from A. uranoscopus, A. leroyi, A. brevidorsalis, A. cubangoensis, A. n. sp. Mangula, A. krefftii, A. chalei, and A. athiensis by its first dorsal pterygiophore intercept count (usually 2 vs. usually 1) and from A. uranoscopus, A. leroyi, A brevidorsalis, A. cubangoensis, A. grandis, A. chalei and A. athiensis by having a shorter body with usually 35 to 36 total vertebrae (vs. usually 37 or more). Anoplopterus n. sp. Ruvu is diagnosed from A. uranoscopus, A. leroyi, A. brevidorsalis, A. har geri, A. cubangoensis, A. grandis, A. chalei, and A. athiensis by having a emarginate caudal fin (vs. forked caudal fin) and from all other species of the Anoplopterus by having a h eavily spotted coloration with spots usually coalesced to form vermiculations (vs. body not finely spotted or if spotted, spots never coalesced to form vermiculations ). It is diagnosed from A. uranoscopus, A. leroyi, and A. hargeri by having a narrower interorbital width (20.724.9% HL vs. 25.231.7% HL) and A. grandis by having distinct pale patches at origin and insertion of dorsal fin (vs. pale patches absent). Anoplopterus

PAGE 113

113 n. sp. Ruvu is further diagnosed from A. n. sp. Mangula and A. chalei by having a longer predorsal length (36.139.9% SL vs. 31.6 34.9 % SL) and from A. cuba ngoensis and A. chalei by having a deeper body (body depth at anus 12.815.7% SL vs. 8.9 12.2% SL). It is further diagnosed from A. n. sp. Mangula and A. chalei by having a shorter caudal peduncle length (16.018.2% SL vs. 18.2 21.9 % SL) and from A. chalei by having a deeper caudal peduncle (10.312.3 vs. 7.99.7 % SL). Anoplopterus n. sp. Ruvu is further diagnosed from A. uranoscopus, A. leroyi, A. hargeri, A. cubangoensis, A. n. sp. Mangula, A. krefftii, A. grandis, and A. athiensis by usually having 6 or 7 gill rakers on the first arch (vs. usually 8 or more) and from A. uranoscopus, A. leroyi, A. brevidorsailis, A. hargeri, A. n. sp. Mangula and A. grandis by usually having 5 branched anal fin rays (vs. usually 6 or 7). It is further diagnosed from all o ther species of the A. uranoscopus, A. leroyi, A. hargeri, A. brevidorsailis, A. cubangoensis, A. n. sp. Mangula, A. grandis, A. chalei, and A. athiensis by having shorter maxillary barbels (37.349.6 % HL vs. 49.799.1 % HL). Description Morphometric d ata in Table 4 9 Body elongate, ventral profile flattened ventrally to anal fin base, then tapered dorsally to end of caudal peduncle. Dorsal profile rising gently from tip of snout to origin of dorsal fin, then nearly horizontal to end of caudal peduncle. Gr eatest body depth at dorsal fin origin. Caudal peduncle laterally compressed. Anus and urogenital openings located just posterior to base of pelvic fin, closer to insertion of pelvic fin than to origin of anal fin. Skin smooth. Lateral line complete, extending from dorsal edge of opercular cavity to base of caudal fin. Total vertebrae 35* (9), 36 (4), or 37 (1). Abdominal vertebrae 20* (9), or 21 (5). Caudal

PAGE 114

114 vertebrae 14 (1), 15* (11), or 16 (2). First dorsal pterygiophore intercept count 1* (2) or 2 (11). Head and anterior part of body depressed and broad. Head wedgeshaped in lateral view. Snout broad, moderately pointed when viewed from above. Head becoming wider from tip of snout to pectoral fin base. Branchiostegal membranes moderately joined at isthmus forming a V shaped or deeply concave connection. Mouth broad, gently curved, subterminal. Lips moderately fleshy, papillate. Rictal lobe large and papillate. Anterior portion of premaxillary tooth band exposed with mouth closed. Premaxillary tooth patches joined, forming U shaped band, with short conical teeth. Dentary teeth short and conical, tooth patches forming U shaped band, separated medially. Three pairs of simple, tapered circumoral barbels. Maxillary barbel large, fleshy and flattened with pointed tip; barbel extending posterolaterally from corner of mouth to just short of pectoral fin base. Outer mandibular barbel thin with pointed tip, origin at posterior corner of lower jaw, extending to edge of branchiostegal membrane. Inner mandibular barbel originates anterolaterally of inner mandibular barbel, extending to edge of branchiostegal membrane. Branchiostegal membrane with 8 (28) rays. Gill rakers on first epibranchial 2* (24) or 3 (5); rakers on first ceratobranchial 4* (14) or 5 (15); total gill rakers on first arch 6* (11), 7 (16), or 8 (2). Eyes small, positioned dorsolaterally approximately midway between tip of snout and posterior margin of operculum. Horizontal diameter of eye slightly wider than vertical diameter. Eye without free orbit; covered with skin confluent with dorsal surface of head. Anterior and posterior nares with prominent tubular rims; nares separate but

PAGE 115

115 relatively close to each other. Posterior nare located about midway between eye and tip of snout. Dorsal fin origin at point over tip of pectoral fin. Dorsal fin with i,6 (29) rays, and fin margin straight. Pectoral fin with i,9* (28) or i,10 (1) rays ; unbranched ray greatly thickened. Pectoral fin with four or five innermost rays progressively shorter making posterior fin mar gin rounded. Pelvic fin inserted posteriorly to dorsal f in base. Pelvic fin with i,5 (29) rays with first ray unbranched and greatly thickened. Pelvic fin with straight posterior margin. Adiposefin base longer than anal fin base; origin anterior to origin of anal fin base; fin extending past anal fin insertion. Margin strongly convex with sharply rounded edge, not deeply incised posteriorly. Caudal fin emarginate, with i,7,8,i* (29) principal rays. Anal fin with short base; origin posterior to origin of adipose fin base; fin with ii,5* (27) or ii,6 (2) rays. Anal fin margin almost straight. Coloration Dorsal and lateral surfaces of head and body heavily spotted, spots usually coalesced to form vermiculations Ventral region light brown. Light dorsal saddles at origin and insertion of dorsal fin and adipose fin. Dorsal, adipose, caudal, and anal fins brown. Pectoral and pelvic fins positioned horizontally with upper surfaces brown and lower surfaces light yellow and with spots Maxillary and mandibular barbel s brown. Caudal fin with dark crescent shaped band at base and broad diffuse band on rays. Juvenile coloration similar to that of adult but spots usually larger. Distribution Known from the Ruvu River basin, Tanzania (Figure 4 3 ).

PAGE 116

116 Discussion The present st udy examined over 1000 museum specimens of Anoplopterus from the Wami River basin south to the Buzi River basin and west to the upper Zambezi River and Okavango River basins in eastern and southern Africa. This study area included the type localities of A. uranoscopus (Wami basin), A. leroyi (Ruvu basin), A. hargeri (Zambezi basin), A. brevidorsalis (Buzi basin), and A. cubangoensis ( Okavango basin) all of which were determined to be valid species based on differences in coloration, body shape, and meristic counts. Additionally, two new species, A. n. sp. Ruvu from the Ruvu River basin and A. n. sp. Mangula from the Rufiji basin, were described. This brings the total number of species of Anoplopterus to 17, eight of which occur in eastern Tanzania and southern Kenya. Of the 17 species, four species ( A. cryptobullatus, A. lampei, A. kivuensis, and A. zairensis ) do not occur in southeastern Africa. Anoplopterus cryptobullatus which is found in the upper Congo basin, is diagnosed from all other species of Anop lopterus by having extremely large bilateral bony swimbladder capsules (see Skelton, 1986 figure 16a) (vs. bilateral bony swimbladder capsules normally developed) and from all species recognized in this study by the combination of an emarginated caudal fin and mottled body coloration. Anoplopterus lampei from Ethiopia, is diagnosed from all species described in this study by having an adipose fin confluent with the caudal fin. Anoplopterus kivuensis, known from the Nile and Congo basins is diagnosed by its posteriorly placed dorsal fin, and first dorsal pterygiophore intercept count usually 4 or more vs. usually 2 or less). Anoplopterus zairensis found in the lower Congo basin, i s diagnosed by its shorter body and 32 35 vs. 3 6 42 total vertebrae in all spe cies except A. n. sp. Ruvu which has 3537 total vertebrae. Anoplopterus zairensis is diagnosed

PAGE 117

117 from A. n. sp. Ruvu by having no dark spots on the body (vs. h eavily spotted with spots usually coalesced to form vermiculations ). Two species found in southeas tern Africa were not compared in the diagnoses because they are morphologically distinct and easily distinguished from other species recognized in this study. Anoplopterus laticaudatus is found in the Buzi basin and is diagnosed from all other species by having an adipose fin confluent with the caudal fin. Anoplopterus natalensis occurs in the Buzi and lower Zambezi basins and is diagnosed by its posteriorly placed dorsal fin, first dorsal pterygiophore intercept count usually 4 or m ore vs. usually 2 or les s. This increase in the number of species of Anoplopterus from five recognized by Skelton (1984) to the current number is due almost entirely to the increased availability of specimens from eastern and southern Africa due to recent collecting. Of the seven new species described since Skeltons revision, six were described entirely from specimens collected after his study. Additionally, at the time of Skeltons study, the only specimens of A. uranoscopus available were the two type specimens and the only sp ecimen of A. leroyi available was the holotype. Anoplopterus uranoscopus and A. leroyi are the two oldest names available for eastern African Amphilius and, until the identities of these two species could be determined, it was not possible to determine the validity of the other nominal species described from eastern Africa. In the fall of 2007, 57 specimens of A. uranoscopus were collected from three localities in the Wami River basin including two localities very close the type locality. Additionally, 33 specimens of A. leroyi were collected from the Ruvu basin. These collections allowed for the first time the identities of A. uranoscopus and A. leroyi to be

PAGE 118

118 determined and for the status of the other nominal species placed in the synonymy of A. uranoscopus to be revisited. The status of three of these nominal species, A. grandis, A. oxyrhinus and A. krefftii were addressed by Thomson & Page (2010), but they did not address the status of A. leroyi, A. hargeri, A. brevidorsalis and A. cubangoensis which have been determined to be valid species in this study. Anoplopterus uranoscopus occurs in the Wami River basin in eastern Tanzania where it is probably the only species of the A. uranoscopus group in the basin. It is also found to be present in the Rufiji R iver basin based on a single collection from 12 Bridges River These specimens have subtle differences in coloration from specimens from the Wami River basin but dont differ in any meristic counts and are supported with molecular data to be most closely r elated to specimens from the Wami basin. Anoplopterus leroyi occurs only in the Ruvu basin where A. n. sp. Ruvu also occurs. Anoplopterus leroyi was found at multiple sites in the Ruvu basin, but A n. sp. Ruvu was found only at the highest elevation site collected in the basin. At the collection site of A. n. sp. Ruvu a large series of specimens that were initially identified as A. leroyi were also collected. W hen examined more closely these specimens were found to differ from A. leroyi and require furth er study Anoplopterus hargeri occurs in the Shire River system of the Lower Zambezi River basin and the Rovuma River basin. It is the only species of Anoplopterus known from the Rovuma River basin, but is found with A. natalenis in the Shire River system. Amphilius natalenis is similar to A. hargeri externally but can be separated from this species by its more posteriorly placed dorsal fin (first dorsal pterygiophore intercept count usually 4 or more vs. usually 2 or less in A. hargeri ).

PAGE 119

119 Anoplopterus brev idorsalis occurs in the Buzi basin in Mozambique where A. natalenis and A. laticaudatus also occur, but can be separated from these species by having an adipose fin that is not confluent with the caudal fin (vs. adipose fin confluent with the caudal fin in A. laticaudatus and Buzi River populations of A. natalenis ). Anoplopterus cubangoensis is found throughout the upper Zambezi River basin and the Okavango River basin. It is also the species of Anoplopterus that occurs at the edge of Victoria Falls identi fied as A. platychir by Balon (1974). Anoplopterus cubangoensis has also been collected from the middle Zambezi River basin in tributaries just below Victoria Falls. It is the only species of Anoplopterus that occurs in the middle and upper Zambezi River basins and the Okavango River basin. Frequency tables were constructed for counts of four meristic characters (Tables 4 10 to 4 13). Total vertebrae count, f irst dorsal pterygiophore intercept count total gill raker count, and branched anal fin ray count w ere found to be useful for distinguishing species. Total vertebrae count varied from 35 to 40 in the specimens examined (Table 4 10). Five species are characterized by a low vertebral count, usually having fewer than 38 total vertebrae: A. leroyi, A. hargeri, A. n. sp. Mangula A. n. sp. Ruvu and A. krefftii Three other species, A. uranoscopus A. chalei and A. athiensis have high vertebral counts, usually having 38 or more total vertebrae. F irst dorsal pterygiophore intercept count varied from zero to three among species, but the count is very stable within species (Table 4 11). Anoplopterus hargeri and A. n. sp. Ruvu have a f irst dorsal pterygiophore intercept count that is almost always two. All other species except A. grandis have a dorsal pterygiophore intercept count that is almost always one. In A. grandis the count is one or two.

PAGE 120

120 Total gill raker count varied from six to 11 (Table 4 12). Counts were quite variable within species, but A. brevidorsalis and A. n. sp. Ruvu mostly had counts of eight or less while A. uranoscopus A. cubangoensis A. chalei and A. athiensis mostly had counts of nine or more. Branched anal fin ray count varied from five to seven, with most species having a mode of six (Table 4 13). Anoplopterus n. sp. Ruvu is the only species with a mode of five branched anal fin rays. Although this study greatly increases the number of species recognized in Anoplopterus taxonomic work on the genus is far from complete. Within the study area, additional undescribed diversity likely exis ts in the tributaries of Lake Malawi. I was able to examine a few specimens from northern Malawi tributaries that appeared to be an undescribed species based on color pattern; however, additional material needs to be examined to confirm the number of speci es in the tributaries. In the Rufiji River basin, Anoplopterus n. sp. Mangula was collected at three sites with a total of four heavily mottled specimens. The two mottled specimens that were collected together appear to be different species based on caudal fin shape. One of the specimens has a deeply forked caudal fin, while the other has an emarginated caudal fin. The other two mottled specimens have forked caudal fins, but the color patterns in all four specimens look very different from one another. Outside the study area, study is needed of material from the Incomatti Maputo, Umbuluzi (Mbuluzi) and Limpopo River b asin s in South Africa and Swaziland This material has been recognized as A. uranoscopus (Skelton, 2001), but is actually more similar to A. brevidorsalis and probably represents one or more undescribed species. Specimens from the Malagarasi River drainage, the Lake Rukwa basin, and southern

PAGE 121

1 21 tributaries of Lake Tanganyik a that have been identified as A. uranoscopus may also represent undescribed species. Several different species similar to A. uranoscopus appear to be present in the Malagarasi River drainage and multiple species also appear to be present in the southern tributaries of Lake Tanganyik a. Comparative Material Examined Anoplopterus athiensis : Galana R iver basin: BMNH 1905.12.11.2, Kenya, Nairobi River, Kikaya, ca. 110'33"S, 3656'24"E (1: 160.3, paratype); BMNH 1909.11.15.2123, Nairobi River, elev. 2500 ft., ca. 110'33"S, 3656'24"E (2: 89.1119.5, paratypes ); BMNH 1910.10.31.31, Nairobi, Kenya, ca. 116'28"S, 3648'47"E (1: 150.0, holotype) ; BMNH 1928.11.10.1112, Nairobi River, elev. 6500 ft., ca. 110'33"S, 3656'24"E (3: 59.6140.5 paratypes ); BMNH 1928.11.10.1318, Kenya, Riara River, elev. 5700 ft., ca. 18'49"S, 3657'47"E (2: 37.687.2 paratypes ); BMNH 1937.12.11.1618, Kenya, Mbakasi River, ca. 118'10"S, 3655'10"E (3: 58.1110.3, paratypes ); BMNH 1937.6.4.43, Kenya, Athi River, Fourteen Falls, ca. 17'40"S, 3721'25"E (1: 44.0, paratype); BMNH 1969.3.24.61 62, Kenya, Ng ong River, tributary of Nairobi River, ca. 118'25"S, 3653'23"E (2: 123.6146.9, paratypes ); BMNH 1987.3.23.4, Kenya, Athi River, S. E. of Nairobi, ca. 126'32"S, 3659'25"E (1: 36.0, paratype); CAS SU 24169, ex. BMNH 1928.11.10.1318, (1: 87.1, paratype) ; MCZ 32518, ex. BMNH 1928.11.10.1318, (1: 96.5 paratype); MRAC A7 25P 1 2, ex. BMNH 1928.11.10.1318, (2: 78.082.0 paratypes ); SAIAB 87475, ex. BMNH 1909.11.15.2123, (1: 106.0, paratype); UF 167873, ex. BMNH 1928.11.10.1318, (2: 79.492.6, paratypes); USNM 72922, Kenya, Nairobi R., near Nairobi, ca. 110'33"S, 3656'24"E (2: 113.4159.4, paratypes ).

PAGE 122

122 Anoplopterus chalei : Rufiji River basin: CU 93744, Tanzania, Little Ruaha River at bridge in Ihembe on road from Iringa to Dabaga, altitude 1680 m, 754'41"S, 3547' 44"E (63: 32.2 108.5); UF 170728, same data as CU 93744, (63: 26.6 123.9). Anoplopterus grandis : E wa so N g iro River basin: BMNH 1912.22.120, Kenya, Eusso Mara, a swift mountain stream, tributary of Eusso Nyiro [= Ewaso Ngiro] ca. 04613N, 37 3327E (1: 166.0, holotype of A. oxyrhinus ); BMNH 1908.9.17.13 18, Kenya, NyiroNarok; Niro Narok system, elev. 40005000 ft., ca. 01518N, 36 3221E ( 5 : 93.4181.4); BMNH 1912.3.22.119, Kenya, Eusso Nyiro [= Ewaso Ngiro] below falls, ca. 04702N, 38 0504E (1: 113.6) SAIAB 87474, ex. BMNH 1908.9.17. 1318, (1 : 135.3) ; UF 177478, ex. BMNH 1908.9.17. 1318, (1 : 120.2) Tana River basin: BMNH 1904.12.23.5052, Kenya, Chania River of Tetse, Kenya, elev. 7000 ft., ca. 10126N, 37 0407E (2: 13 6.1 160.0, Syntypes); BMNH 1937.6.4.36 42, Kenya, Thika River, Ndula Falls, ca. 10258N, 37 0534E (12: 32.1 65.9); BMNH 1965.12.7.125, Kenya, Rogati River, Sagana, ca. 03953N, 37 1207E (1: 42.9); BMNH 1965.12.7.126128, same locality (3: 60.9 89.7); BMNH 1965.12.7.129, same locality (1: 156.8); BMNH 1965.12.7.130 131, same locality (2: 74.9 98.7); BMNH 1965.12.7.132, same locality (1: 51.8); BMNH 1966.6.28.23, same locality (2: 65.9102.6); BMNH 1966.8.25.18, same locality (1: 34.8); MRAC 7448P 12 13, Kenya, upper Tana, side creek, ca. 0050'S, 3715'E (2: 52.660.3). Unknown river basin: CAS SU 66020, Kenya, East Macania [Makania] River (1: 43.5). Anoplopterus krefftii : Galana River basin: BMNH 1969.3.24.6369, Kenya, Tsabo [Tasvo] River tributary of Athi [Galana] River, mountain Mbololo, ca. 31450 S 38 2745E (7: 43.6 99.5); BMNH 1969.3.24.70, Kenya, Voi River, Teita hills, Voi

PAGE 123

123 District, 32651 S 38 2916E (1: 106.6). Lake Jipe basin: BMNH 1966.6.28.1, Tanzania, Jipe River, ca. 343 45 S 37 4506E (1: 69.7). Pangani River basin: BMNH 1905.7.25.4142, Kenya, Kibosho, Kilimandjaro, ca. 3 15' S 37 19' E (2: 77.9 94.9); BMNH 1968.10.25.3, Tanzania, River Lume (upper reaches of Pangani River) at source, 5 miles east of Lake Chala, ca. 32322 S 37 4345E (1: 106.6); BMNH 1968.10.25.8, Tanzania, River Lume (upper reaches of Pangani River) east of Taveta, ca. 32322 S 37 4345E (1: 67.2); BMNH 1969.1.15.1, Tanzania, From a stream at Arusha in the foothills of Mount Meru, ca. 31723 S 36 4537E (1: 84.7); CU 93726, Tanzania, Una River directly below Kinukamori Falls near Marangu, altitude 1343 m, 31639 S 37 3110E (5: 37.1 208.0); CU 93728, Tanzania, Kikuletwa River at below dam along road from Moshi to Samanga, altitude 757 m, 32630 S 37 1811E (1 3 : 30.1 80.3); CU 93729, Tanzania, Kikavu River at bridge on road from Moshi to Arusha, altitude 413 m, 31908 S 37 1305E (5: 36.8 85.1); CU 93736, Tanzania, Mkuzi River directly below Soni Falls on road from Mombo to Losh oto, altitude 1428 m, 45051 S 38 2158E (17 : 37.3 118.3); CU 93737, Tanzania, Mkuzi River along road from Mombo to Loshoto, altitude 384 m, 45204S, 38 2052E (34: 34.7 109.4) ; CU 93738, Tanzania, Pangani River at bridge in Hale on road from Muhez a to Segera, altitude 226 m, 51750 S 38 3613E (6: 34.9 60.7); FMNH 111684, Tanzania, West Usambara Mts., in river (stream) near Ambangulu Tea Estate factory, ca. 50454 S 38 2555E (1: 29.0); MRAC 2010 08P 3 5, ex. UF 170744, (3: 46.651.5); SAI AB 87473, ex. UF 170744, (3: 43.053.2); UF 170704, Tanzania, Pangani River along road from Same to Korogwe, altitude 350 m, 50806 S 38 2341E (1:65.3); UF 170712, same data as CU 93738, (6: 32.4 72.2); UF 170719, same data as CU 93729, (5: 42.9 89. 7); UF 170721,

PAGE 124

124 same data as CU 93728, (12 : 30.0 115.9); UF 170722, same data as CU 93736, (18 : 40.5 139.3); UF 170724, same data as CU 93726, (7: 39.9 201.5) ; UF 170744, same data as CU 93737, (28: 33.4 123.9) Sigi River basin: BMNH 1909.10.19.2627, T anzania, Usambara, in rivulet running from Anani Hills to Sigi R., ca. 5 0' S 38 48'E (2: 47.5 86.0, Syntypes); BMNH 1968.10.25.56, Tanzania, Kisiwani, near Amani, East Usambura mountains, elev. 1,475 ft., ca. 51134 S 38 3739E (2: 36.2 43.1); BM NH 1968.10.25.7, Tanzania, River Sigi, east Usumbura Mountains, ca. 5 0' S 38 48'E (1: 72.7); CAS 63737, Tanzania, Zigi [Sigi] R., Corn mill, in the Usambara Mountains, 10 km southeast (via the Amani Muheza Road) of Amani, ca. 50744 S 38 4257E (1: 47.8); CU 93735, Tanzania, Kihuhwi River at bridge on road from Muheza to Amani, altitude 225 m, 50739 S 38 4123E (1: 63.0); FMNH 111678, Tanzania, East Usambara Mountains, 4.5 km ESE Amani, Monga Tea Estate, ca. 5 05' S 38 36'E (1: 112.0); MCZ 51041, Tanzania, Amani, Usambara Mts., ca. 5 05' S 38 36'E (1: 90.0); UF 170713, same data as CU 93735, (2: 70.3 73.0).

PAGE 125

125 Table 41 Morphometric data for Anoplopterus uranoscopus. Range (n=51) MeanSD %SL Head length 24.1 26.7 25.30.7 Head width 19. 6 22.4 20.90.7 Head height 10.0 11.9 11.40.4 Body depth 11.7 16.4 13.51.1 Body depth at anus 12.7 16.3 14.20.6 Predorsal length 35.0 38.6 36.40.9 Prepectoral length 17.6 20.5 19.50.7 Preanal length 72.0 77.1 74.71.1 Dorsal fin base length 8.3 13.6 10.71.1 Adipose fin base length 23.4 28.9 25.91.4 Anal fin base length 7.0 12.1 10.10.9 Pelvic fin length 17.3 22.5 20.01.2 Pectoral fin length 20.2 24.6 22.21.0 Anal fin length 14.9 19.8 17.60.9 Caudal peduncle length 15.1 17.9 16.60.8 Caudal peduncle depth 10.9 14.3 11.90.8 Anus to anal fin length 9.2 13.8 11.71.2 Prepelvic length 48.2 54.1 51.61.3 Postpelvic length 47.2 52.8 49.21.3 Dorsal fin insertion to adipose 42.3 49.5 46.31.9 Dorsal fin origin to caudal 64.1 69.8 67.1 1.3 Preanus length 59.4 64.9 62.41.1 %HL Snout length 43.7 52.7 48.12.2 Interorbital distance 25.2 28.0 26.50.8 Maxillary barbel length 69.5 99.1 83.68.0 Inner mandibular barbel length 24.4 44.0 32.95.3 Outer mandibular barbel length 42.8 69 .8 55.66.5 Eye diameter 10.2 17.5 13.81.6

PAGE 126

126 Table 42 Meristic traits, caudal fin shape, and coloration diagnostic for species of Anoplopterus Species Gill rakers on first arch Branched anal fin rays Total vertebrae 1st dorsal pterygiophore intercept count Caudal fin shape Coloration Anoplopterus uranoscopus 9 11, rarely 8 6 or 7, rarely 5 usually 3839, rarely 40 1 Forked Grey with diffuse stripe along side; distinct pale patches at origin and insertion of dorsal fin Anoplopterus leroyi 8 10, r arely 7 6 or 7 usually 37, rarely 36 1, rarely 2 Forked Brown/yellow with diffuse stripe along side; distinct pale patches at origin and insertion of dorsal fin Anoplopterus hargeri 8 or 10, rarely 7 or 11 6, rarely 5 or 7 36 37 2, rarely 3 Forked Brown ; black spots/blotches usually present; distinct pale patches at origin and insertion of dorsal fin; juveniles with large black blotches and may appear mottled Anoplopterus brevidorsalis 7 or 8, rarely 6 or 9 6 or 7, rarely 5 37 38 1 Forked Brown/yellow ; black spots/blotches usually present; distinct pale patches at origin and insertion of dorsal fin; juveniles with large black blotches and may appear mottled Anoplopterus n. sp. Mangula 8 or 9, rarely 7 6, rarely 5 36 37 1 Emarginate Grey/yellow; no d ark marking on body; distinct pale patches at origin and insertion of dorsal fin

PAGE 127

127 Table 42 Continued Species Gill rakers on first arch Branched anal fin rays Total vertebrae 1st dorsal pterygiophore intercept count Caudal fin shape Coloration Anoplopter us n. sp. Ruvu 6 or 7, rarely 8 5, rarely 6 35 36, rarely 37 2, rarely 1 Emarginate Heavily spotted, spots usually coalesced to form vermiculations; distinct pale patches at origin and insertion of dorsal fin Anoplopterus grandis 8 10, rarely 6, 7 or 11 6 usually 3739, rarely 40 1 or 2 Forked Uniformly light brown or brown with large dark spots or blotches; no distinct pale patches at origin and insertion of dorsal fin Anoplopterus krefftii 8 10, rarely 7 or 11 5 or 6, rarely 7 usually 36, rarely 37 or 38 1, rarely 2 Emarginate Brown/yellow; black spots/blotches rarely present; distinct pale patches at origin and insertion of dorsal fin Anoplopterus chalei 9 10, rarely 11 5 or 6 usually 38, rarely 39 1 Forked Usually dark blotches alongside but may be uniformly brown; usually distinct pale patches at origin and insertion of dorsal fin Anoplopterus athiensis 9 or 10, rarely 11 5 or 6, rarely 7 usually 38, rarely 39 or 40 1, rarely 2 Forked Brown and finely spotted; no distinct pale patches at origin and insertion of dorsal fin

PAGE 128

128 Table 43 Morphometric traits diagnostic for species of the Anoplopterus Species Caudal peduncle depth (% SL) Caudal peduncle length (% SL) Interorbital width (% HL) Head depth (% SL) Predorsal length (% SL) Body depth at an us (% SL) Maxillary barbel length (% HL) Anoplopterus uranoscopus 10.9 14.3 15.1 17.9 25.2 28.1 10.0 11.9 35.0 38.6 12.7 16.3 69.5 99.1 Anoplopterus leroyi 10.0 13.4 14.9 17.8 25.8 30.2 11.7 13.8 34.7 36.8 13.0 16.0 69.0 94.8 Anoplopterus hargeri 11.8 13.5 14.4 16.8 25.7 31.7 12.3 13.2 37.3 38.4 13.0 14.4 62.2 95.9 Anoplopterus brevidorsalis 10.7 13.7 16.4 17.3 22.2 25.3 11.3 13.9 35.9 38.9 12.5 14.0 64.9 86.0 Anoplopterus cubangoensis 8.1 10.9 17.7 21.7 19.8 23.9 10.0 13.3 36.0 38.5 10.2 12.2 51.4 78.4 Anoplopterus n. sp. Mangula 9.9 12.4 18.2 20.3 24.3 27.2 10.8 13.1 31.6 34.9 11.6 13.1 71.4 92.6 Anoplopterus n. sp. Ruvu 10.3 12.3 16.0 18.2 20.7 24.9 11.2 13.6 36.1 39.9 12.8 15.7 37.3 49.6 Anoplopterus grandis 9.0 12.6 14.6 19.0 22.1 32.6 11.3 1 4.8 33.7 40.4 11.0 14.4 49.7 83.8 Anoplopterus krefftii 10.2 13.5 14.5 22.7 22.5 29.6 9.1 14.1 31.8 38.3 11.1 16.5 41.3 93.6 Anoplopterus chalei 7.1 9.7 19.0 21.9 17.9 24.2 9.7 12.4 31.5 36.4 8.9 10.8 50.6 90.7 Anoplopterus athiensis 10.1 12.6 14.6 18.3 23.5 32.5 11.3 14.9 35.2 39.6 11.1 15.3 69.2 88.0

PAGE 129

129 Table 44 Morphometric data for Anoplopterus leroyi Range and mean include the holotype. Holotype Range (n=31) MeanSD %SL Head length 25.2 24.1 26.5 25.40.8 Head width 22.0 20.3 23.6 21.40.7 Head height 11.7 11.7 13.8 12.70.5 Body depth 14.8 12.6 15.9 14.20.9 Body depth at anus 13.5 13.0 16.0 14.40.7 Predorsal length 34.9 34.7 36.8 35.90.6 Prepectoral length 19.4 18.0 22.4 19.81.2 Preanal length 75.6 71.2 77.7 73.81.6 Dorsal fin ba se length 9.9 9.9 13.0 11.30.8 Adipose fin base length 25.3 21.9 25.7 24.01.1 Anal fin base length 8.8 8.8 13.4 11.21.1 Pelvic fin length 18.9 18.1 21.5 19.81.0 Pectoral fin length 19.7 19.7 23.7 22.31.0 Anal fin length 16.1 16.1 19.1 17.80.9 C audal peduncle length 16.6 14.9 17.8 16.50.6 Caudal peduncle depth 11.7 10.0 13.4 12.20.8 Prepelvic length 52.0 47.1 54.9 50.22.1 Postpelvic length 47.9 47.3 53.2 50.31.8 Dorsal fin insertion to adipose 50.9 44.5 50.4 46.41.4 Dorsal fin origin to caudal 67.3 67.1 70.9 68.81.1 Preanus length 61.0 59.2 63.9 61.51.3 %HL Snout length 49.6 43.1 51.6 47.42.1 Interorbital distance 26.8 24.6 30.2 27.01.3 Maxillary barbel length 78.5 69.0 94.8 78.88.1 Inner mandibular barbel length 46.3 27.8 46.3 34.44.9 Outer mandibular barbel length 58.3 50.6 69.9 57.94.8 Eye diameter 9.5 9.5 19.8 13.82.2

PAGE 130

130 Table 4 5 Morphometric data for Anoplopterus hargeri Range and mean include the holotype. Holotype Range (n=20) MeanSD %SL Head length 2 5.2 24.6 26.7 25.70.7 Head width 23.0 20.6 23.9 21.91.0 Head height 13.1 12.3 13.2 12.70.3 Body depth 13.1 12.3 15.9 14.41.1 Body depth at anus 13.1 13.0 14.4 13.60.4 Predorsal length 38.3 36.2 39.2 37.80.8 Prepectoral length 18.4 18.4 21.5 20. 00.9 Preanal length 75.6 72.7 77.7 74.91.4 Dorsal fin base length 8.8 8.7 12.1 10.31.1 Adipose fin base length 20.3 18.0 22.9 20.31.7 Anal fin base length 10.9 8.1 10.9 9.81.1 Pelvic fin length 18.2 18.5 20.9 19.70.8 Pectoral fin length 20.3 20 .2 25.4 22.51.4 Anal fin length 18.9 16.4 19.9 17.61.0 Caudal peduncle length 16.4 14.4 16.8 16.10.8 Caudal peduncle depth 11.9 11.8 13.5 12.50.5 Prepelvic length 52.3 50.2 53.2 51.60.9 Postpelvic length 50.1 46.8 50.8 49.11.2 Dorsal fin insert ion to adipose 44.3 41.6 48.1 44.12.1 Dorsal fin origin to caudal 66.2 64.5 69.0 66.11.3 Preanus length 61.5 59.9 63.8 62.01.2 %HL Snout length 52.5 45.9 53.9 50.42.5 Interorbital distance 31.7 25.7 31.7 28.62.0 Maxillary barbel length 95.9 62.2 95.9 74.111.0 Inner mandibular barbel length 46.3 26.6 51.6 34.16.3 Outer mandibular barbel length 69.1 41.7 70.0 54.48.1 Eye diameter 13.8 11.0 15.9 13.21.5

PAGE 131

131 Table 4 6 Morphometric data for Anoplopterus brevidorsalis Range and mean include the holotype. Holotype Range (n=13) MeanSD %SL Head length 26.4 23.9 27.0 25.90.8 Head width 22.2 20.0 22.8 21.30.9 Head height 13.8 11.3 13.9 12.90.8 Body depth 16.4 12.8 16.4 14.21.1 Body depth at anus 13.3 12.5 14.0 13.30.5 Predorsal l ength 38.6 35.9 38.9 37.41.1 Prepectoral length 20.8 18.7 22.9 20.81.1 Preanal length 75.7 70.6 75.7 73.11.4 Dorsal fin base length 9.6 12.0 11.00.8 Adipose fin base length 19.8 24.2 22.61.4 Anal fin base length 9.8 12.2 11.20.7 Pelvic fi n length 18.2 17.0 21.7 19.51.3 Pectoral fin length 20.3 20.3 23.9 22.41.1 Anal fin length 17.3 17.3 19.9 18.80.9 Caudal peduncle length 17.3 16.4 17.3 16.80.3 Caudal peduncle depth 10.7 10.7 13.7 12.10.9 Prepelvic length 54.7 50.1 54.7 52.41.2 Postpelvic length 48.8 47.2 50.7 48.71.2 Dorsal fin insertion to adipose 40.6 45.5 43.11.7 Dorsal fin origin to caudal 61.6 66.5 65.21.4 Preanus length 59.7 64.0 61.71.5 %HL Snout length 43.4 43.1 50.6 47.12.7 Interorbital distance 24 .8 22.2 25.3 23.80.9 Maxillary barbel length 69.0 64.9 86.0 74.85.8 Inner mandibular barbel length 26.7 59.8 36.58.1 Outer mandibular barbel length 37.8 67.1 53.08.3 Eye diameter 10.6 10.6 16.0 13.21.6

PAGE 132

132 Table 4 7 Morphometric data for Anopl opterus cubangoensis Range and mean include the holotype. Syntypes (n=5) Range (n=59) MeanSD %SL Head length 25.3 26.7 24.6 28.2 26.60.8 Head width 19.7 23.6 19.0 23.6 21.00.9 Head height 10.2 12.2 10.0 13.3 12.10.8 Body depth 9.6 12.4 9.3 1 5.5 12.91.4 Body depth at anus 10.2 11.9 10.2 12.2 11.50.6 Predorsal length 36.9 38.5 36.0 38.5 37.20.8 Prepectoral length 19.3 22.6 18.9 24.6 20.81.3 Preanal length 73.5 77.6 69.7 77.6 72.81.7 Dorsal fin base length 9.4 10.2 9.9 13.0 11.30.8 A dipose fin base length 17.4 20.0 17.4 24.3 20.31.5 Anal fin base length 5.6 8.7 5.6 12.1 9.91.2 Pelvic fin length 16.7 18.9 15.3 20.9 18.11.2 Pectoral fin length 19.1 21.8 18.2 23.8 20.91.3 Anal fin length 15.8 18.2 14.8 20.0 17.51.2 Caudal pedun cle length 17.8 19.4 17.7 21.7 19.11.0 Caudal peduncle depth 8.1 9.6 8.1 10.9 9.70.6 Prepelvic length 50.0 55.1 50.0 5.7 51.91.2 Postpelvic length 46.8 50.7 45.2 51.9 49.41.4 Dorsal fin insertion to adipose 40.7 40.3 45.8 42.71.4 Dorsal fin origi n to caudal 63.5 62.3 70.3 66.51.8 Preanus length 59.2 58.4 63.7 60.91.3 %HL Snout length 43.9 49.2 43.4 51.1 47.41.9 Interorbital distance 19.2 23.6 17.5 23.9 22.11.3 Maxillary barbel length 58.7 70.7 51.4 78.4 63.56.3 Inner mandibular barb el length 24.4 22.4 40.9 30.44.2 Outer mandibular barbel length 34.8 33.5 66.2 47.17.4 Eye diameter 10.6 12.7 9.5 19.1 13.52.0

PAGE 133

133 Table 4 8 Morphometric data for Anoplopterus n. sp. Mangula. Range and mean include the holotype. Holotype Range (n=16) MeanSD %SL Head length 23.1 22.3 26.3 24.41.2 Head width 19.3 19.2 21.4 20.20.6 Head height 10.8 10.8 13.1 12.00.8 Body depth 13.1 12.1 15.2 13.10.9 Body depth at anus 12.6 11.6 13.1 12.70.5 Predorsal length 32.1 31.6 34.9 33.41.2 Prepec toral length 17.3 17.3 20.3 19.20.9 Preanal length 69.5 69.5 73.5 71.71.2 Dorsal fin base length 11.1 10.0 12.2 11.00.7 Adipose fin base length 26.5 25.3 27.2 26.40.6 Anal fin base length 10.6 9.1 11.1 10.10.7 Pelvic fin length 19.9 18.6 21.6 20. 10.9 Pectoral fin length 24.0 22.4 25.4 23.70.8 Anal fin length 17.0 16.0 18.9 17.30.7 Caudal peduncle length 20.0 18.2 20.3 19.00.6 Caudal peduncle depth 10.0 9.9 12.4 11.10.7 Prepelvic length 48.3 48.2 51.0 49.50.8 Postpelvic length 53.3 49.3 53.3 50.91.2 Dorsal fin insertion to adipose 48.5 44.5 49.7 47.21.7 Dorsal fin origin to caudal 70.2 67.0 71.1 69.01.5 Preanus length 59.9 58.6 61.9 60.51.0 %HL Snout length 49.6 44.7 51.3 48.12.0 Interorbital distance 26.8 24.3 27.2 25.81 .1 Maxillary barbel length 78.5 71.4 92.6 79.87.0 Inner mandibular barbel length 46.3 24.6 31.6 27.52.4 Outer mandibular barbel length 58.3 39.6 57.3 47.85.9 Eye diameter 9.5 8.2 12.5 10.01.1

PAGE 134

134 Table 4 9 Morphometric data for Anoplopterus n. sp. Ruvu Range and mean include the holotype. Holotype Range (n=16) MeanSD %SL Head length 23.8 23.5 25.8 24.70.7 Head width 20.7 19.9 22.4 21.10.6 Head height 12.2 11.2 13.6 12.30.6 Body depth 13.0 12.8 15.8 13.80.8 Body depth at anus 14.6 12. 8 15.7 14.20.8 Predorsal length 37.1 36.1 39.9 37.51.0 Prepectoral length 17.0 17.0 20.8 18.60.9 Preanal length 72.7 72.2 77.3 74.41.4 Dorsal fin base length 10.5 9.6 12.5 10.80.8 Adipose fin base length 23.6 21.4 26.1 23.91.2 Anal fin base len gth 9.4 8.2 10.0 9.10.6 Pelvic fin length 19.5 18.1 21.5 19.70.9 Pectoral fin length 22.6 21.0 24.9 22.61.0 Anal fin length 17.2 14.6 18.8 16.70.9 Caudal peduncle length 17.2 18.2 20.3 19.00.6 Caudal peduncle depth 11.7 16.0 18.2 17.30.5 Prepel vic length 50.4 48.1 52.3 50.41.2 Postpelvic length 51.2 48.8 52.7 50.61.0 Dorsal fin insertion to adipose 46.8 41.8 48.4 45.71.7 Dorsal fin origin to caudal 65.4 62.4 66.9 65.11.2 Preanus length 59.8 58.0 63.2 61.11.3 %HL Snout length 48.1 45.7 50.4 48.41.2 Interorbital distance 23.7 20.7 24.9 23.51.2 Maxillary barbel length 46.6 37.3 49.6 44.23.3 Inner mandibular barbel length 21.1 16.3 22.3 20.31.5 Outer mandibular barbel length 30.8 28.8 40.4 33.83.0 Eye diameter 16.5 14.2 19.1 15.71.4

PAGE 135

135 Table 4 10 Distribution of t otal vertebrae counts in species of Anoplopterus Total Vertebrae 35 36 37 38 39 40 Anoplopterus uranoscopus 9 5 Anoplopterus leroyi 3 13 Anoplopterus hargeri 2 5 Anoplopterus brevidorsalis 1 6 Anoplopterus cubangoensis 4 5 Anoplopterus n. sp. Mangula 4 7 Anoplopterus n. sp. Ruvu 9 4 1 Anoplopterus grandis 8 15 11 1 Anoplopterus krefftii 9 1 1 Anoplopterus chalei 9 1 Anoplopterus athiensis 8 3 2 Table 4 11 Distribution of 1st dors al pterygiophore intercept counts in species of Anoplopterus 1st dorsal pterygiophore intercept count 0 1 2 3 Anoplopterus uranoscopus 14 Anoplopterus leroyi 15 1 Anoplopterus hargeri 6 1 Anoplopterus brevidorsalis 1 6 Anoplopterus cubangoensis 9 Anoplopterus n. sp. Mangula 11 Anoplopterus n. sp. Ruvu 2 11 Anoplopterus grandis 22 13 Anoplopterus krefftii 11 1 Anoplopterus chalei 10 Anoplopterus athiensis 12 1

PAGE 136

136 Table 4 12 Distribution of t otal gillraker counts in species of Anoplopterus Total Gill Rakers 5 6 7 8 9 10 11 Anoplopterus uranoscopus 5 23 20 7 Anoplopterus leroyi 1 17 17 9 Anoplopterus hargeri 1 2 29 8 1 Anoplopterus brevidorsalis 1 8 20 3 Anoplopterus cubang oensis 1 8 57 26 7 Anoplopterus n. sp. Mangula 4 11 8 Anoplopterus n. sp. Ruvu 11 16 2 Anoplopterus grandis 6 3 5 18 7 1 Anoplopterus krefftii 11 37 89 39 3 Anoplopterus chalei 2 13 6 Anoplopterus athiensis 7 12 2 Table 4 13 Distribution of b ranched anal fin ray counts in species of Anoplopterus Branched anal fin rays 4 5 6 7 8 Anoplopterus uranoscopus 4 47 6 Anoplopterus leroyi 35 12 Anoplopterus hargeri 3 34 2 Anoplopterus brevidorsalis 3 22 7 Anoplopter us cubangoensis 33 62 Anoplopterus n. sp. Mangula 2 21 Anoplopterus n. sp. Ruvu 27 2 Anoplopterus grandis 36 Anoplopterus krefftii 32 133 21 Anoplopterus chalei 11 11 Anoplopterus athiensis 7 13 2

PAGE 137

137 Figure 41 Species of the Anoplopterus uranoscopus group recognized by Thomson & Page (2010). Lateral views of A: Anoplopterus uranoscopus UF 170723, 96.2 mm SL, Wami River basin, Tanzania; B: Anoplopterus cryptobullatus CU 91072, 88.5 mm SL, Congo River basin, Zambia; C: Anoplopterus chalei UF 170728, 123.9 mm SL Rufiji River basin, Tanzania; D: Anoplopterus grandis UF 177478, 120.2 mm SL, Ewaso Nigiro River basin, Kenya; E. Anoplopterus krefftii UF 170722, 139.3 mm SL, Pangani River basin, Tanzania; F. Anoplopterus athiensi s BMNH 1910.10.31.31, 150.0 mm SL, Galina River basin, Kenya. Scale bars equal 1 cm.

PAGE 138

138 Figure 42 Anoplopterus uranoscopus UF 170723 96.2 mm SL; lateral, dorsal and ventral view s.

PAGE 139

139 Figure 43 Distribution of Anoplopterus n. sp. Ruvu A. n. sp. Mangul a A. leroyi and A. uranoscopus

PAGE 140

140 Figure 44 Anoplopterus leroyi CU 93725 83.9 mm SL; lateral, dorsal and ventral view s.

PAGE 141

141 Figure 45 Anoplopterus hargeri SAIAB 34332 80.3 mm SL; lateral, dorsal and ventral view s.

PAGE 142

142 Figure 46 Distributions of Anopl opterus brevidorsailis A. hargeri and A. cubangoensis

PAGE 143

143 Figure 47 Anoplopterus brevidorsailis SAIAB 53691 85.8 mm SL; lateral, dorsal and ventral view s.

PAGE 144

144 Figure 48 Anoplopterus cubangoensis SAIAB 80539 69.7 mm SL; lateral, dorsal and ventral view s.

PAGE 145

145 Figure 49 Anoplopterus n. sp. Mangula, UF 170702 95.6 mm SL, holotype; lateral, dorsal and ventral view s.

PAGE 146

146 Figure 410 Anoplopterus n. sp. Ruvu UF 1707 27 75.4 mm SL, holotype; lateral, dorsal and ventral view s.

PAGE 147

147 CHAPTER 5 REVISION OF THE AMPHILIUS JACKSONII COMPLEX (SILURIFORMES: AMPHILIIDAE), WITH THE DESCRIPTION OF FIVE NEW SPECIES. Background The African catfish genus Amphilius includes 18 species of the subfamily Amphiliinae that are primarily distributed in western Africa. It is distinguishe d from the closely related genus Anoplopterus by having an epidermal fold at the base of the caudal fin and 6 + 7 or 7 + 8 principal caudal fin rays (vs. 8+9) ( Chapter 3 ) The distribution of Amphilius is primarily Low African while the distribution of Ano plopterus is primarily High African. Low Africa is northern and western Africa in which most of the land is at elevations well between 500 and 1000 ft. and High Africa is southern and eastern Africa in which most of the land is at elevations well above 100 0 ft. and much of it is at elevations above 4000 ft. (Roberts, 1975). Only two species of Amphilius occur in High Africa, Amphilius lentiginosus from Angola and Amphilius jacksonii which is widely distributed in eastern Africa. Amphilius jacksonii was desc ribed from a single specimen collected from the Hima River, a tributary of Lake George in western Uganda (Boulenger, 1912). Subsequent authors have recorded this species throughout much of eastern Africa including the Lake Edwards drainage in western Uganda (Walsh et al., 2000), the Malagarasi drainage in Burundi & western Tanzania ( David, 1937 ; De vos et al., 2001), the Rizizi drainage as Amphilius platychir ( Marlier, 1953), the Rufiji basin in eastern Tanzania as Amphilius platychir ( Matthes, 1967; Bailey 1969), the Lake Rukwa basin in Tanzania (Seegers, 1996), and the Upper Congo drainage in Zambia (Seegers, 1996). Beyond distributional data, very little published information is available on A mphilius jacksonii The species was described and figured in both editions of Fishes

PAGE 148

148 of Uganda (Greenwood, 1958, 1966), Seegers (1996) figured the type specimen and described specimens from the Lake Rukwa basin, and Walsh et al. (2000) redescribed the species and provided habitat and lifehistory notes from the Lake George drainage, Uganda. The Weberian complex of this species was described and figured by Chardon (1968), and Diogo & Chardon (2000) described the mandibular barbel structure There have been no studies that have examined geographic variation in A jac ksonii A study of geographic variation in specimens identified as A mphilus jacksonii revealed substantial morphological differences among populations from different drainages. Th e aim of this study is to determine the number of species that material ident ified as A. jacksonii represents, determine the distribution of each of the species, and formally describe each species. Materials and Methods Measurements were made point to point with digital calipers, and data were recorded to hundred th s of a millimeter Subunits of the head are presented as proportions of head length (HL). Head length and measurements of other body parts are given as proportions of standard length (SL). Use of the terms origin and insertion to designate, respectively, the most anterior and posterior points on the bases of all fins follows Cailliet et al. (1986). Counts and measurements were made on the left side of a specimen when possible and follow Skelton (1981, 1984, 1986) with the additional measurements of body depth at anus, prepectoral fin length, preanal fin length, dorsal fin base length, adiposefin base length, pelvic and pectoral fin lengths, prepelvic fin length, postpelvic fin length, and preanus length. Only specimens 40.0 mm SL or larger were measured, but counts were made on all type specimens. Descriptions of new species are based on the holotype and the paratypes. Specimens from drainages that

PAGE 149

149 do not include the type locality and specimens that were examined but from which no morphometric or meristic data was taken are listed as nontypes. For fin ray counts, numbers of unbranched soft rays are indicated by lower case Roman numerals, and branched soft rays by Arabic numerals. The number of anterior unbranched rays in the anal fin is difficult to determine, and the count s were checked with radiographs whenever possible. Amphiliids typically have a small spinelet in front of the first unbranched dorsal fin ray. The spinelet is not included in the counts. Branchiostegal ray counts include all b ranchiostegal ray s. Material e xamined is given under each species account and is listed by drainage followed by catalog number, country, locality, geographic coordinates and, in parentheses, the number of specimens and the size range in mm SL. Geographic coordinates preceded by ca. are estimated from other original locality data. Materials examined in this study are deposited in the following institutions: the American Museum of Natural History, New York, New York ( AMNH ), the Auburn University Museum Fish Collection, Auburn, Alabama ( AU M ), the Natural History Museum, London ( BMNH), the Cornell University, Vertebrate Collections, Ithaca, New York ( CU), the Harvard Museum of Comparative Zoology Cambridge, Massachusetts ( MCZ ), the Royal Museum of Central Africa, Tervuren, Belgium ( MRAC ) the South African Institute for Aquatic Biodiversity Grahamstown, South Africa ( SAIAB ), and the Florida Museum of Natural History, Gainesville, Florida ( UF ). Synonymies include all references to the species in east central Africa. The first page of the ref erence to the species and all figures are listed. If the species is also listed in a key on a separate page from the account, that page is also listed. The type of

PAGE 150

150 information in the reference is given followed by the locality for the species as given in t he account. Additionally, any specimens on which the account is known to be based are listed. If the account is only based in part on the species, only the information that is applicable to that species is listed. Accounts of species outside east central A frica that give only a general distribution in east central Africa are excluded. Taxonomic Descriptions Amphilius jacksonii complex The Amphilius jacksonii complex is distinguished from all other species of Amphilius by the following combination of charac ters: a variably mottled body coloration that includes dark saddles (vs. body coloration not mottled with dark saddles), head and body not heavily spotted (vs. head and body heavily spotted), cren e lations present on epidermal fold (vs. epidermal fold smoot h), lobes formed by epidermal fold crenelations rounded (vs. lobes formed by epidermal fold crenelations elongated and pointed), b ranchiostegal ray s usually 6 to 8 (vs. greater than 8), and total gill rakers usually 6 to 11 (vs. greater than 11). The Amphi lius jacksonii complex differs from all other species of the genus ( A. atesuensis, A. brevis, A. caudosignatus, A. dimonikensis, A. grammatophorus A. kakrimensis, A. korupi, A. lamani, A. longirostris, A. maesii, A. mamonekenensis, A. nigricaudatus A. op isthophthalmus, A. platychir A. pulcher, A. rheophilus ) except A. lentiginosus by its variably mottled body coloration that includes dark saddles that are joined laterally (vs. body coloration not mottled, without dark saddles that are joined laterally ). It differs from A. lentiginosus by having its head and body not spotted (vs. head and body heavily spotted) and by having fewer total gill rakers on the first gill arch ( 6 11, rarely 5 or 12 vs. 14 16). The Amphilius jacksonii complex differs from A.

PAGE 151

151 longi rostris and A. opisthophthalmus by having 6+7 principal caudal fin rays (vs. 7 + 8 ) and a well developed crenelated epidermal fold (vs. fold rudimentary ) and differs from A. brevis A. dimonikensis A. korup i and A. maesii by having distinct crenelations on the epidermal fold (vs. crenelations absent or very weak ). If differs from A. kakrimensis, A. platychir and A. rheophilus by the shape of the lobes formed by the crenelations on the epidermal fold (lobes rounded vs. elongate and pointed) and from A. brev is, A. caudosignatus A. dimonikensis, A. korupi, A. lamani, A. maesii, A. mamonekenensis, A. nigricaudatus and A. pulcher by having fewer branchiostegal ray s (usually 68, rarely 5 or 9 vs. 9 10 in A. pulcher more than 9 in the other species). Amphilius jacksonii Boulenger 19 12 (Fig ure 5 1; Table 5 1 ) Amphilius jacksonii Boulenger 19 12: 602 Original description, Type locality: Hima R. flowing into Lake George (Ruisamba), eastern Uganda, elev. 3500 ft. [ Lake George drainage, Nile basin] h olotype: BMNH 1912.10.15.47 ; Boulenger, 1916b: 307, fig. 182, description based on type; David & Poll, 1937: 255, description, Rutshuru River [Lake Edward drainage, Nile basin]; Poll, 1939: 17, 62, records from Albert National Park [Lake George and Lake Edwards drainages ]; Harry, 1953: 190, synonymy; Greenwood, 1957: 77, figs. 49, 80 description, size, habitat, distribution [Lake George and Lake Edwards drainages]; Copely, 1958: 154, Hima R.; Greenwood, 1958: 89, fig. 49, description, size, habitat, distribution [Lake Ge orge and Lake Edwards drainages]; Whitehead, 1958: 198 western Uganda rivers ; Corbet, 1961: 81 [ not examined ] ; Greenwood, 1966: 93, fig. 49, description, size, habitat, distribution [Lake George and Lake Edwards drainages]; Seegers, 1996: 188, figs. 133 1 34 (in part), type information, holotype figured, Seegers, 1996b: 251 (in part), type information, distribution; Wal sh et al., 2000: 166, redescription, diagnosis, description, lifehistory aspects, habitat, d istribution [Lake George drainage]. Material examined Lake George drainage: AMNH 97419, Uganda, Dura River, Dura station, south end of Kibale Forest Reserve near Queen Elizabeth National Park ca. 0 12' 00N, 30 22' 00"E ( 6 : 41.979.5); BMNH 1912.10.15.47 Uganda, Hima River, ca. 0 17' 19N, 30 10' 23"E (1: 86.7, holotype); BMNH 1971.1.5.3738, Uganda, Mapanga River, about 8

PAGE 152

152 km upstream just below rapids ca. 0 05' 10.6N, 30 23' 03.8"E ( 2 : 58.760.1); BMNH 1971.1.5.39, Uganda, Sibwe River ca. 0 10' 01.4N, 30 12' 36.8"E ( 1 : 87.0); BMNH 1971.2.19.1415 Uganda, Sibwe River ca. 0 10' 01.4N, 30 12' 36.8"E ( 1 : 87.0); CU 97331, ex. UF 110746 (4: 37.692.7); MCZ 100587, Uganda, Middle Dura River, Kanyanchu, Kibale Forest Reserve, ca. 0 27' 00N, 30 22' 00"E ( 1 : 76.7); MRAC 90 046P 68 78, Ugan da, Ruimi (Rwimi) River, road between Fort Portal Kasese [route A 109] ca. 0 22' 23.5N, 30 12' 40.8"E ( 11: 44.0 84. 2); MRAC 90 046 P 7987, Uganda, Ruimi (Rwimi) River, road between Fort Portal Kasese [route A 109] ca. 0 22' 23.5N, 30 12' 40.8"E ( 1 1 : 4 7 2 8 3 0); SAIAB 187245, ex. UF 110746 (42.3 97.0); UF 110743 Uganda, Middle Dura River, Kanyanchu, Kibale Forest Reserve, ca. 0 27' 00N, 30 22' 00"E ( 12: 48.6 104.3); UF 110744, Uganda, Middle Dura River, Kanyanchu, Kibale Forest Reserve, ca. 0 2 7' 00N, 30 22' 00"E ( 19: 23.6 115.9); UF 11074 5, Uganda, Middle Dura River, Kanyanchu, Kibale Forest Reserve, ca. 0 27' 00N, 30 22' 00"E ( 5 : 54.688.6 ); UF 11074 6, Uganda, Middle Dura River, Kanyanchu, Kibale Forest Reserve, ca. 0 27' 00N, 30 22' 00"E ( 36 : 34.1 102 .0); UF 110747, Uganda, Dura River Mainaro, ca. 0 37' 00N, 30 16' 00"E ( 34 : 54.4 111.1); UF 110748 Uganda, Ruimi River at bridge on road between Fort Portal and Hima, ca. 0 22' 22.2N, 30 12' 41"E ( 41: 42.7 92.8 ); UF 110749, Uganda, Ruimi River at bridge on road between Fort Portal and Hima, ca. 0 22' 22.2N, 30 12' 41"E ( 1 : 47.6); UF 110750, Uganda, Mubuku River at bridge on road between Fort Portal and Kasese, ca. 0 15' 30.8N, 30 07' 9.7"E ( 7 : 52.885.1 ). Lake Edward drainage: AUM 47138, Uganda, Munyage River, below Butogota town, DRC Uganda border foot path, elev. 1190m 0 53' 00S, 29 37' 50"E ( 3 : 60.169.9 ); AUM 4713 9, Uganda, Tributary of Ishasha at bridge between Karangara

PAGE 153

153 and Butogota. Outside BINP, sources in park, elev 1180m 0 52' 18S, 29 39' 59"E ( 2 : 73.4 1 80.8); AUM 47143, Uganda, River Ishash, below bridge in cattle farm on ButogotaKanungu road. elev. 1300m 0 49' 50S, 29 38' 21"E ( 2 : 65.773.4); AUM 47146, Uganda, Bwindi Impenetrable Forest National Park c a. 0 1' 26S, 29 41' 23"E ( 4 : 58.886.6 ); MRAC 22551, Uganda, Rutshuru, Buseregenyi ca. 1 04' 00S, 29 26' 00"E ( 1 : 80.0); MRAC 23233, Uganda, Rutshuru, Buseregenyi ca. 1 04' 00S, 29 26' 00"E ( 1 : 82.2); MRAC A5 019 P 00480056, Uganda, Munyage Riv er, below Butogota town, DRCUganda border foot path, elev. 1190m ca. 1 04' 00S, 29 26' 00"E ( 9 : 39.497.3 ); UF 169241 ex. AUM 47143 (3: 65.9 70.7 ); UF 169242 ex. AUM 47139 (3: 80.787.3 ); UF 169243 ex. AUM 47138 (2: 67.1 86.3 ); UF 169257 ex. AUM 4 7146 (3: 60.775.1 ). Kagera River drainage: MRAC 164876, Rwanda, Mwogo River, 18 km NE of Butare ca. 2 29' 00S, 29 38' 00"E ( 1 : 77.2 ); MRAC 71783850, Rwanda, Mwogo River, the source of the Nile, ss tributary of Kagera River ca. 2 22' 52S, 29 41' 4 8"E ( 61: 38.5 107.0); MRAC 71851925, Rwanda Mwogo River, the source of the Nile, ss tributary of Kagera River ca. 2 22' 52S, 29 41' 48"E ( 74 : 38.5 106.4 ); MRAC 83417 Burundi, Urigoli, Nyalugogo [Nyabugogo?] River, Urigoli territory ca. 3 25' 26S, 29 53' 30"E ( 1 : 34. 7); MRAC 86 27 P 00880095, Rwanda, Rubondo River, t rib utary of Kiryango River, near town of Mukugi ca. 2 12' 00S, 29 41' 00"E ( 8 : 51.8 86.1 ); MRAC 91 30P 02960299, Burundi Nyakijanda River, t rib utary of Ruvubu River, the bridge on Road of General interest 4, 38 km from Kinyinya, ca. 3 36' 00S, 30 07' 00"E ( 4 : 49.565.0 ); MRAC 91 34P 01470150, Burundi Kavuruga River, tributary of Ruvubu River, near town of Buhinyuza, ca. 3 02' 00S, 30 21' 00"E ( 4 : 49.4 74.9); MRAC 91 34P 01510152, Burundi Kinyanderama River, t rib utary of Ruvubu River, near town

PAGE 154

154 Buhinyuza, ca. 3 02' 00S, 30 21' 00"E ( 2 : 66.6 71.1 ); MRAC 91 34 P 01600164, Burundi Nyabiko River,on MuyingaGitega Road, small t rib utary rocky and sandy bottom ca. 5 2 9' 00S, 30 14' 00"E ( 5 : 35.1 92.3); MRAC 91757763, Rwanda, Nyabugogo River at confluence of Lusine River ca. 1 47' 41S, 30 07' 23"E ( 7 : 58.996.8); MRAC 91764836, Rwanda, Nyabugogo River at confluence of Lusine River ca. 1 47' 41S, 30 07' 23"E ( 72 : 34.9 79.8 ); MRAC 93097120, Rwanda, Nyabugogo River, at outlet of Lake Mohasi ca. 1 46' 45S, 30 07' 58"E ( 23 : 28.2 77.0); MRAC 94272276, Rwanda, ca. 2 32' 15S, 29 40' 16"E ( 5 : 40.182.9); MRAC 96031.16111614, Burundi Karuzi, small tributary stream of the fishfarming ponds, Ndurumu system, ca. 3 06' 05S, 30 09' 53"E ( 4 : 29.3 78.4). Diagnosis Diagnostic characters are summarized in Table 5 2 Amphilius jacksonii is diagnosed from all other species of the Amphilius jacksonii complex by its m ore slender caudal peduncle (4.87.9% SL vs. 8.1 12.3% SL). It is further distinguished from A. n. sp. Malagarasi A. n. sp. Congo, and A. n. sp. Rufiji by its longer caudal peduncle (caudal peduncle length 16.720.6% SL vs. 13.318.8% SL), and from A. n. sp. Congo A. n. sp. Rufiji, and A. n. sp. Lake Kyogo by its more slender body (body depth at anus 9.6 13.2 % SL vs. 13.5 17.4% SL). It is further distinguished from A. n. sp. Congo by having fewer total gill rakers on the first gill arch ( 6 9, rarely 10 vs 10 11, rarely 9 or 12) and from A. n. sp. Ruzizi by having a wider interorbital width ( 26.732.0 % HL vs. 23.4 25.1% HL). Description Morphometric data are in Table 2 Body elongate, ventral profile flattened ventrally to anal fin base, then tapered dorsally to end of caudal peduncle. Dorsal profile

PAGE 155

155 rising gently from tip of snout to origin of dorsal fin, then nearly horizontal to end of caudal peduncle. Greatest body depth at dorsal fin origin. Caudal pedunc le laterally compressed, with crenellated epider mal fold. Anus and urogenital opening located at midpoint of adpressed pelvic fin, closer to insertion of pelvic fin than to origin of anal fin. Skin smooth. Lateral line complete, extending from dorsal edge of opercular cavity to base of caudal fin. Head and anterior part of body depressed and broad. Head wedgeshaped in lateral view. Snout broad, blunt when viewed from above. Head becoming wider from tip of snout to pectoral fin base. Branchiostegal membranes moderately joined at isthmus forming a U shap ed connection. Mouth broad, gently curved, subterminal. Lips moderately fleshy, slightly papillate. Rictal lobe large and papillate. Anterior portion of premaxillary tooth band exposed with mouth closed. Premaxillary tooth patches joined, forming V shaped band with anterior broad protrusion. Premaxillary and d entary t eeth short conical. Dentary tooth patches forming U shaped band, separated medially. Three pairs of simple, tapered circumoral barbels. Maxillary barbel large, fleshy and flattened with pointed tip; barbel extending posterolaterally from corner of mouth to pectoral fin base. Outer mandibular barbel thin with pointed tip, origin at posterior corner of lower jaw, extending to origin of pectoral fin. Inner mandibular barbel originates anterolaterally of inner mandibular barbel, extending to edge of branchiostegal membrane. Branchiostegal membrane with 6 (5), 7 ( 85), 8 (147 ) or 9 ( 1 ) rays. Gill rakers on first epibranchial 1 (4), 2 ( 178 ) or 3 (47); rakers on first ceratobranchial 4 ( 22),

PAGE 156

156 5( 88), 6 ( 99), 7 (1 5 ) or 8 (1); total gill rakers on first arch 6 ( 25), 7 ( 74 ), 8 (83 ), 9 (38 ), or 10 ( 6 ) Eye small, positioned dorsolaterally approximately midway between tip of snout and posterior margin of operculum. Horizontal diameter of eye slightly wider than vertical diameter. Eye without free orbit covered with skin confluent with dorsal surface of head. Anterior and posterior nares with prominent tubular rims; nares separate but relatively close to each other. Posterior nare located about midway between eye and tip of snout. Dorsal fin origin at point over tip of adpressed pectoral fin. Dorsal fin with i,6 ( 243) rays, margin straight Pectoral fin with i,7 (1), i, 8 ( 104 ), or i, 9 ( 137 ) rays ; unbranched ray greatly thickened. Pectoral fin with four or five innermost rays progressively shorter making posterior fin margin rounded. Origin of p elvic fin posterior to dorsal fin insertion Pelvic fin with i,5 ( 243 ) rays with first ray unbranched and greatly thickened. Pelvic fin with straight posterior margin. Adi posefin base longer than anal fin base, origin anterior to origin of anal fin base; fin extending past anal fin insertion. Margin strongly convex with sharply rounded edge, not deeply incised posteriorly. Caudal fin deeply forked with tips of lobes rounde d ; fin with i, 5 5 ,i ( 8 ) i, 5 6 ,i ( 230) or i, 6 6 ,i (1) principal rays. Anal fin with short base, origin posterior to origin of adiposefin base, with ii,6 (62), ii,7 (107), ii,8 (9), iii,6 ( 48) iii,7 (16), or iv7 (1) rays. Anal fin margin almost straight. Coloration Body variably mottled and with dark saddles. First saddle immediately posterior of head, second saddle under dorsal fin, third saddle between dorsal and adipose fins, fourth saddle under anterior of adipose fin, and fifth saddle on caudal peduncle. All

PAGE 157

157 saddles connected laterally by broad stripe. Ventral region light brown with fourth and fifth saddles extending around ventral side. Dorsal and anal fins light brown with dark medial band. Adipose fin dark brown, cream colored distally. Pectoral and pelvic fins positioned horizontally with upper surfaces brown and lower surfaces light yellow ; each with dark medial band. Caudal fin brown with medial cream patch and tips of upper and lower lobes cream. Caudal coloration asymmetrical, lower lobe with more pigment than upper lobe. Distribution Lake George, Lake Edwards, and Kagera River drainages, Nile basin, eastern U ganda, Rwanda, and Burundi (Figure 5 2) Amphilius n. sp. Ruzizi (Figure 5 3 ; Table 5 3 ) Amphilius platychir (non Gnther) Marlier, 1953: 194 Distribution in Ruzizi drainage; Distribution mapped. Holotype: MRAC 93294 Rwanda, Lufiro River, upper reaches of Matchuza, Ruzizi River drainage, ca. 2 44' 00S, 29 02' 99"E ( 1 : 82.8) Paratypes: CU 97332, same data as holotype (1: 47.7); MRAC 91 529 531, Democratic Republic of the Congo, Kiliba River ca. 3 14' 30S, 29 09' 36E (3: 31.264.9 ); MRAC 93295297, same data as holotype, ( 3 : 41.3 45.3) ; MRAC 93126 128, Burundi Nyakagunda River ca. 2 47'S, 29 04'E (3: 37.5 66.5 ); MRAC 93129130, B urundi Nyamagana River ca. 2 55'S, 29 08'E (2: 116.1120.6); MRAC 93131, Burundi Lua River ca. 2 46'S, 29 02'E (1: 79.5 ); SAIAB 187262, same data as holotype (1:64.2); UF 184236 (69.8).

PAGE 158

158 Non types Mutimbuzi River drainage: MRAC 91330389, Burundi M urago River ca. 3 17'S, 29 23'E ( 59: 29.271.4); MRAC 93076096, Burundi Musazi River [=Muzazi River ], ca. 3 17'S, 29 25'E (20: 33.191.0); MRAC 126266267, Burundi Murago River ca. 3 17'S, 29 23'E ( 7 : 41.3 82.8). Ruzizi River drainage: MRAC 91034.01530159, Burundi Nyamagana River near town of Mabayi ca. 2 42'S, 29 15'E (7: 47.779.9); MRAC 91390, Burundi Nyakagunda River, Bugarama, ca. 2 47'S, 29 04'E (1: 25.6); MRAC 93051054, Democratic Republic of the Congo Muniowe River, Luvungi ca. 2 56'S, 28 57'E (4: 43.9 95.2 ); MRAC 9305 5 075, Democratic Republic of the Congo, Luvubu River ca. 2 52'S, 29 02'E (20: 27.3108.5); MRAC 93121125, Burundi Nyamagana River ca. 2 55'S, 29 08'E (5: 44.1 97.6 ), MRAC 96031.15711576, Burundi Nyamag ana River, 3km after Mabayi, center Kivogero, ca. 2 42' 56S, 29 14' 40E (5: 46.288.0 ); MRAC 96031.1590 1595, Ikibenga River, 26 km on Cibitoke Mabayi road, ca. 2 43' 03S, 29 10' 05E (6: 50.5 69.3 ). Unkn o w n drainage: MRAC 91392395, Burundi Kikoma River ca. 3 20'S, 29 18'E (4: 45.386.6 ). Diagnosis Diagnostic characters are summarized in Table 5 2 A. n. sp. Ruzizi is diagnosed from A. n. sp. Malagarasi A. n. sp. Congo, and A. n. sp. Rufiji by its more slender caudal peduncle (caudal peduncle depth 8.6 9.5 % SL vs. 9.712.3% SL) and from A. jacksonii A. n. sp. Malagarasi A. n. sp. Rufiji, and A. n. sp. Lake Kyogo by its narrower interorbital width ( 23.4 25.1% HL vs. 26.735.8% HL). It is further diagnosed from A. n. sp. Congo, A. n. sp. Rufiji, and A. n. sp. Lake Kyogo by its more slender body (body depth at anus 11.1 12.9% SL vs. 13.5 17.4% SL). A. n. sp. Ruzizi is further diagnosed from A. n. sp. Congo by having fewer branchiostegal rays (67 vs. 8 9), fewer total gill

PAGE 159

159 rakers on the first gill arch ( 6 8, rarely 5 or 9 vs. 10 11, rarely 9 or 12), and a longer caudal peduncle (caudal peduncle length 18.720.3 % SL vs. 14.416.4 % SL). It is further diagnosed from A. n. sp. Rufiji by a longer caudal peduncle (caudal peduncle length 18.7 20.3 % SL vs. 13.315.5% SL) and shorter dorsal fin insertion to adiposefin insertion length ( 38.2 41.5 % SL vs. 42.2 44.6% SL). It is further diagnosed from A. jacksonii by its deeper caudal peduncle (caudal peduncle depth 8.6 9.5 % SL vs. 4.8 7.9 % SL). Description Morphometric data are in Table 5 3 Body elongate, ventral profile flattened ventrally to anal fin base, then tapered dorsally to end of caudal peduncle. Dorsal profile rising gently from tip of snout to origin of dorsal fin, then nearly horizontal to end of caudal peduncle. Greatest body depth at dorsal fin origin. Caudal pedunc le laterally compressed, with crenellated epidermal fold. Anus and urogenital openings located at midpoint of adpressed pelvic fin, much closer to insertion of pelvic fin than to origi n of anal fin. Skin smooth. Lateral line complete, extending from dorsal edge of opercular cavity to base of caudal fin. Head and anterior part of body depressed and broad. Head wedgeshaped in lateral view. Snout broad, blunt to rounder when viewed from above. Head becoming wider from tip of snout to pectoral fin base. Branchiostegal membranes moderately to broadly joined at isthmus forming a U shaped connection. Mouth broad, gently curved, subterminal. Lips moderately fleshy, slightly papillate. Rictal lobe large and slightly papillate. Anterior portion of premaxillary tooth band exposed with mouth closed. Premaxillary tooth patches joined, forming cresent

PAGE 160

160 shaped band. Premaxillary and dentary t eeth short conical. Dentary tooth patches forming U shaped band, separated medially. Three pairs of simple, tapered circumoral barbels. Maxillary barbel large, fleshy and flattened with pointed tip; barbel extending posterolaterally from cor ner of mouth, not reaching pectoral fin base. Outer mandibular barbel thin with pointed tip, origin at posterior corner of lower jaw, extending to just short of origin of pectoral fin. Inner mandibular barbel originates anterolaterally of inner mandibular barbel, extending to about twothirds of distance to edge of branchiostegal membrane. Branchiostegal membrane with 6 (7), or 7 ( 8) rays. Gill rakers on first epibranchial 2 ( 15) or 3 (1 ); rakers on first ceratobranchial 3 (1), 4 ( 8 ), 5(2 ), 6 (4 ), or 7 (1) ; total gill rakers on first arch 5 (1), 6 ( 8 ), 7 (1 ), 8 ( 5 ), or 9 ( 1). Eye small, positioned dorsolaterally approximately midway between tip of snout and posterior margin of operculum. Horizontal diameter of eye slightly wider than vertical diameter. Eye without free orbit; covered with skin confluent with dorsal surface of head. Anterior and posterior nares with prominent tubular rims; nares separate but relatively close to each other. Posterior nare located about midway between eye and tip of snout. Dorsal fin origin at point over or just posterior to tip of pectoral fin. Dorsal fin with i,6 ( 16) rays, and fin margin straight Pectoral fin with i, 8 ( 2 ), or i, 9 ( 14 ) rays ; unbranched ray greatly thickened. Pectoral fin with four or five innermost rays progressively shorter making posterior fin margin rounded. Origin of p elvic fin p osterior of dorsal fin insertion Pelvic fin with i,5 ( 16) rays with first ray unbranched and greatly thickened. Pelvic fin with straight posterior margin.

PAGE 161

161 Adiposefin base longer than anal fin base, origin anterior to origin of anal fin base, fin extending past anal fin insertion. Margin strongly convex with sharply rounded edge, deeply incised posteriorly. Caudal fin moderately forked with tips of lobes rounded; fin with i, 5 6 ,i ( 15) principal rays. Anal fin with short base, origin posterior to origin of adiposefin base, with ii,5 (3), ii,6 (2), iii5 (1), iii,6 ( 7 ) or iii,7 (3) rays. Anal fin margin almost straight. Coloration Body variably mottled with dark saddles. First saddle posterior of head, second saddle at dorsal fin, third saddle between dorsal and adipose fins, fourth saddle under anterior of adipose fin, and fifth saddle on caudal peduncle. All saddle connected laterally by broad stripe. Ventral region light brown with fourth saddle extending around ventral side. Dorsal, and anal fins brown wi th dark medial band (band on anal fin often indistinct on small specimens) Adipose fin dark brown, cream colored anteriorly, posteriorly and on distal edge. Pectoral and pelvic fins positioned horizontally with upper surfaces brown with cream colored dist al edge. L ower surfaces light yellow. Caudal fin cream colored with medial dark band. Distribution Ruzizi drainage, eastern Rwanda and Burundi, western Democratic Republic of the Congo. Also, northeastern tributaries of Lake Tanganyika in Burundi (Figure 5 2). Amphilius n. sp. Malagarasi (Figure 5 4; Table 5 4) Amphilius jacksonii David, 1937: 418, Malagarasi River; De vos et al., 2001: 131 Malagarasi River. Holotype: CU 97334, ex. CU 95207, Tanzania Malagarasi River at Lower Igamba Falls (Kasagwe) Mal agarasi River drainage, 5 10' 48S, 30 03' 03"E ( 1 : 52.3).

PAGE 162

162 Paratypes: Malagarasi River drainage: CU 90414, Tanzania, Downstream of bridge across Malagarasi River at Uvinza, 5 07' 03.7S, 30 22' 27.1"E ( 1 : 50.8); CU 95207, same data as holotype, ( 6 : 41. 3 69.8) ; CU 95208, Tanzania, Malagarasi River up river 12 km E of Uvinza near Village of Kanzibwe, 5 08' 23.5S, 30 29' 21.6"E ( 1: 55.3) ; CU 95209 Tanzania, Malagarasi River downriver 4 km W from Uvinza; village of Nkwasa 5 05' 52.6S, 30 21' 16.2"E ( 5: 38.1 74.7 ); CU 95210, Tanzania, Malagarasi River near fork in road to Ngutu, 5 06' 56.3S, 30 17' 48.1"E ( 1: 67.6), CU 95211 Tanzania, Malagarasi River in first site in gorge, 5 13' 44.2S, 30 13' 26.5"E ( 6: 41.252.9); CU 95212 Tanzania, Malagar asi River in gorge 5 12' 10.2S, 30 10' 11.1"E ( 1: 44.8); MRAC B3 06P 1 3, ex CU 95209 (3: 46.3 54.6); SAIAB 187275, ex CU 95209 (3: 45.652.6); UF 184234, ex CU 95209 (3: 41.864.7). Non types Malagarasi River drainage: CU 90426, Tanzania, Malagarasi River at bridge between Kafuru and Makere, 4 01' 32.5S, 30 33' 06.5"E (3:24.629.3), CU 90469, Tanzania, Malagarasi River at bridge between Kafuru and Makere, 4 01' 32.5S, 30 33' 06.5"E (26: 24.746.7); MRAC 47343362, Burundi, Malagarazi River and i ts tributaries, (19: 38.796.2), MRAC 91030.0217, Burundi, Nyesasa River, Trib of Muyovozi River (which is a tributary of the Malagarazi), at bridge on RN8 [now RN11] just before Gihofi, ca. 3 59'S, 30 09'E (1: 51.4); MRAC 91030.02180225, Burundi, Kinyw a River, Trib of Malagarazi, at the bridge on Route Pr85, near town of Buyaga, ca. 3 55'S, 30 10'E (7: 16.269.7); MRAC 91030.02260248, Burundi, Nyankanda River, Trib of Malagarazi, in the Fault of the Germans, on Route Pr 85, ca. 3 54'S, 30 13'E (23: 33.5 65.4); MRAC 91030.02490295, Burundi, Ntanga River, Trib of Malagarazi, at

PAGE 163

163 the bridge on Road of General interest 4, 7 km from Kinyinya, ca. 3 37' 39.3S, 30 17' 54.5E (47: 26.476.6); MRAC 91030.03000309, Burundi, Musasa River, Trib of Muyovozi, at the bridge on RN8, 5 km Rutana, ca. 3 58' 44.9S, 30 01' 56.0E (10: 45.671.5); MRAC 91030.03100315, Burundi, Mutsindozi River, trib of Malagarazi at bridge on Route RN12, km 39 from Lake Nyanza, ca. 4 05'S, 29 57'E (6: 54.767.5); MRAC 91061.0 106, Burundi, Musasa River, Trib of Muyovozi, at the bridge on RN8, 5 km Rutana, ca. 3 58' 44.9S, 30 01' 56.0E (1: 55.3); MRAC 91062.08960923, Burundi, Mutsindozi River, near the sugar refinery, Gihofi, ca. 4 02'S, 30 09'E (28: 29.673.1); MRAC 91062.09240929, Burundi, Musasa River, Trib of Muyovozi, at the bridge on RN8, 5 km Rutana, ca. 3 58' 44.9S, 30 01' 56.0E (6: 36.485.9); MRAC 91062.09300932, Burundi, Mashuro River, about 5 km from Gihofi towards Giharo, 3 57' 37S, 30 09' 54.6E (3: 28.9 48.7); MRAC 91062.0933, Burundi, Ruru River, trib of Rumpungwe River, on the road towards Cendajuru, ca. 3 19'S, 30 34'E (1: 79.2); MRAC 91062.09500957, Burundi, Kabingo, Mazimero River, on RutanaKinyinya Road, 3 53' 01.5S, 30 11' 54.2E ( 8: 37.163.1); MRAC 91062.09581056, Burundi, Rugoma River, trib of Rumpungwe River, 4 km from Kinyinya, ca. 3 38'S, 30 23'E (99: 24.6100.9); MRAC 91062.10571081, Burundi, Rugaragara River, on KinyinyaGisuru Road, about 20 km from Gisuru, ca. 3 27' S, 30 24'E (25: 25.967.7); MRAC 91062.10821094, Burundi, Kiruhura River, trib of Mwambu River, zone Muyange, ENE of Cendajuru, ca. 3 16'S, 30 40'E (13: 32.6 71.8); MRAC 91062.10951107, Burundi, Nyabigosi River, trib of Rumpungwe River, near Gisuru, c a. 3 29'S, 30 30'E (13: 28.368.5); MRAC 91062.11391168, Burundi, Gitinwa River, trib of Ruru River, trib of Rumpungwe River, near Cendajuru, before village of Gusiana, ca. 3 20'S, 30 34'E (30:

PAGE 164

164 24.683.9); MRAC 91062.11691245, Burundi, Mukitcha River 8 km towards Cendajuru from Gasenyi, ca. 3 17'S, 30 36'E (77: 28.774.6); MRAC 91079.00550058, Burundi, Musasa river, (local name Uruhuzi) 78 km SSE of Gitega, 10 km SE of Rutana, 4 00' 00S, 30 06' 47.3E (5: 25.971.9); MRAC 93150.01840186, B urundi, Mazimero River, on road from RutanaKinyinya, ca. 3 54'S, 30 13'E (3: 63.575.9); MRAC 93150.01870238, Burundi, Ruru River, about 9 km from Muyaga towards Cendajuru, ca. 3 18'S, 33 13'E (52: 35.480.4); MRAC 93150.0239, Burundi, Ntanga River, Trib of Malagarazi, at the bridge on Road of General interest 4, 7 km from Kinyinya, ca. 3 38'S, 30 18'E (1: 52.7); MRAC 93150.02400260, Burundi, Mukazye River (Nyamabuye), 10 km from Giharo towards Kinyinya, ca. 3 47'S, 30 18'E (21: 26.972.9); MRAC 93150.02610333, Burundi, Idumaniro River,on CendajuruGitwenge road, ca. 3 15'S, 30 39'E (73: 25.881.1); MRAC 93150.03510359, Burundi, Rumpungwe River, near Gisuru,on road from KinyinyaGisuru, ca. 3 27'S, 30 29'E (9: 26.274.3); MRAC 93150.03600378, Burundi, Kiruhura River,on CendajuruGitwenge road, ca. 3 15'S, 30 39'E (18: 35.865.8), MRAC 93150.03790382, Burundi, Nyanzari River, on CankuzoMushiha road, near Gishungo, 3 04' 25S, 30 39' 22E (4: 48.368.6); MRAC 93150.03830392, Burundi, Mutsindozi River, near the sugar refinery, Gihofi, ca. 4 00'S, 30 40'E (12: 44.280.0); MRAC 93152.06130622, Tanzania, Malagarazi River, rapids at Uvinza, near salt mine, 5 06' 07S, 30 21' 56E (11: 37.159.9); MRAC 93152.06230643, Tanzania, Nyamgongo River, trib of Malagarazi River, km 46 from Uvinza to Kasulu, 4 48' 34S, 30 12' 51E (21: 28.369.0); MRAC 96031.15481549, Burundi, Nyarugunga River, km 30 after Kinyinya, on KinyinyaGisuru road, ca. 4 52'S, 29 50'E (2: 53.758.6); MRAC 96031.1550 1562, Burundi,

PAGE 165

165 Nyarugunga River, km 30 after Kinyinya, on KinyinyaGisuru road, ca. 4 52'S, 29 50'E (13: 47.686.6); MRAC 96031.15631570, Burundi, Nyarubare River, km 15 on Kinyinya Gisuru road, ca. 3 34'S, 30 26'E (8: 45.781.8); MRAC 96083.1141, T anzania, Ruchigi River, 7km after Kasulu to Kibando, trib of Malagarasi, 4 32' 01.4S, 30 08' 58.6E (1: 71.5); MRAC 96083.11421145, Tanzania, Mgandazi River, near Kasulu, ca. 4 37'S, 30 06'E (4: 48.772.6). Luiche River drainage: MRAC 93152.06440649, Tanzania, Mukuti River, trib. of Luiche River, on road from Kigoma to Kasulu, 4 53' 12.1S, 29 52' 12E (6: 35.0118.8); MRAC 93152.06500651, Tanzania, Mungonya River, trib of Luiche River, 10 km from Kigoma, route KigomaKasulu, 4 52' 28S, 29 49' 52E (2: 65.290.9); MRAC 93152.06520661, Tanzania, Kidahwe River, km 34 on Kigoma Uvinza road, trib of Luiche River, ca. 4 53'S, 29 48'E (12: 68.6112.0). Lake Rukwa Basin: MRAC 191055057, Zambia, Saisi and Kalambo river, Tunduma road, ca. 9 06'S, 31 29'E (3: 30.639.1); MRAC 191092113, Zambia, Saisi River, at crossing of AbercornTunduma road, 9 05' 38.1S, 31 29' 20.5E (22: 27.5104.8); MRAC 191423424, Zambia, Lumi river, Kawimbe road, ca. 8 50'S, 31 32'E (2: 31.635.4); MRAC 94034.08220831, Tanzania, Piti River, 63 km south of Rungwa River on road to Makongolosi, 7 26' 48.5S, 33 25' 22.1E (10: 32.966.6); SAIAB 37397, Zambia, Near Mbala, Saisi River at Tunduma Road near Neilsons Farm, ca. 9 05'S, 31 32'E (1: 55.4); SAIAB 38096, Zambia, Saisi River at Tunduma Road bridge, 9 05' 39S, 31 29' 22E (1: 35.8); SAIAB 39557, Zambia, Saisi River at Tunduma Road bridge, ca. 9 05' 39S, 31 29' 22E (1: 57.0); SAIAB 38108, Zambia, Chitungulu stream, Nsunzu Farm, trib of Saisi River, ca. 9 05'S, 31 32'E (2: 54.370.3); SAIAB 39543, Zambia, Chitungulu stream, Nsunzu Farm, trib of Saisi River, ca. 9 05'S, 31 32'E (1: 37.6);

PAGE 166

166 SAIAB 50333, Tanzania, Lupa River at Lupatingatinga, on Makongolsi Rungwa Road, 8 01' 48.9S, 33 16' 18.9E (1: 42.0). Rufiji basin, Upper Great Ruaha River system: SAIAB 59388, Tanzania, Bridge near Chinata on MbeyaIringa road, 8 51' 35.7S, 34 01' 33E (31: 49.2111.7); SAIAB 59397, Tanzania, Great Ruaha River at Rte A104 crossing just east of Chimala, 8 51' 16.4 S, 34 05' 7.0E (32: 30.3 72.6). Diagnosis Diagnostic char acters are summarized in Table 52 Amphilius n. sp. Malagarasi is diagnosed from A. jacksonii, A. n. sp. Ruzizi and A. n. sp. Lake Kyogo by its deeper caudal peduncle (caudal peduncle depth 9.912.3% SL vs. 4.8 9.5% SL), and from A. n. sp. Ruzizi and A. n. sp. Congo by its wider interorbital width (28.135.8% HL vs. 23.427.7% HL). It is distinguished from A. n. sp. Rufiji by its longer caudal peduncle (caudal peduncle length 16.018.8% SL vs. 13.315.5% SL) and shorter dorsal fin insertion to adiposefin insertion length ( 37.541.7% SL vs. 42. 2 44.6% SL). It is further diagnosed from A. n. sp. Congo by having fewer total gill rakers on the first gill arch (79, rarely 6 or 10 vs. 10 11, rarely 9 or 12). Description Morphometric data are in Table 4 Body elongate, ventral profile flattened ventrally to anal fin base, then tapered dorsally to end of caudal peduncle. Dorsal profile rising gently from tip of snout to origin of dorsal fin, then nearly horizontal to end of caudal peduncle. Greatest body depth at dorsal fin origin. Caudal pedunc le laterally compressed, with crenellated epidermal fold. Anus and urogenital openings located at midpoint of adpressed pelvic fin, much closer to insertion of pelv ic fin than to origin of anal fin. Skin smooth. Lateral line complete, extending from dorsal edge of opercular cavity to base of caudal fin.

PAGE 167

167 Head and anterior part of body depressed and broad. Head wedgeshaped in lateral view. Snout broad, blunt when viewed from above. Head becoming wider from tip of snout to pectoral fin base. Branchiostegal membranes moderately joined at isthmus forming a V shaped connection. Mouth broad, gently curved, subterminal. Lips moderately fleshy, slightly papillate. Rictal lob e large and slightly papillate. Anterior portion of premaxillary tooth band exposed with mouth closed. Premaxillary tooth patches joined, forming cresent shaped band. Premaxillary and dentary t eeth short conical. Dentary tooth patches forming U shaped ban d, separated medially. Three pairs of simple, tapered circumoral barbels. Maxillary barbel large, fleshy and flattened with pointed tip; barbel extending posterolaterally from cor ner of mouth, to pectoral fin base. Outer mandibular barbel thin with pointed tip, origin at posterior corner of lower jaw, extending to origin of pectoral fin. Inner mandibular barbel originates anterolaterally of inner mandibular barbel, extending to edge of branchiostegal membrane. Branchiostegal membrane with 7 (17), or 8 ( 24) rays. Gill rakers on first epibranchial 2 ( 20) or 3 (21); rakers on first ceratobranchial 4 ( 1 ), 5 ( 8 ), 6 (25), or 7 ( 7 ) ; total gill rakers on first arch 6 ( 1 ), 7 (4 ), 8 (16), 9 ( 16) or 10 (4). Eye small, positioned dorsolaterally approximately midway betw een tip of snout and posterior margin of operculum. Horizontal diameter of eye slightly wider than vertical diameter. Eye without free orbit; covered with skin confluent with dorsal surface of head. Anterior and posterior nares with prominent tubular rims; nares separate but relatively close to each other. Posterior nare located about midway between eye and tip of snout.

PAGE 168

168 Dorsal fin origin at point over or just posterior to tip of pectoral fin. Dorsal fin with i,6 ( 41) rays, and fin margin straight Pectoral fin with i, 8 ( 35), or i, 9 ( 6 ) rays ; unbranched ray greatly thickened. Pectoral fin with four or five innermost rays progressively shorter making posterior fin margin rounded. Origin of p elvic fin posterior of dorsal fin insertion Pelvic fin with i,5 ( 41) rays with first ray unbranched and greatly thickened. Pelvic fin with straight posterior margin. Adiposefin base longer than anal fin base, origin anterior to origin of anal fin base, fin extending past anal fin insertion. Margin strongly convex with sharply rounded edge, deeply incised posteriorly. Caudal fin deeply forked with tips of lobes rounded; fin with i, 5 6 ,i ( 40) principal rays. Anal fin with short base, origin posterior to origin of adiposefin base, with ii,6 (7), ii,7 (2), iii5 (4), iii,6 ( 27) or iii,7 (1) rays. Ana l fin margin almost straight. Coloration Body variably mottled with dark saddles. First saddle posterior of head, second saddle at dorsal fin, third saddle between dorsal and adipose fins, fourth saddle under anterior of adipose fi n, and fifth saddle on caudal peduncle. All saddle connected laterally by broad stripe. Ventral region light brown with fourth and fifth saddle extending around ventral side. Dorsal, and anal fins light brown with dark medial band. Adipose fin dark brown t o black, with posterior distal edge cream colored. Pectoral and pelvic fins positioned horizontally with upper surfaces cream colored with dark fin base and dark medial band. Lower surfaces light yellow. Caudal fin cream colored with dark markings on upper and lower lobes. Caudal coloration asymmetrical, with lower lobe almost completely dark (except for small cream colored mark at base of caudal and cream

PAGE 169

169 colored tip). Upper lobe with less dark pigment, with dark blotch covering dorsal edge, but caudal bas e, medial ventral edge, and tip cream colored. Distribut ion Malagarasi River drainage, western Tanzania and southern Burundi. Also, Luiche River drainage western Tanzania, Lake Rukwa basin, Tanzania and Zambia, and upper Great Ruaha River system, Rufiji ba sin, eastern Tanzania (Figure 52). Amphilius n. sp. Congo (Fig. 5 5; Table 55) Amphilius platychir (non Gnther) Boulenger, 1920: 26 Lubumbashi lisabethville (MRAC P 6787) [Kafubu River system, Upper Congo basin] Holotype: CU 97335, ex. CU 91052, Za mbia Tswishi Stream at bridge under construction Lwela River system 11 33' 26S, 29 09' 59"E ( 1 : 83.1). Paratypes: Lwela River system : AUM 57570, ex. CU 91049 (3: 36.351.7); CU 91049, Zambia, Lwela River at bridge on Mansa? Road 11 33' 25S, 29 1 0' 10"E ( 21: 32104.5); CU 91050 Zambia Ngo Stream, about 40 km SW of Mansa, 11 31' 33S, 29 09' 03"E ( 4: 34.246.8 ); CU 91052 same data as holotype, (11: 31.744.9); SAIAB 76634 same data as CU 91049 (12: 32.786.7 ); SAIAB 76789 same data as CU 910 50 (5: 36.047.5); MRAC B3 06P 4 6 ex. CU 91049 (3: 41.947.4); SAIAB 76799, same data as holotype (13: 27.3 57.2 ); UF 184235 (3: 39.050.8). Non types Chambeshi River system: CU 91056, Zambia, Kanchibiya Stream at bridge on Kasama Mpika road, 11 29 44.2S, 31 16' 46.6"E (28: 32.557.0); MRAC 96031.15401547, Zambia, Musombizi River, trib of Chambeshi, 55km on the road MbalaKasama, ca. 09 18S, 31 16.5'E (8: 25.857.8); MRAC 96031.16201625, Zambia, Chambeshi

PAGE 170

170 River, at bridge km 40 on road MbalaK asama, 09 12 15.3S, 31 20' 49.8"E (6: 57.7135.6); SAIAB 40125, Zambia, Chambeshi River, rapids near Kapoloso Stream, ca. 10 43S, 31 15'E (1: 47.8); SAIAB 46788, Zambia, Mansha River, Shiwa Hot Springs & road bridge, ca. 11 10S, 31 35'E ( 16: 32.0 64.8); SAIAB 77139 Zambia, Samfa Rapids at pontoon on Chambeshsi River 10 51 07.6S, 31 10' 02.3"E (1: 69.5); SAIAB 77142 same data as CU 91056 (28: 29.485.3 ). Kafubu River system : MRAC 6787, Democratic Republic of the Congo, Lubumbashi River, at E lisabethville [Lubumbashi], ca. 11 39S, 27 28'E ( 1: 39.1); MRAC 183365366, Democratic Republic of the Congo, at the bridge on Lubumbashi River, forested camp of the Kipopo, ca. 11 33S, 27 22'E ( 2: 31.733.1); MRAC 183367369, Democratic Republic of the Congo Lubumbashi River, 25 km upstream of Elisabethville, ca. 11 33S, 27 22'E ( 3: 33.9 39.2 ); MRAC 73025.0803, Democratic Republic of the Congo, Lubumbashi, in front of Lido, Lubumbashi River ca. 11 39S, 27 27'E ( 1: 118.8). Luapula River syst em : SAIAB 76705 Zambia Mambilima Falls on Luapula River 10 32 22.2S, 28 39' 40.3"E ( 1: 57.0). Lufira River system : MRAC 165341, Democratic Republic of the Congo, Gombela, Kafila River, Katanga, 10 46 31S, 27 47' 38"E ( 1: 60.8); SAIAB 77478 Demo cratic Republic of Congo, Diptera River near Fungurume, 10 36 20.9S, 26 16' 40.1"E ( 4: 43.5 77.0). Luongo River system : CU 91051 Zambia, Luongo River, below Musonda Dam, on road from Mansa to Kashiba (route D79) 10 42 15.5S, 28 48' 03.6"E ( 2: 60. 0 62.0); CU 91053, Zambia Luongo River above pontoon south of Musonda Falls, MansaSerenje Road (route D235) 10 40 51.6S, 28 43' 09.1"E ( 2: 57.078.7 ); CU 91054 Zambia Luongo River at bridge on KashibaMwenda road (route M3) 10 28 12.7S, 31 01' 28.2"E ( 3: 44.0105.1);

PAGE 171

171 SAIAB 76663 same data as CU 91053 (3: 54.6 94.3 ); SAIAB 76670 same data as CU 91051 (1: 83.2 ); SAIAB 76729 same data as CU 91054 (5: 39.9 65.4 ); SAIAB 76925 Zambia, Luongo River at Mukonshi Bridge on MwendaKawambwa road (rout e M13) 10 08 39S, 29 10' 01.2"E ( 41: 32.2 53.8 ). Upper Congo (Lualaba) River system: SAIAB 81501 Democratic Republic of the Congo, Bona River [Tributary of Lac Delcommune (= Lac Nzilo)], Near Lenge Village, 10 36 34S, 25 49' 15"E ( 1: 78.8); SAIAB 81566 Democratic Republic of the Congo, Large Stream 1, Bona River system 10 35 49S, 25 53' 23"E ( 1: 43.0 ); SAIAB 82855 Democratic Republic of the Congo, Kisanfu River, Upstream of Nayebe crossing 10 48 01.1S, 25 58' 52.6"E ( 2: 46.8 85.4); SAI AB 82877 Democratic Republic of the Congo, Kisanfu River, Just below bridge by hydroelectric station, 10 45 49.3S, 25 57' 49.3"E ( 6: 36.542.1); SAIAB 82898, Democratic Republic of the Congo, Kisanfu River, Channel downstream of hydro station 10 45 49.3S, 25 57' 49.3"E ( 6: 37.5 74.3). Diagnosis Diagnostic characters are summarized in Table 52. Amphilius n. sp. Congo is diagnosed from all other species of the Amphilius jacksonii complex by having more gill rakers on the first gill arch (1011, rar ely 9 or 12 vs. 69, rarely 5 or 10). It is further diagnosed from A. jacksonii by its deeper caudal peduncle (caudal peduncle depth 9.711.4% SL vs. 4.87.9% SL), shorter caudal peduncle (caudal peduncle 14.416.4% SL vs. 16.720.6% SL), and less slender body (body depth at anus 13.914.9% SL vs. 9.6 13.2% SL). It is further diagnosed from A. n. sp. Ruzizi by having more branchiostegal rays (8 9 vs. 6 7), a deeper caudal peduncle (caudal peduncle depth 9.711.4% SL vs. 8.6 9.5 % SL), shorter caudal peduncle (caudal peduncle 14.4 16.4% SL vs. 18.7 20.3 % SL), and less slender body (body depth at anus 13.914.9% SL vs. 11.1 12.9 % SL).

PAGE 172

172 Amphilius n. sp. Congo is further diagnosed from A. n. sp. Malagarasi A. n. sp. Rufiji, and A. n. sp. Lake Kyogo by its narrower interorbital width ( 25.127.7 % HL vs. 28.135.8% HL), and from A. n. sp. Rufiji by having more branchiostegal rays (89 vs. 6 7, rarely 8), more branched pectoral fin rays (910 vs. 78, rarely 9), and a shorter dorsal fin insertion to adipose fin insert ion length ( 37.5 41.2% SL vs. 42.244.6 % SL). It is further diagnosed from A. n. sp. Lake Kyogo by its deeper caudal peduncle (caudal peduncle depth 9.711.4% SL vs. 8.1 9.5 % SL), shorter caudal peduncle (caudal peduncle 14.416.4% SL vs. 16.820.9% SL). D escription Morphometric data are in Table 5 5 Body elongate, ventral profile flattened ventrally to anal fin base, then tapered dorsally to end of caudal peduncle. Dorsal profile rising gently from tip of snout to origin of dorsal fin, then nearly horizontal to end of caudal peduncle. Greatest body depth at dorsal fin origin. Caudal pedunc le laterally compressed, with crenellated epidermal fold. Anus and urogenital openings located at midpoint of adpressed pelvic fin, closer to insertion of pelvic fin than to origin of anal fin. Skin smooth. Lateral line complete, extending from dorsal edge of opercular cavity to base of caudal fin. Head and anterior part of body depressed and broad. Head wedgeshaped in lateral view. Snout broad, blunt when viewed from above. Head becoming wider from tip of snout to pectoral fin base. Branchiostegal membranes moderately joined at isthmus forming a V shaped connection. Mouth broad, gently curved, subterminal. Lips moderately fleshy, slightly papillate. Rictal lobe large and slightly papillate. Anterior portion of premaxillary tooth band exposed with mouth closed. Premaxillary tooth patches joined, forming crescent

PAGE 173

173 shaped band. Premaxillary and dentary t eeth short conical. Dentary tooth patches forming U shaped band, separat ed medially. Three pairs of simple, tapered circumoral barbels. Maxillary barbel large, fleshy and flattened with pointed tip; barbel extending posterolaterally from cor ner of mouth, to pectoral fin base. Outer mandibular barbel thin with pointed tip, orig in at posterior corner of lower jaw, extending to origin of pectoral fin. Inner mandibular barbel originates anterolaterally of inner mandibular barbel, extending to edge of branchiostegal membrane. Branchiostegal membrane with 8 (25), or 9 ( 12) rays. Gill rakers on first epibranchial 2 ( 1 ) or 3 (37); rakers on first ceratobranchial 7 ( 22), or 8 ( 15) ; total gill rakers on first arch 10 ( 22) or 11 (15). Eye small, positioned dorsolaterally approximately midway between tip of snout and posterior margin of operculum. Horizontal diameter of eye slightly wider than vertical diameter. Eye without free orbit; covered with skin confluent with dorsal surface of head. Anterior and posterior nares with prominent tubular rims; nares separate but relatively close to each other. Posterior nare located about midway between eye and tip of snout. Dorsal fin origin at point over tip of pectoral fin. Dorsal fin with i,6 ( 37 ) rays, and fin margin straight Pectoral fin with i, 9 ( 37 ) rays ; unbranched ray greatly thickened. Pector al fin with four or five innermost rays progressively shorter making posterior fin margin rounded. Origin of pelvic fin posterior of dorsal fin insertion Pelvic fin with i,5 ( 37) rays with first ray unbranched and greatly thickened. Pelvic fin with straig ht posterior margin.

PAGE 174

174 Adiposefin base longer than anal fin base, origin anterior to origin of anal fin base, fin extending past anal fin insertion. Margin strongly convex with sharply rounded edge, deeply incised posteriorly. Caudal fin deeply forked with tips of lobes rounded; fin with i, 5 6 ,i ( 36) principal rays. Anal fin with short base, origin posterior to origin of adiposefin base, with ii,6 (2), ii,7 (2), iii,6 ( 24) or iii,7 (10) rays. Anal fin margin almost straight. Coloration Body variably mottled with dark saddles. First saddle posterior of head, second saddle at dorsal fin, third saddle between dorsal and adipose fins, fourth saddle under anterior of adipose fin, and fifth saddle on caudal peduncle. All saddle connected laterally by broad stripe. Ventral region light brown with fourth and fifth saddle extending around ventral side. Dorsal, and anal fins light brown with dark medial band. Adipose fin dark brown to black, with anterior and posterior distal edges cream colored. Pectoral and pelvic f ins positioned horizontally with upper surfaces cream colored with dark fin base and dark medial band. L ower surfaces light yellow. Caudal fin cream colored with dark markings on upper and lower lobes. Caudal coloration asymmetrical, with lower lobe almost completely dark (except for small cream colored mark at base of caudal and cream colored tip). Upper lobe with less dark pigment, with dark blotch covering dorsal edge, but caudal base with large cream colored patch and tip cream colored. Distribution Upp er Congo basin upstream of Kabalp, Democratic Republic of the Congo including tributaries of the Upper Congo mainstem, the Lufira River drainage, and the Kafubu, Chambeshi, Lwela, and Luongo River systems of the Luapula River drainage (Figure 5 2).

PAGE 175

175 Amphili us n. sp. Rufiji (Figure 5 6; Table 56) Amphilius platychir (non Gnther) Bailey, 1969: 192 (in part), Kilombero ( BMNH 1969.2.11.192194) [Rufiji basin] Holotype: UF 184237, ex. UF 170743, Tanzania, Sonjo River at bridge in Man'gula on road from Mikumi to Ifakara, altitude 302 m Rufiji River basin, 07 48' 29.6S, 36 53' 47.6"E ( 1 : 86.3). Paratypes: Rufiji River basin : AMNH 258334, same data as holotype, ex. CU 93734 (10: 43.461.5); AUM 57571, same data as holotype, ex. CU 93734 (10: 40.766.8); BMNH 1969.2.11.192194, Tanzania, Near Kiberege, trib of Kilombero River ca. 07 57'S, 36 52'E ( 3: 30.831.7); CU 93731 Idete River at bridge in Idete on road from Ifakara to Taveta, altitude 310 m 08 06' 14S, 36 29' 17"E ( 7: 27.3 67.3); CU 93732, Tanzania 12 Bridges River at overhead bridge for train on road from Mikumi to Ifakara, altitude 419 m 07 27' 53S, 37 00' 52"E ( 4: 49.7 72.4 ); CU 93734, same data as holotype, (40: 36.769.9); MRAC B3 06P 7 16 same data as holotype, ex. UF 170743 (10: 43.360.5); SAIAB 187279, same data as holotype, ex. UF 170743 (10: 45.157.0); UF 170706, same data as CU 93732 (5: 47.680.3); UF 170729, same data as CU 93731 (8: 28.9 84.7); UF 170743, same data as holotype (40: 39.884.7). Non types Wami basin: CU 93730 Tanzania Divue River above & below falls along road from Dumila to Turiani, altitude 374 m 06 10' 26.8S, 37 34' 59.8"E ( 10: 27.1 57.5); CU 93733, Tanzania, Wami River at bridge in Rudewa on road from Dumila to Kilosa, altitude 433 m 06 40' 45.2S, 3 7 07' 27.0"E ( 8: 27.141.2 ); UF 170707, same data as CU 93733 (10: 25.9 77.0 ); UF 170708, same data as CU 93730 (10: 40.059.4).

PAGE 176

176 Diagnosis Diagnostic characters are summarized in Table 5 2 Amphilius n. sp. Rufiji is diagnosed from A jacksonii, A. n. sp Ruzizi A. n. sp. Malagarasi and A. n. sp. Lake Kyogo by its shorter caudal peduncle (caudal peduncle length 13.315.5% SL vs. 16.020.9% SL), and from A. jacksonii, A. n. sp. Ruzizi and A. n. sp. Lake Kyogo by its deeper caudal peduncle (caudal pedunc le depth 10.112.0% SL vs.4.8 9.5% SL). It is diagnosed from A. n. sp. Ruzizi A. n. sp. Malagarasi and A. n. sp. Congo by its longer d orsal fin insertion to adiposefin insertion length (42.244.6 % SL vs. 37.541.7% SL). Amphilius n. sp. Rufiji is furth er diagnosed from A. jacksonii by its deeper body (body depth at anus 14.5 17.4% SL vs. 9.6 13.2 % SL), and from A. n. sp. Ruzizi by wider interorbital width ( 28.9 34.1% HL vs. 23.425.1% HL) and deeper body (body depth at anus 14.517.4% SL vs. 11.112.9% SL). It is further diagnosed from A. n. sp. Lake Kyogo by having fewer branched pectoral fin rays ( 7 8, rarely 9 vs. 9), and from A. n. sp. Congo by its fewer branchiostegal rays ( 6 7, rarely 8 vs. 8 9), fewer branched pectoral fin rays ( 7 8, rarely 9 vs. 9 10), fewer gill rakers on the first gill arch ( 6 8, rarely 9 vs. 10 11, rarely 9 or 12), and wider interorbital width ( 28.934.1% HL vs. 25.127.7 % HL). Description Morphometric data are in Table 5 6. Body elongate, ventral profile flattened ventrally t o anal fin base, then tapered dorsally to end of caudal peduncle. Dorsal profile rising steeply from tip of snout to origin of dorsal fin, then nearly horizontal to end of caudal peduncle. Greatest body depth at dorsal fin origin. Caudal peduncle laterally compressed, with crenellated epidermal fold. Anus and urogenital openings located at midpoint of adpressed pelvic fin, closer to insertion of pelvic fin than to origin of anal fin.

PAGE 177

177 Skin smooth. Lateral line complete, extending from dorsal edge of opercular cavity to base of caudal fin. Head and anterior part of body depressed and broad. Head wedgeshaped in lateral view. Snout broad, blunt when viewed from above. Head becoming wider from tip of snout to pectoral fin base. Branchiostegal membranes moderately joined at isthmus forming a V shaped connection. Mouth broad, gently curved, subterminal. Lips moderately fleshy, slightly papillate. Rictal lobe large and slightly papillate. Anterior portion of premaxillary tooth band exposed with mouth closed. Premax illary tooth patches joined, forming crescent shaped band. Premaxillary and dentary teeth short, conical. Dentary tooth patches forming U shaped band, separated medially. Three pairs of simple, tapered circumoral barbels. Maxillary barbel large, fleshy and flattened with pointed tip; barbel extending posterolaterally from corner of mouth, to pectoral fin base. Outer mandibular barbel thin with pointed tip, origin at posterior corner of lower jaw, extending to origin of pectoral fin. Inner mandibular barbel originates anterolaterally of inner mandibular barbel, extending to edge of branchiostegal membrane. Branchiostegal membrane with 6 (22), 7 (123), or 8 (2) rays. Gill rakers on first epibranchial 1 (1), 2 (122) or 3 (24); rakers on first ceratobranchial 4 (19), 5 (95), or 6 (33); total gill rakers on first arch 6 (19), 7 (79) 8 (42) or 9 (7). Eye small, positioned dorsolaterally approximately midway between tip of snout and posterior margin of operculum. Horizontal diameter of eye slightly wider than vertic al diameter. Eye without free orbit; covered with skin confluent with dorsal surface of head. Anterior and posterior nares with prominent tubular rims; nares separate but

PAGE 178

178 relatively close to each other. Posterior nare located about midway between eye and t ip of snout. Dorsal fin origin at point over tip of pectoral fin. Dorsal fin with i,6 (147) rays, and fin margin straight. Pectoral fin with i,7 (13), i,8 (130), or i,9 (4) rays ; unbranched ray greatly thickened. Pectoral fin with four or five innermost rays progressively shorter making posterior fin margin rounded. Origin of pelvic fin posterior of dorsal fin insertion. Pelvic fin with i,5 (147) rays with first ray unbranched and greatly thickened. Pelvic fin with straight posterior margin. Adiposefin bas e longer than anal fin base, origin anterior to origin of anal fin base, fin extending past anal fin insertion. Margin strongly convex with sharply rounded edge, deeply incised posteriorly. Caudal fin deeply forked with tips of lobes rounded; fin with i,5, 5,i (4), i,5,6,i (141) or i,6,6,i (2) principal rays. Anal fin with short base, origin posterior to origin of adiposefin base, with ii,6 (56), ii,7 (34), iii,5 (3), iii,6 (50), or iii,7 (5) rays. Anal fin margin almost straight. Coloration Body variably m ottled with dark saddles. First saddle posterior of head, second saddle at dorsal fin, third saddle between dorsal and adipose fins, fourth saddle under anterior of adipose fin, and fifth saddle on caudal peduncle. All saddles connected laterally by broad stripe. Ventral region light brown with fourth and fifth saddle extending around ventral side. Dorsal, and anal fins light brown with dark medial band. Adipose fin dark brown to black, with anterior and posterior distal edges cream colored. Pectoral and pelvic fins positioned horizontally with upper surfaces cream colored with dark fin base and dark medial band. Lower surfaces light yellow. Caudal fin cream colored with

PAGE 179

179 dark markings on upper and lower lobes. Caudal coloration variable but usually spotted a nd asymmetrical, with lower lobe more pigmented. Distribution Rufiji and Wami River basins, eastern Tanzania (Figure 52). Amphilius n. sp. Lake Kyogo (Figure 5 7; Tables 57) Amphilius jacksonii Seegers et al., 2003: 37, Lake Victoria drainage (affluent rivers) [Nile basin] Holotype: UF 184238, ex. UF 169233, Uganda Manafwa River at Bumwangu, Lake Kyogo drainage, Nile basin, 00 56' 26.2N, 34 16' 49.2"E ( 1 : 104.1). Paratypes: Lake Kyogo drainage: AUM 47147, same data as holotype (16: 37.7 112.3); AUM 47149 Uganda, Malaba River at Lwakhakha, border between Uganda and Kenya, 00 47' 04N, 34 22' 44"E ( 7: 46.398.1); BMNH 1961.6.13.19, Uganda Malawa [=Malaba] River ca. 00 35'N, 34 03'E (1: 99.4); BMNH 1962.2.6.6263 Uganda, Malawa [=Malaba] River ca. 00 35'N, 34 03'E (2: 80.0 107.9 ); BMNH 1965.10.15.1821, Kenya Malikisi River ca. 01 37'N, 34 13'E ( 4: 35.7 48.0); CU 97333, same data as holotype, ex. AUM 47147 (3: 50.390.5); MRAC B3 06P 17 19, same data as holotype, ex. UF 169233 (3: 56.088.1); SAIAB 187280, same data as holotype, ex. UF 169233 (3: 47.998.0); UF 1 69233, same data as holotype (15: 43.0 107.6); UF 169235, same data as AUM 47149 (7: 45.4 90.1). Non types Lake Kyogo drainage: BMNH 1965.10.15.11 13, Uganda Sironko River ca. 01 29'N, 34 14'E ( 2: 103.7119.5); BMNH 1965.10.15.1415 Uganda Sironko River ca. 01 29'N, 34 14'E ( 1: 101.1); BMNH 1965.10.15.1617 Uganda, Zuzu River, tributary of the Manafwa River ca. 01 13'N, 34 21'E ( 2: 40.942.1). Lake Manyara

PAGE 180

180 Basin: BMNH 1 969.2.20.1 Tanzania Stream entering Lake Manyara, ca. 03 38'S, 35 41'E ( 1: 45.0). Nzoia River system (NE Lake Victoria tributary) : SAIAB 65048 Kenya, Moi Brgidge, Little Nzoia River ca. 00 55'N, 35 07'E ( 1: 74.5 ). Diagnosis Diagnostic characters ar e summarized in Table 52. Amphilius n. sp. Lake Kyogo is diagnosed from A. n. sp. Malagarasi A. n. sp. Congo, and A. n. sp. Rufiji by having a more slender caudal peduncle (depth 8.19.5 % SL vs. 9.912.3% SL). It is diagnosed from A. jacksonii by having a deeper caudal peduncle (depth 8.1 9.5 % SL vs. 4.8 7.9 % SL) and a deeper body (body depth at anus 13.5 15.8 % SL vs. 9.6 13.2 % SL). Amphilius n. sp. Lake Kyogo is diagnosed from A. n. sp. Ruzizi by having a wider interorbital width ( 28.3 31.7% HL vs. 23.425.1% HL) and a deeper body (body depth at anus 13.515.8% SL vs. 11.112.9% SL). It is further diagnosed from A. n. sp. Congo by having fewer gill rakers on the first gill arch (68, rarely 5 or 9 vs. 10 11, rarely 9 or 12) and a longer caudal peduncle (l ength 16.8 20.9 % SL vs. 14.416.4 % SL). Amphilius n. sp. Lake Kyogo is further diagnosed from A. n. sp. Rufiji by having more branched pectoral fin rays (9 vs. 7 8, rarely 9) and a longer caudal peduncle (length 16.8 20.9 % SL vs. 13.315.5% SL). Descript ion Morphometric data are in Table 5 7 Body elongate, ventral profile flattened ventrally to anal fin base, then tapered dorsally to end of caudal peduncle. Dorsal profile rising gently from tip of snout to origin of dorsal fin, then nearly horizontal to end of caudal peduncle. Greatest body depth at dorsal fin origin. Caudal pedunc le laterally compressed, with crenellated epidermal fold. Anus and urogenital openings located at

PAGE 181

181 midpoint of adpressed pelvic fin, closer to insertion of pelvic fin than to ori gin of anal fin. Skin smooth. Lateral line complete, extending from dorsal edge of opercular cavity to base of caudal fin. Head and anterior part of body depressed and broad. Head wedgeshaped in lateral view. Snout broad, blunt when viewed from above. He ad becoming wider from tip of snout to pectoral fin base. Branchiostegal membranes moderately joined at isthmus forming a V shaped connection. Mouth broad, gently curved, subterminal. Lips moderately fleshy, strongly papillate. Rictal lobe large and papill ate. Anterior portion of premaxillary tooth band exposed with mouth closed. Premaxillary tooth patches joined, forming crescent shaped band. Premaxillary and dentary t eeth short conical. Dentary tooth patches forming U shaped band, separated medially. Thr ee pairs of simple, tapered circumoral barbels. Maxillary barbel large, fleshy and flattened with pointed tip; barbel extending posterolaterally from cor ner of mouth, to pectoral fin base. Outer mandibular barbel thin with pointed tip, origin at posterior corner of lower jaw, extending to origin of pectoral fin. Inner mandibular barbel originates anterolaterally of inner mandibular barbel, extending to edge of branchiostegal membrane. Branchiostegal membrane with 7 (21), or 8 ( 25) rays. Gill rakers on first epibranchial 2 ( 36) or 3 (10); rakers on first ceratobranchial 4 ( 2 ), 5 ( 22), 6 ( 1 ) or 7 (1); total gill rakers on first arch 6 ( 2 ), 7 ( 16) 8 (23) or 9 (5). Eye small, positioned dorsolaterally approximately midway between tip of snout and posterior marg in of operculum. Horizontal diameter of eye slightly wider than vertical diameter. Eye without free orbit; covered with skin confluent with dorsal surface

PAGE 182

182 of head. Anterior and posterior nares with prominent tubular rims; nares separate but relatively clos e to each other. Posterior nare located about midway between eye and tip of snout. Dorsal fin origin at point over tip of pectoral fin. Dorsal fin with i,6 ( 45 ) or i,7 (1) rays, and fin margin straight Pectoral fin with i, 9 ( 46) rays ; unbranched ray great ly thickened. Pectoral fin with four or five innermost rays progressively shorter making posterior fin margin rounded. Origin of p elvic fin posterior of dorsal fin insertion Pelvic fin with i,5 ( 46) rays with first ray unbranched and greatly thickened. Pe lvic fin with straight posterior margin. Adiposefin base longer than anal fin base, origin anterior to origin of anal fin base, fin extending past anal fin insertion. Margin strongly convex with sharply rounded edge, deeply incised posteriorly. Caudal fin deeply forked with tips of lobes rounded; fin with i, 5 6 ,i ( 46) principal rays. Anal fin with short base, origin posterior to origin of adiposefin base, with ii,6 (10), ii,7 (12), iii,6 ( 19 ) or iii,7 (6) rays. Anal fin margin almost straight. Coloration Body variably mottled with dark saddles. First saddle posterior of head, second saddle at dorsal fin, third saddle between dorsal and adipose fins, fourth saddle under anterior of adipose fin, and fifth saddle on caudal peduncle. All saddles connected lat erally by broad stripe. Ventral region light brown with fourth and fifth saddle extending around ventral side. Dorsal, and anal fins light brown with faint medial bands Adipose fin dark brown to black, with posterior distal edge cream colored. Pectoral an d pelvic fins positioned horizontally with upper surfaces cream colored with darkened fin base and faint medial bands. Lower surfaces light yellow. Caudal fin cream colored with dark

PAGE 183

183 markings on upper and lower lobes. Caudal coloration asymmetrical, with l ower lobe almost completely dark (except for small cream colored mark at base of caudal and cream colored tip). Upper lobe with less dark pigment, with dark blotch covering dorsal edge, but caudal base with large cream colored patch and tip cream colored. Distribution Lake Kyogo drainage, western Uganda. Also known from the Nzoia River, a northeastern tributary of Lake Victoria and the Lake Manyara basin, Tanzania (Figure 52). Discussion The present study examined over 2000 museum specimens of the Amphilius jacksonii complex and has determined that it consists of six species, five of which were previously undescribed. Species of the complex are distinguished primarily by differences in body shapes, most notably by differences in body depth and caudal peduncle length and depth. Amphilius jacksonii is the species with the most slender body and the longest and most slender caudal peduncle. Amphilius n. sp. Rufiji has the deepest body and the shortest and deepest caudal peduncle and appears to have a much small er maximum body size then the other species in the complex. The largest Amphilius n. sp. Rufiji specimen examined is only 86.3 mm SL while specimens greater than 115 mm SL were examined for all other species. Additionally only six of the 186 specimens exam ined for this species had a standard length greater than 70 mm while specimens greater than 80 mm SL were common in collections of all the other species with many specimens greater than 100 mm SL. Frequency tables were constructed for counts of four differ ent meristic characters (Tables 7 to 10). Branchiostegal ray counts, total gill raker counts, and branched

PAGE 184

184 pectoral fin ray counts were found to be useful for distinguishing species. The high number of branchiostegal ray s (8 or 9) found in A. n. sp. Congo distinguishes it from both A. n. sp. Rufiji and A. n. sp. Ruzizi which typically have 6 or 7 branchiostegal ray s (Table 58 ). Amphius n. sp. Congo also had higher total gill raker counts than all the other species in the complex. Almost all specimens of A. n. sp. Congo examined had 10 or 11 gill rakers on the first ceratobranchial while almost all specimens examined of the other species had less than 10 (Table 5 9 ). The low number of branched pectoral fin rays found in A. n. sp. Rufiji distinguishes it from A. n. sp. Congo and A. n. sp. Lake Kyogo In A. n. sp. Rufiji, 180 of the 185 specimens examined had less than nine branched pectoral fin rays while the 47 specimens of A. n. sp. Lake Kyogo and the 88 specimens of A. n. sp. Congo examined all had nine or more branched pectoral fin rays (Table 5 10 ). No clear patterns were observed in the frequency distribution table of branched anal fin rays (Table 5 11 ). In addition to the characters in the diagnoses, there are coloration differences between species. In A n. sp. Congo and A. n. sp. Lake Kyogo the pigment on the caudal fins of juvenile and adult specimens are strongly asymmetrical, with the lower lobe almost completely pigmented (except for small cream colored mark at base of caudal and cream colored tip) and the upper lobe with less pigment, with dark blotch covering dorsal edge, but caudal base with large cream colored patch and tip cream colored. In A. jacksonii, A. n. sp. Malagarasi and A. n. sp. Rufiji, juveniles have similar caudal fin coloration, but adults have spotted caudal fins that are less strongly asymmetrical. Amphilius n. sp. Ruzizi differs in caudal fin coloration from all the other species in the complex. It does not have an asymmetrical caudal fin pigmentation

PAGE 185

185 pattern; instead it has dark a single dark band on the caudal fin with similar amounts of pigment on the upper and lower lobes. Amphilius n. sp. Ruzizi also differs from other species in the complex in paired fin coloration and body coloration. The paired fins in A. n. sp. Ruzizi do not have dark medial bands as in the other species. Instead, the paired fins are almost entirely brown with only the distal edge cream colored. The body coloration in A. n. sp. Ruzizi is similar to other species in the complex in having the same variably mottled coloration with dark saddles, but the areas in between the saddles are uniformly cream colored (vs. these areas cream colored but also stippled with darker pigment in other species of the complex). All specimens examined of A. n. sp. Congo have muc h bolder dark pigment than what was observed in specimens of all the other species of the complex. The dark pigment the fins and on the body of specimens is consistently much bolder in both juvenile and adult specimens of A. n. sp. Congo than juvenile and adult specimens of all other species in the complex. There is at least one additional undescribed species of the A. jacksonii complex that is not described in this study. This species occurs only in the Ivi River system, Lake Edwards drainage, in southwest ern Uganda and is being studied by another researcher. This undescribed species was included in a phylogenetic analysis of Amphlius that included five of the six species of the complex that are described in this study (Chapter 2). In the analyses that incl uded MtDNA, this undescribed species could not be distinguished from A. jacksonii but in the analyses that included only nuclear DNA, it was recovered as sister to all other members of the complex.

PAGE 186

186 There may be addition undescribed species in the complex in the Rufiji basin. Four specimens examined from the Little Ruaha River system (SAIAB 59406) were very similar to A. n. sp. Rufiji but differ from it by having eight unbranched anal fin rays, more than any of the 186 specimens examined of A. n. sp. Rufiji and more that the number in any other species in the complex with the exception of some specimens of A. jacksonii from the Kagera drainage, Nile basin (Table 10). These specimens are recognized here as A. cf. n. sp. Rufiji but are not formally described because of the small number of specimens examined and because they appear to differ from A. n. sp. Rufiji by just this single character. Several lots of specimens collected from the Rufiji and deposited at the American Museum of Natural History (AMNH 215916, 215922, 215930, and 215936) may also represent an undescribed species. These specimens were determined not to be A. n. sp. Rufiji but were not studied further because the locality data for them is vague and it is unclear what part of the river basin they were collected from. Further study is needed of the material from the Upper Malagarasi deposited in MRAC and considered here to be conspecific with A. n. sp. Malagarasi Although all these specimens were examined during a visit to MRAC, meristics were not taken on them and were identified as A. n. sp. Malagarasi without direct comparison to them Comparative Material Examined Amphilius cf. n. sp. Rufiji: Little Ruaha River system, Rufiji R iver basin: SAIAB 59406, Tanzania, About 5 km on road Iringa to Rua ha, 7 46'58"S, 3545'54"E (4: 54.765.2). Amphilius sp.: Kilombero River system, Rufiji R iver basin: AMNH 215916 Tanzania, Udzungwa mountains national park, Man'gula camp site no. 3 on Mwaya

PAGE 187

187 River ca. 7 51'S, 3653'E (5: 3741); AMNH 215922, same data as AMNH 215916 (1: 58.9); AMNH 215930, same data as AMNH 215916 (2: 60.936.3); AMNH 2159 36, same data as AMNH 215916 (1: 38.6).

PAGE 188

188 Table 51 Morphometric data for Amphilius jacksonii Range and mean include the holotype. Holotype Range (n=31) MeanSD %SL Head length 22.2 21.2 26.6 24.01.3 Head width 18.2 18.1 21.7 19.41.0 Head height 11.8 11.5 14.0 12.50.8 Body depth 13.0 11.5 17.0 14.41.3 Body depth at anus 11.5 9.6 12.0 11.30.7 Predorsal length 37.7 31.7 39.8 36.21.5 Prepectoral length 18 .5 17.4 22.7 19.51.4 Preanal length 73.1 67.8 77.4 71.22.0 Dorsal fin base length 9.8 8.5 13.4 10.81.2 Adipose fin base length 18.2 25.9 21.72.1 Anal fin base length 10.6 10.2 18.0 12.01.5 Pelvic fin length 19.3 17.9 23.6 20.51.3 Pectoral fin length 22.8 19.7 27.7 23.61.8 Anal fin length 15.6 13.7 22.5 19.31.9 Caudal peduncle length 18.6 16.7 20.1 18.70.8 Caudal peduncle depth 7.3 4.8 7.9 7.10.6 Anus to anal fin length 17.3 11.8 17.9 14.51.5 Prepelvic length 50.5 43.9 50.9 48.41.7 Postpelvic length 49.7 47.0 56.2 52.42.2 Dorsal fin insertion to adipose 40.6 43.7 42.20.9 Dorsal fin origin to caudal 59.7 71.7 66.22.1 Preanus length 51.7 60.0 56.62.0 %HL Snout length 44.8 44.8 55.1 48.32.2 Interorbital distance 29. 2 26.7 32.2 28.31.2 Maxillary barbel length 65.1 64.8 90.2 75.56.5 Inner mandibular barbel length 34.1 47.2 40.83.1 Outer mandibular barbel length 49.8 73.1 62.26.1 Eye diameter 15.6 13.4 19.2 16.51.4

PAGE 189

189 Table 5 2 External traits diagnostic for species of the Amphilius jacksonii complex. Character Amphilius jacksonii A. n. sp. Ruzizi A. n. sp. Malagarasi A. n. sp. Congo A. n. sp. Rufiji A. n. sp. Lake Kyogo Branchiostegal rays 7 8, rarely 6 or 9 6 7 7 8 8 9 6 7, rarely 8 7 8 Branched P ectoral Rays 8 9 rarely 7 8 9 8 10 9 10 7 8, rarely 9 9 Total Gill Rakers 6 9, rarely 10 6 8, rarely 5 or 9 7 9, rarely 6 or 10 10 11 rarely 9 or 12 6 8, rarely 9 6 8, rarely 5 or 9 G ill rakers on first ceratobranchial 4 7, rarely 8 4 6, rarely 3 or 7 5 6, rarely 4 or 7 7 8, rarely 6 or 9 4 6 rarely 3 5 6 rarely 4 or 7 Caudal peduncle depth 4.8 7.9 8.6 9.5 9.9 12.3 9.7 11.4 10.1 12.0 8.1 9.5 Caudal peduncle length 16.7 20.6 18.7 20.3 16.0 18.8 14.4 16.4 13.3 15.5 16.8 20.9 Interorbital width 26.7 32.0 23.4 25.1 28.1 35.8 25.1 27.7 28.9 34.1 28.3 31.7 Dorsal fin insertion to adiposefin insertion 40.6 43.7 38.2 41.5 37.5 41.7 37.5 41.2 42.2 44.6 38.2 42.0 Body depth at anus 9.6 13.2 11.1 12.9 11.4 17.7 13.9 14.9 14.5 17.4 13.5 15.8

PAGE 190

190 Tabl e 53 Morphometric data for Amphilius n. sp. Ruzizi Range and mean include the holotype. Holotype Range (n=7) MeanSD %SL Head length 21.6 21.6 26.9 24.52.1 Head width 17.9 17.9 20.3 19.21.1 Head height 11.5 11.5 13.7 12.50.8 Body depth 14.0 12.9 14 .2 13.60.5 Body depth at anus 11.1 11.1 12.9 12.10.8 Predorsal length 37.8 35.2 40.1 37.61.7 Prepectoral length 16.6 15.1 22.4 18.92.7 Preanal length 69.2 68.8 72.4 70.11.2 Dorsal fin base length 8.6 8.6 10.7 9.70.7 Adipose fin base length 15.0 15.0 18.9 16.41.3 Anal fin base length 8.1 8.1 10.5 8.90.9 Pelvic fin length 17.9 16.6 21.3 19.21.7 Pectoral fin length 20.9 19.9 24.9 22.52.1 Anal fin length 15.7 15.7 19.3 17.71.8 Caudal peduncle length 20.3 18.7 20.3 19.40.6 Caudal peduncle depth 8.6 8.6 9.5 9.10.3 Anus to anal fin length 17.9 13.8 17.9 15.01.5 Prepelvic length 47.4 46.5 50.3 48.21.3 Postpelvic length 51.4 51.4 53.6 52.50.9 Dorsal fin insertion to adipose 40.6 38.2 41.5 39.91.3 Dorsal fin origin to caudal 64.5 62.2 66.2 64.11.5 Preanus length 53.7 53.7 58.9 56.31.8 %HL Snout length 51.5 45.6 51.5 48.12.3 Interorbital distance 23.7 23.4 25.1 24.20.7 Maxillary barbel length 59.5 43.2 59.5 53.06.2 Inner mandibular barbel length 26.2 25.6 33.3 29.22.8 O uter mandibular barbel length 42.3 40.3 47.3 44.22.7 Eye diameter 12.7 11.6 14.3 12.80.8

PAGE 191

191 Table 54 Morphometric data for Amphilius n. sp. Malagarasi Range and mean include the holotype. Holotype Range (n=32) MeanSD %SL Head length 25.7 23.2 26.8 25.10.9 Head width 22.0 20.2 22.6 21.00.6 Head height 15.3 12.8 16.8 14.60.9 Body depth 20.0 14.1 20.2 17.91.5 Body depth at anus 16.9 13.7 17.7 15.70.9 Predorsal length 37.4 32.7 39.5 37.01.3 Prepectoral length 19.5 18.6 22.7 20.11.0 Pr eanal length 72.7 70.1 76.7 72.51.7 Dorsal fin base length 10.9 9.1 13.3 11.41.0 Adipose fin base length 16.8 16.7 23.5 20.21.7 Anal fin base length 11.7 10.0 14.1 11.91.1 Pelvic fin length 19.0 17.7 20.7 19.20.7 Pectoral fin length 21.6 19.7 24. 3 22.01.1 Anal fin length 19.3 17.7 21.3 19.80.9 Caudal peduncle length 17.9 16.0 18.8 16.80.8 Caudal peduncle depth 11.4 10.1 12.3 11.30.5 Anus to anal fin length 15.5 12.4 16.9 14.91.1 Prepelvic length 50.7 47.3 52.1 49.61.2 Postpelvic length 50.6 49.2 55.1 51.61.3 Dorsal fin insertion to adipose 40.5 37.5 41.7 40.31.0 Dorsal fin origin to caudal 66.6 62.4 67.9 65.51.5 Preanus length 56.5 55.8 61.4 58.31.2 %HL Snout length 48.5 41.7 51.2 46.72.4 Interorbital distance 29.8 28.1 3 5.8 31.62.0 Maxillary barbel length 72.5 61.4 98.0 77.38.2 Inner mandibular barbel length 40.0 28.3 54.3 41.55.7 Outer mandibular barbel length 60.4 44.0 81.0 62.58.3 Eye diameter 15.1 14.0 19.3 16.11.3

PAGE 192

192 Table 55 Morphometric data for Amphiliu s n. sp. Malagarasi Range and mean include the holotype. Holotype Range (n=20) MeanSD %SL Head length 25.6 25.4 28.0 27.10.7 Head width 19.8 19.8 22.2 21.10.6 Head height 12.5 12.1 14.1 13.50.5 Body depth 15.3 15.0 17.3 16.00.7 Body depth a t anus 14.1 13.9 14.9 14.30.3 Predorsal length 37.2 36.2 37.8 37.20.6 Prepectoral length 20.4 19.7 21.3 20.60.5 Preanal length 72.4 71.2 73.6 72.50.7 Dorsal fin base length 11.8 9.6 12.7 11.01.0 Adipose fin base length 20.0 18.6 22.9 20.51.1 An al fin base length 11.6 11.5 14.3 12.60.8 Pelvic fin length 18.1 17.9 19.2 18.40.4 Pectoral fin length 19.2 19.2 23.5 21.81.0 Anal fin length 19.1 19.0 22.3 20.20.9 Caudal peduncle length 15.7 14.4 16.4 16.90.7 Caudal peduncle depth 10.2 9.9 12.3 15.40.5 Anus to anal fin length 14.5 11.7 15.7 13.80.9 Prepelvic length 49.6 48.6 50.9 50.00.7 Postpelvic length 50.8 50.3 53.1 51.10.7 Dorsal fin insertion to adipose 40.7 37.5 41.2 39.51.1 Dorsal fin origin to caudal 66.5 63.5 66.5 64.90.9 P reanus length 56.3 56.4 58.7 57.70.8 %HL Snout length 47.3 43.2 47.4 45.41.3 Interorbital distance 27.1 25.1 27.7 26.30.7 Maxillary barbel length 72.7 59.4 72.7 66.53.8 Inner mandibular barbel length 40.1 31.3 40.6 36.12.4 Outer mandibular b arbel length 59.8 47.0 63.6 54.34.3 Eye diameter 13.3 12.1 16.2 14.31.1

PAGE 193

193 Table 56 Morphometric data for Amphilius n. sp. Rufiji. Range and mean include the holotype. Holotype Range (n=68) MeanSD %SL Head length 25.9 24.0 26.7 25.40.7 Head w idth 21.1 19.6 22.1 20.60.5 Head height 13.5 12.9 15.2 13.80.5 Body depth 17.3 15.6 19.3 17.00.8 Body depth at anus 17.0 14.5 17.4 15.70.7 Predorsal length 37.7 35.6 39.7 37.41.1 Prepectoral length 19.2 18.4 21.1 19.80.6 Preanal length 74.1 70. 7 75.0 72.91.1 Dorsal fin base length 12.2 10.1 12.3 11.10.6 Adipose fin base length 22.7 19.3 24.2 21.21.2 Anal fin base length 10.9 10.5 13.8 11.90.7 Pelvic fin length 20.6 18.2 20.9 19.80.7 Pectoral fin length 21.7 20.5 24.1 22.20.7 Anal fin length 19.2 18.1 21.4 19.70.8 Caudal peduncle length 15.3 13.3 15.5 14.80.6 Caudal peduncle depth 10.5 10.1 12.0 10.90.4 Anus to anal fin length 14.9 12.4 16.6 14.20.9 Prepelvic length 49.8 46.7 51.6 49.21.1 Postpelvic length 50.7 49.3 54.6 51.4 .1 Dorsal fin insertion to adipose 43.3 42.2 44.6 43.10.6 Dorsal fin origin to caudal 67.8 63.2 68.3 66.21.2 Preanus length 58.5 56.8 60.9 58.81.1 %HL Snout length 50.5 45.3 50.7 48.21.2 Interorbital distance 28.9 28.9 34.1 31.01.3 Maxill ary barbel length 76.1 65.5 91.4 75.05.6 Inner mandibular barbel length 34.3 33.7 50.7 39.83.7 Outer mandibular barbel length 60.4 51.1 82.2 64.16.0 Eye diameter 12.5 12.1 16.7 13.81.1

PAGE 194

194 Table 57 Morphometric data for Amphilius n. sp. Lake Kyogo. Range and mean include the holotype. Holotype Range (n=45) MeanSD %SL Head length 24.2 22.4 25.0 24.11.3 Head width 19.5 19.1 20.7 19.50.9 Head height 13.3 11.7 14.3 13.21.3 Body depth 16.7 14.4 18.5 16.72.1 Body depth at anus 15.6 13.5 15. 8 14.31.2 Predorsal length 36.0 34.4 38.3 36.52.0 Prepectoral length 18.9 17.6 20.5 19.21.5 Preanal length 72.9 69.0 74.4 71.52.7 Dorsal fin base length 11.4 10.3 13.3 11.51.5 Adipose fin base length 22.5 18.1 23.1 20.32.5 Anal fin base length 12.2 10.3 14.0 11.81.8 Pelvic fin length 17.8 17.5 21.5 19.22.0 Pectoral fin length 21.8 20.0 23.6 22.31.8 Anal fin length 16.8 16.8 20.9 18.82.1 Caudal peduncle length 17.1 16.8 20.9 18.12.1 Caudal peduncle depth 9.6 8.1 9.5 8.907 Anus to anal fin length 15.1 11.8 16.2 14.42.2 Prepelvic length 50.1 46.5 50.3 48.61.9 Postpelvic length 51.8 49.3 56.2 52.13.5 Dorsal fin insertion to adipose 41.8 38.2 42.0 40.61.9 Dorsal fin origin to caudal 66.2 65.3 69.0 66.81.9 Preanus length 58.7 54.8 60.6 57.22.9 %HL Snout length 50.5 42.9 52.9 49.55.1 Interorbital distance 28.7 28.3 31.7 29.21.8 Maxillary barbel length 83.3 61.6 85.0 73.211.7 Inner mandibular barbel length 38.7 35.1 44.6 40.64.8 Outer mandibular barbel length 61.9 48. 2 72.4 61.912.2 Eye diameter 13.8 12.6 18.0 15.22.7

PAGE 195

195 Table 5 8 Branchiostegal ray counts in A. jacksonii, A. n. sp. Ruzizi A. n. sp. Malagarasi, A. n. sp. Congo, A. n. sp. Rufiji, and A. n. sp. Lake Kyogo Branchiostegal Rays 5 6 7 8 9 A. jacksoni 5 85 147 1 A. n. sp. Ruzizi 7 10 A. n. sp. Malagarasi 37 138 A. n. sp. Congo 72 14 A. n. sp. Rufiji 25 158 2 A. n. sp. Lake Kyogo 22 25 Table 5 9 Total gill raker counts in A. jacksonii, A. n. sp. Ruzizi A. n. sp. Malagarasi A. n sp. Congo, A. n. sp. Ruzizi and A. n. sp. Lake Kyogo Total Gill Rakers 5 6 7 8 9 10 11 12 A. jacksoni 25 74 83 38 6 A. n. sp. Ruzizi 1 8 2 6 1 A. n. sp. Malagarasi 5 27 72 59 10 A. n. sp. Congo 2 50 31 3 A. n. sp. Rufiji 1 25 102 48 8 A. n. sp. Lake Kyogo 2 16 24 5 Table 5 10 Branched pectoral fin ray counts in A. jacksonii, A. n. sp. Ruzizi A. n. sp. Malagarasi A. n. sp. Congo, A. n. sp. Ruzizi and A. n. sp. Lake Kyogo Branched Pectoral Fin Rays 6 7 8 9 10 A. jackso ni 1 104 137 A. n. sp. Ruzizi 4 14 A. n. sp. Malagarasi 81 94 5 A. n. sp. Congo 84 4 A. n. sp. Rufiji 14 166 5 A. n. sp. Lake Kyogo 47 Table 5 11 Branched anal fin ray A. jacksonii, A. n. sp. Ruzizi A. n. sp. Malagarasi A. n. s p. Congo, A. n. sp. Ruzizi and A. n. sp. Lake Kyogo Branched Anal Fin Rays 5 6 7 8 A. jacksoni 110 124 9 A. n. sp. Ruzizi 4 9 5 A. n. sp. Malagarasi 6 97 63 A. n. sp. Congo 4 69 15 A. n. sp. Rufiji 5 134 47 A. n. sp. Lake Kyogo 29 18

PAGE 196

196 Figure 5 1 Amphilius jacksoni UF 110743 XXmm SL; lateral, dorsal and ventral view.

PAGE 197

197 Figure 5 2 Distributions of species of the Amphilius jacksonii complex.

PAGE 198

198 Figure 5 3 Amphilius n. sp. Ruzizi MRAC 9329493300 82.8 mm SL, Holotype; lateral, dorsal and ventral view.

PAGE 199

199 Figure 5 4 Amphilius n. sp. Malagarisi, CU 97334, 52.3 mm SL, Holotype; lateral, dorsal and ventral view.

PAGE 200

200 Figure 5 5 Amphilius n. sp. Congo CU 97335, 83.1 mm SL, Holotype; lateral, dorsal and ventral view.

PAGE 201

201 Figure 5 6 Amphi lius n. sp. Rufiji, UF 184237, 86.3 mm SL, Holotype; lateral, dorsal and ventral view.

PAGE 202

202 Figure 57 Amphilius n. sp. Lake Kyogo, UF 184238, 104.1 mm SL, Holotype; lateral, dorsal and ventral view.

PAGE 203

203 CHAPTER 6 CONCLUSIONS Although African fishes have been the subject of a great amount of research, the taxonomy and phylogenetic relationships of many African fish groups remain poorly known. One of the most widely distributed and diverse families of African fishes is the catfish family Amphiliidae. Cha pter 2 is the first phylogenetic analysis of the family using molecular data, and the taxonomic sampling is much greater than in the previous morphological studies of the family. The molecular analyses provided strong support two of the three currently rec ognized subfamilies but all analyses recovered Paramphilius baudoni as sister to all remaining amphiliids (Doumeinae + Leptoglanidinae + Amphilius ), rendering the subfamily Amphiliinae polyphyletic. These results indicate that Paramphilius should be placed in its own subfamily However, Diogo (2005) found strong support for a monophyletic Amphiliinae. The studies are not directly comparable because the species of Paramphilius used in each study differ, and because P. baudoni is not the type species of the genus P lacing P. baudoni in a new subfamily based on the results of the current study would be premature and create uncertainty about what to do with remaining species of Paramphilius The extensive taxonomic sampling in the study allowed for the monophyly of several genera to be tested for the first time. In the subfamily Leptoglanidinae, most species are placed in a single genus, Zaireichthys Although only four of the 18 currently recognized species were included in the study, the genus was not recovered as monophyletic in any of the analyses. In the subfamily Doumeinae, all genera that had more than one species included in the study were not recovered as monophyletic in most analyses. Further research on the phylogenetic relationships of these two

PAGE 204

204 sub families will benefit from increased taxon sampling as additional tissue samples are obtained. In the absence of sequences from additional taxa, research on the phylogenetic relationships of theses subfamilies will continue to be primarily based on morphol ogy. Although all the genera in these subfamilies have been deliminated based on morphology, none of the characters used were identified using explicit phylogenetic methods. For example, the two largest d oumein genera, Doumea and Phractura, are primarily d efined by the absence of characters that are found in some of the other d oumein genera. I n his phylogenetic analysis of Amphiliidae, Diogo (2003) identif ied characters for all genera recognized at that time, but his taxon sampling was very limited (one of eight species of Doumea, two of 13 species of Phractura, and two of 18 species of Zaireichthys ) and it is unknown how useful they may be in defining genera. Chapter 3 is a phylogenetic analysis of Amphilius the largest genus in the family. The genus was recovered as a well supported clade in the concatenated analyses but was not recovered as monophyletic in any of the nuclear gene analyses. Additionally, in the analyses of the S7 intron, the High African species were recovered as sister to the genus Dolic hamphilius a genus in the subfamily Leptoglanidinae. Two morphological groups have been recognized in Amphilius a primarily Low African group of species that have an epidermal fold at the base of the caudal fin and 6 + 7 or 7 + 8 principal caudal fin ray s, and a primarily High African group of species that lack the epidermal fold and have 8 + 9 principal caudal fin rays (Skelton, 19 84; Thomson & Page, 2010). The H igh African Amphilius species were recovered as a well supported c lade in all analyses. The Low African Amphilius species were recovered as a clade in all the analyses that included nuclear genes but not in the two analyses that included only the

PAGE 205

205 cyt b locus. Given the strong evidence that the High African Amphilius is monophyletic the name Anop lopterus Pfeffer, 1889 was resurrected for the group. All the Low African species were retained in Amphilius Chapters 4 and 5 focus on the taxonomy of Anoplopterus and Amphilius respectively. The main goal of C hapter 4 was to determine the status of sever al nominal species that have been placed in the synonymy of Anoplopterus uranoscopus The nominal species were all determined to be valid species and redescribed. Additionally, two new species were described from eastern Tanzania. The additional diversity described in this chapter substantially increased the recognized number of species in Anoplopterus but there is still a large number of undescribed species awaiting descriptions. Chapter 5 focused on the taxonomy of the Amphilius jacksonii complex. Five n ew species of Amphilius are described and A. jacksonii is redescribed. Additional diversity may be present in the complex as discussed in the chapter. There is also a large amount of undescribed diversity in the genus outside the A. jacksonii complex tha t awaits description.

PAGE 206

206 LIST OF REFERENCES Bailey, R. G. 1969. The non cichlid fishes of the eastward flowing rivers of Tanzania, East Africa. Revue de Zoologie et Botanique Africaines. Tervuren 80: 170 199. Bailey, R. M., and D. J. Stewart. 1984. Bagrid catfishes from Lake Tanganyika, with a key and descriptions of new taxa. Miscellaneous Publications, Museum of Zoolog y of the University of Michigan. no. 168: i iv + 1 41. Balon, E. K. 1974. Fishes of the edge of Victoria Falls, Africa: demise of a physic al barrier for downstream invasions. Copeia, 1974 (no. 3): 643660, pls. 15.Barnard, K. H. 1942. A note on Amphilius natalensis Blgr. (Siluroidae, Amphiliidae). An nals of the Natal Museum 10: 257259. Bell Cross, G. 1973. The fish fauna of the Buzi River System in Rh odesia and Mozambique. Arnoldia 6: 1 14. Bell Cross, G. and R. A. Jubb, 1973. The Amphiliidae of Southern Africa and record of Amphilius lampei Pietschmann, 1913, from the Inyang a Mountains, Rhodesia. Arnoldia 6: 1 9. Bernacek, G. M. 1980. Introduction to the freshwater fishes of Tanzania. University of Dar es Salaa m, Dept. Zoology, Dar es Salaam. 79 p. Bertin, L. and R. Estve 1950. Catalogue des types de poissons du musum National d'Histoire Naturelle. 5e partie. Ostariophysaires (Siluri formes). Imprimerie Nationale, Paris. Ca t. Fish Types, Paris 5e partie: 1 85. Boulenger, G. A. 1898a. On the habit of the Siluroid fish Anoplopterus platychir Gthr. Annals and Mag azine of Natural History (Series 7) 8: 254 255. Boulenger, G. A. 1898b. Repo rt on the collection of fishes made by Mr. J. E. S. Moore in Lake Tanganyika during his expedition, 189596. Transactions of the Zoological Society of London 15: 1 30, pls. 18. Boulenger, G. A., 1901. Description of a new silurid fish of the genus Anoplopterus *, from Cameroon. Annals and M agazine of Natural History (Series 7) 8: 447 448. Boulenger, G. A. 1902. Additions la faune ichthyologique de bassin du Congo. Matriaux pour la faune du Congo. Annales du Muse du Congo, Zoologie 2: 19 57, pls. 716. Boulenger, G. A. 1905a. On a second collection of fishes made by Mr. S. L. Hinde in the Kenya District, East Africa. Proceedings of t he Zoological Society of London 1 : 62 64, pl. 7.

PAGE 207

207 Boulenger, G. A. 1905b. A list of freshwaters fishes of Africa. Annals and M agazine of Natural History (Series 7) 16: 3660. Boulenger, G. A. 1907. Descriptions of three new fishes from Central Africa. Annals and Magazine of Natural Hist ory (Series 7) 20:487 489. Boulenger, G. A. 1911. Catalogue of the freshwater fishes of Africa in the British Museum (Natural History), Vol. 2. British Museum (Natural History) Trustees, London. xii + 529 p. Boulenger, G. A. 1912. Descriptions of two new fishes from the Nile system. Annals and Magazine of Natural History (Series 8) 10: 6016 02. Boulenger, G. A. 1916. Catalogue of the freshwater fishes of Africa in the British Museum (Natural History), Addenda. British Museum (Natural History) Trustees, London. Boulenger, G. A. 1917. Description of a new silurid fish from Natal. Annals of the Durban Museum 1: 432. Boulenger, G. A. 1920d. Poissons de la mission Stappers, 19111913 pour lexploration hydrographique et biologique des Lacs Tanganyika et Moro. Revue de Zoologie Africaine 8: 1 57. Cailliet, G. M., M. S. Love, & A. W. Ebeling. 1986. Fishes: A Field and Laboratory Manual on Their Structure, Identification, and Natural History. Wadsworth Publishing Company, Belmont, CA, 194 p. Chardon, M. 1968. Anatomie compare de lappareil de Weber et des structures connexes chez les Siluriformes. Annales du Muse Royal de l'Afrique Centrale. Srie in 8, Sciences Zoolog iques. Tervuren, Belgique, 169: 1 285. Chow, S., and K. Hazama. 1998. Universal PCR primers for S7 ribosomal protein gene introns in fish. Molecular Ecology 7:1255 1256. Copley, H. 1958. Common freshwater fishes of East Africa. H. F. & G. Witherby Ltd., London, 172 p. Corbet, P. S. 1961. The food of noncichlid fishes in the Lake Victoria basin, with remarks on their evolution and adaptation to lacustrine conditions. Proceeding o f the Z oological Society of London 136: 1 101. Crass, R. S. 1960. Notes on the freshwater fishes of Natal with descriptions of four new species. Annals of the Natal Museum 14: 405 458.

PAGE 208

208 David, L. 1937. Poissons de l'Urundi. Revue de Zool ogie et Botanique Af ricaines 29: 413 420. David, L., and M. Poll. 1937. Contribution la faune ichthyologique du Congo Belge: Collections du Dr. H. Schouteden (19241926) et d'autres rcolteurs. Annales du Mus e du Congo Belge, Zoologie (Series 1) 3: 189 294, pl. 12. De Vos, L., L. Seegers, L. Taverne and D. F. E. Thys van den Audenaerde. 2001. Lichtyofaune du bassin de la Malagarasi (Systme du Lac Tanganyika): une synthse de la connaissance actu elle. Annales Muse Royal de lAfrique Centrale, Srie in 4, Sciences Zoolog iques 285: 117 135. Diogo, R. 2003. Anatomy, phylogeny, and taxonomy of Amphiliidae Pp. 401438, in: G. Arratia, B. G. Kapoor, M. Chardon, and R. Diogo (eds.), Catfishes. Science Publishers, Enfield, NH, USA. Diogo, R. 2005. Morphological Evolution, Aptat ions, Homoplasies, Constraints and Evolutionary Trends : Catfishes as a Case Study on General Phylogeny and Macroevolution. Science Publishers, Enfield, NH, USA, 491 pp. Diogo, R. and M. Chardon. 2000. The structures associated with catfish (Teleostei: Si luriformes) mandibular barbels: origin, anatomy, function and synonymy. Netherlands Journal of Zoology 50: 455 478. Edgar R. C. 2004. "MUSCLE: multiple sequence alignment with high accuracy and high throughput." Nucleic Acids Research 32 :1792 1797. Ferraris, C. J., Jr., P. Skelton, and R. P. Vari. 2010. Species of the Doumea chappuisi complex (Siluriformes, Amphiliidae) with the descriptions of new species from the upper Sanaga River and the Nyong R iver basins. Copeia 2010: 705715. Ferraris, C. J., Jr., R P. Vari, and P. Skelton. 2011. A new genus of the African loach catfish (Siluriformes: Amphiliidae) from the Congo River basin, the sister group to all other genera of the Doumeinae, with the description of two new species. Copeia 2011: 477 489. Galtier, N., M. Gouy, and C. Gautier. 1996. SEAVIEW and PHYLO_WIN: two graphic tools for sequence alignment and molecular phylogeny. Comput. Appl. Biosci., 12:543548. Greenwood, P. H. 1957. The fishes of Uganda. III. Uganda Journal 21: 47 80. Greenwood, P. H. 1958. The fishes of Uganda, Uganda Society, Kampala, ii + 1224. Greenwood, P. H. 1966. The fishes of Uganda, 2nd edition. Uganda Society, Kampala, 131 pp.

PAGE 209

209 Gnther, A. 1864. Catalogue of the fishes in the British Museum, vol. 5. Catalogue of the Physostomi containing the families Siluridae, Characinidae, Haplochitonidae, Sternoptychidae, Scopelidae, Stomiatidae in the collection of the British Museum. Trustees, London. xxii + 455 p. Hardman, M. 2004. The phylogenetic relationships among Noturus catfishes (Siluriformes: Ictaluridae) as inferred from mitochondrial gene cytochrome b and nuclear recombination activating gene 2. Molecular Phylogenetics and Evolution 30: 395 408. Hardman, M., and L. M. Page. 2003. Phylogenetic relationships among bullhead catfis hes of the genus Ameiurus (Siluriformes: Ictaluridae). Cope ia 2003: 2033. Harry, R. R. 1953. A contribution to the classification of the African catfishes of the family Amphiliidae, with description of collections from Cameroon. Revue de Zoologie et de Botanique Afr icaines 47: 177 200; 201232. He, S. P., M. Gayet, and F. J. Meunier. 1999. Phylogeny of the Amphiliidae (Teleostei: Siluriformes). Annales des Sciences Naturelles 20: 117146. Jubb, R. A. 1961. An illustrated guide to the freshwater fishes of t he Zambezi River, Lake Kariba, Pungwe, Subi, Lundi and Limpopo Rivers. Bulawayo: Stuart Manning. iiiix + 1 171. Jubb, R. and I. G. Gaigher 1971. Check list of the fis hes of Botswana. Arnoldia 5: 1 22. Mahnert, V. 1976. Catalogue des types de poissons, amphibiens, et reptiles du Musum d'Histoire naturelle de Genve. Revue Suisse de Zoologie, Annales de la Socit zoologique suisse et du Musum d'Histoire naturelle de Genve, 83: 471 496. Marlier, G. 1953. Etude biogographique du bassin de la Rizizi base sur la distribution des poissons. Annales de la Socit royal e zoologique de Belgique 84: 175 224. Matthes, H. 1967. The fishes and fisheries of the Ruaha River basin, Tanzania. East African Freshwater Fisheries Research Organisation (EAFFO) Occasional P apers No. 9:119. Mo, T. 1991. Anatomy, relationships and systematics of the Bagridae (Teleostei: Siluroidei) with a hypothesis of Siluroid phylogeny. Theses Zoologicae 17: 1 216, 63 fig. Pellegrin, J. 1905. Mission scientifique de Ch. Alluaud en Afrique orientale (Juin 1903 Mai 1904). Poissons. II. Systmatique. Mmoires de la Socit Zoologie de France 17: 167 185, pl. 16.

PAGE 210

210 Pellegrin, J. 1907. Silurid nouveau du FoutaDjalon. Bulletin de Museum National du Histoire Naturelle, Paris 13: 23 25. Pellegrin, J. 1919. Poissons nouveaux du Mozambique. Bulletin de la Socit Zoologique de France 44:397401. Pellegrin, J. 1933. Poissons de la rgion du Kivu adresss par M. Guy Babault. Bulletin de la Socit Zoologique de France 58: 169 175. Pellegrin, J. 192 8. Description d'un poisson nouveau du Gabon appartenant au genre Amphilius Bulletin de la Socit Zoologique de France 53: 453 454. Pellegrin, J. 1936. Contribution l'Ichtyologie de l'Angola. Arquivos do Museu Bocage No. 7: 4562. Pfeffer, G. J. 1889. bersicht der von Herrn Dr. Franz Stuhlmann in gypten, auf Sanzibar und dem gegenberliegenden Festlande gesammelten Reptilien, Amphibien, Fische, Mollusken und Krebse. Jahrbuch der Hamburgischen Wissenschaftlichen Anstalten, Hamburg 6:1 36. Pfeffer, G. J. 1893. Ostafrikanische Fische gesammelt von Herrn Dr. F. Stuhlmann im Jahre 1888 und 1889. Jahrbuch der Hamburgischen Wissenschaftlichen Anstalten. Hamburg 10:131177, 3 pls. Pfeffer, G. J. 1896. Die Thierwelt Ost Afrikas und der Nachbargebiete. Lief. v. Die Fische Ost Afrikas. Fische Ost Afrikas 3: i xviii + 1 72. Poche, F. 1902a. ber das Vorkommen von Anoplopterus platychir (Gthr) in We stafrika. Zoologischer Anzeiger 25: 121 122. Poche, F. 1902b. Richtigstellung eines Gattungsnamens unter den S ilur iden. Zoologischer Anzeiger 25: 211. Poll, M. 1939. Exploration du Parc National Albert. Mission G. F. De Witte (19331935). Poissons. Inst. P arcs Nat. Congo Belge No. 24: 81, pls. 111. Posada, D., and K. A. Crandall. 1998. MODELTEST: testing the model of DNA substitution. Bioinformatics 14:817 818. Rambaut, A., and A. J. Drummond. 2004. Tracer. University of Oxford, Oxford. Regan, C. T. 1911. The classification of the teleostean fishes of the order Ostariophysi. 2. Siluroidea. Annals and M agazine of Natural History (Series 8) 8: 553 577.

PAGE 211

211 Roberts, T. R. 1967. Rheoglanis dendrophorus and Zaireichthys zonatus bagrid catfishes from the lower rapids of the Congo River. Ichthyologia, The Aquarium Journal 39: 119131. Roberts, T. R. 1975. Geographical distr ibution of African freshwater fishes. Zoological Journal of the Linnean Society 57:249 319. Roberts, T. R. 2003. Systematics and osteology of Leptoglaninae, a new subfamily of the African catfish family Amphiliidae, with descriptions of three new genera a nd six new species. Proceedings of the California Acad emy of Sciences 54: 81 132. Schmidt, R. C. and F. Pezold. 2011. Morphometric and molecular variation in mountain catfishes (Amphiliidae: Amphilius ) in Guinea, West Africa. Journal of Natural History 45: 521552. Seegers, L. 1996. The fishes of the Lake Rukwa drainage. Annales de la Socit Zoologique de Belgique 278: 1 407. Seegers, L., L. De Vos and D. O. Okeyo. 2003. Annotated checklist of the freshwater fishes of Kenya (excluding the lacustrine haploc hromines from Lake Victoria). Journal of East African Natural History 92:1147. Skelton, P. H. 1981. The description and osteology of a new species of Gephyroglanis (Siluriformes, Bagridae) from the Olifants River, South West Cape, South Africa. Annals of the Cape Provincial Museums, Natural History 13:217249. Skelton, P. H. 1984. A systematic revision of species of the catfish genus Amphilius (Siluroidei, Amphiliidae) from east and southern Africa. Annals of the Cape Provincial Museums, Natural History 16:4171. Skelton, P. H. 1986. Two new Amphilius (Pisces, Siluroidei, Amphiliidae) from the Zare River system, Africa. Revue de Zoologie Africaine 99:263 291. Skelton, P. H. 1989. Descriptions of two new species of West African amphiliid catfishes (Sil uroidei: Amphiliidae). J. L. B. Smith Institute of Ichthyolog y, Special Publication no. 48: 1 13. Skelton, P. H. 1994. Diversity and distribution of freshwater fishes in East and Southern Africa. p.95131. In G.G.Teugels, J.F. Gugan and J.J. Albaret (eds. ) Biological diversity of African freshand brackish water fishes. Geographical overviews presented at the PARADI Symposium, Senegal, 1520 November 1993. Annales du Muse Royal de l 'Afrique Centrale, Sciences Zoologiques 275 p. Skelton, P. H. 2001. A c omplete guide to the freshwater fishes of southern Africa, second edition. Struik Publishers, Cape Town. xiv + 395 p.

PAGE 212

212 Skelton, P. H. 2007a New species of the amphiliid catfish genera Amphilius, Doumea and Phractura and the taxonomy of Paramphilius from w est central Africa (Siluriformes, Amphiliidae). Zootaxa No. 1578: 4168. Skelton, P. H. 2007b. Amphiliidae (p. 753789). In: The Fresh and Brackish Water Fishes of Lower Guinea, West Central Africa / Poissons d'eaux douces et saumtres de basse Guine, oues t de l'Afrique centrale. Volume 1. Paris, Tervuren: IRD, MRAC, MNHN, 2007, 800 p. Skelton, P. H., M. N. Bruton, G. S. Merron, and B. C. W. van der Waal. 1985. The fishes of the Okavango drainage system in Angola, South West Africa and Botswana: Taxonomy and distribution. Ichthyological Bulletin of the J. L. B. Smith In stitute of Ichthyology No. 50: 1 21. Sullivan, J. P., J. G. Lundberg and M. Hardman, 2006. A phylogenetic analysis of the major groups of catfishes (Teleostei: Siluriformes) using rag1 and rag2 nuclear gene sequences. Molecular Phylogenetics and Evolution 22: 870 878. Swofford, D. L. 2003. PAUP* 4.0b10: phylogenetic analysis using parsimony (*and other methods). Sinauer Associates, Sunderland, Massachusetts. Tan, H. H. & M. Kottelat, 1998. Redescription of Betta picta (Teleostei: Osphronemidae) and description of Betta falx sp. n. from central Sumatra. Revue Suisse de Zoology 105: 557 568. Thomson, A. W. and L. M. Page. 2010. Taxonomic revision of the Amphilius uranoscopus group (Teleostei: Si luriformes) in Kenya, with the description of a new species form the Athi River. Bulletin of the Florida Museum of Natural History 49: 45 66. Tweddle, D. and N. G. Willoughby, 1976. Dry season fish population in the Shire V alley Game Reserve. Nyala 2: 3 14. Tweddle, D. and N. G. Willoughby, 1979. An annoted checklist of the fish fauna of the River Shir South of Kapachira Falls, Malawi. Ichthyological Bulletin, Rhodes University 39: 1122. Vaillant, L. L. 1897. Silurode nouveau de l'Afrique orientale ( Chim arrhoglanis leroyi ). Bulletin du Musum dHistoire Naturelle, Paris 3:81 84. Van der Waal, B. C. W. and P. H. Skelton 1984. Check list of fishes of Caprivi. Madoqua 13: 303320, 4 tables. Walsh, S. J., L. J. Chapman, A. E. Rosenberger and C. A. Chapman. 2000. Redescription of Amphilius jacksonii (Siluriformes: Amphiliidae) with habitat and life history notes. Ichthyological Ex ploration of Freshwaters 11: 163 174.

PAGE 213

213 Weber, C. 1998. Catalogue rvis des types primaires de la collection ichthyologique du Mus um d'histoire naturelle de la Ville de Genve (MHNG). Revue Suisse de Zoologie, Annales de la Socit zoologique suisse et du Musum d'Histoire naturelle de Genve 105: 3 14. Whitehead, P. J. P. 1958. A new species of Chiloglanis (Pisces, Mochocidae) in Kenya. Annals and Magazine of Natural History (Series 13) 1: 197 208. Zwickl, D. J. 2006. Genetic algorithm approaches for the phylogenetic analysis of large biological sequence datasets under the maximum likelihood criterion. Ph.D. dissertation, The Univers ity of Texas at Austin.

PAGE 214

214 BIOGRAPHICAL SKETCH Alfred W. Thomson was born December 30, 1978, in Englewood, New Jersey. One of three children, he was interested in the natural world from an early age and has a lways been especially interested in fishes. He received his Bachelor of Science degree from Mansfield University in August 2001, and moved to Kenosha, Wisconsin to work as a research technician for the Illinois Natural History Survey at the Lake Michigan B iological Station in Zion, Illinois. In January 2004 he moved to Gainesville, Florida and worked as a research technician on the All Catfish Species Inventory Project at the Florida Museum of Natural History. In August 2005 he enrolled in UFs Department o f Zoology as a full time graduate student with a research assistantship from the All Catfish Species Inventory Project. He received his masters degree from the Zoology in August 2007 and was then admitted to the department as a Ph.D. student. He received his Ph.D. from the University of Florida in the spring of 2013. He is currently the collection manager for the ichthyology collection at the Florida Fish and Wildlife Research Institute in Saint Petersburg, FL.