u. HARVARD UNIVERSITY Library of the Museum of Comparative Zoology UNIVERSITY OF KANSAS MUSEUM OF NATURAL HISTORY PUBLICATIONS The University ot Kansas Publications, Museum of Natural History, beginning with volume 1 in 1946, was discontinued with volume 20 in 197 1 . Shorter research papers formerly published in the above series are now published as The University of Kansas Museum of Natural History Occasional Papers. The University of Kansas Museum of Natural History Miscellaneous Publications began with number 1 in 1946. Longer research papers are published in that series. Monographs of the Museum of Natural History were initiated in 1970. Authors should contact the managing editor regarding style and submission procedures before manuscript submission. All manuscripts are subjected to critical review by intra- and extramural specialists; final acceptance is at the discretion of the Director. This publication is printed on acid-free paper. Occasional Papers and Miscellaneous Publications are typeset using Microsoft" Word and Aldus PageMaker" on a Macintosh computer. Institutional libraries interested in exchanging publications may obtain the Occasional Papers and Miscellaneous Publications by addressing the Exchange Librarian, The University of Kansas Library, Lawrence, Kansas 66045-2800, USA. Individuals may purchase separate numbers from the Office of Publications, Museum of Natural History, The University of Kansas, Lawrence, Kansas 66045-2454, USA. Editor. Linda Trueb Managing Editor. Joseph T. Collins Formatting: Kate Shaw Printed by University of Kansas Printing Service Lawrence, Kansas MCZ LIBRARY OCCASIONAL PAPERS ^^^^ 0 9 1992 of the HARVARD MUSEUM OF NATUttSlt'fft^bRY The University of Kansas Lawrence, Kansas NUMBER 152. PAGES 1-18 28 OCTOBER 1992 Phylogenetic Relationships among Members OF THE Hybopsis dorsalis Species Group (Teleostei: Cyprinidae) E. O. Wiley and Tom A. Titus' Division of Ichthyology, Museum of Natural History and Department ofSysteniatics and Ecology. The University of Kansas. Lawrence, Kansas 66045-2454. USA The Hybopsis dorsalis species group includes five described species that formerly were placed in Notropis (Mayden, 1989). Hybopsis dorsalis Agassiz (1854) is the northernmost species of the group, having a range centered in the upper Mississippi River Drainage with several disjunct populations to the east (Lee et al., 1980). Hybopsis sabinae (Jordan and Gilbert, 1886) occurs in western Gulf Coast drainages from the San Jacinto Drainage of Texas east to the Calcasieu and lower Red River drainages of Louisiana, with disjunct populations in northeastern Arkansas, southeast- ern Missouri, and the Yazoo River of Mississippi (Suttkus, 199 1 ). Hybopsis ammophiliis (Suttkus and Boschung, 1990) is largely confined to the Mo- bile Bay Drainage, with populations in adjacent tributaries of the Yazoo. Hatchee, and Tennessee drainages. Hybopsis longirostris (Hay, 1881), the longnose shiner, occupies the northern Gulf Coastal Plain from the lower Mississippi River Drainage of Louisiana and Mississippi east to the Apalachicola River Drainage of Florida, Alabama, and Georgia and the 'Present address: Department of Biology, Washington University, St. Louis, Missouri 63130-4899, USA. © Museum of Natural History. The University of Kansas. Lawrence. ISSN:009 1-7958 2 UNIV. KANSAS MUS. NAT. HIST. OCC. PAP. No. 152 upper Altamaha River Drainage of Georgia. An apparently disjunct popula- tion is found in the Ouachita River Drainage of Louisiana (Gilbert and Burgess, 1980). Although largely absent from the Mobile Bay Drainage, where it is replaced by H. ammophilns, H. l<>ni>irostris is known from two tributaries of the lower Alabama River (Suttkus and Boschung. 1990; Suttkus. 1991). Hyhopsis rafinesque (Suttkus, 1991) is endemic to the upper tributaries of the Yazoo River Drainage of the Bluff Hills and North Central Plateau physiographic provinces of Mississippi (Suttkus, 1991). Although much is known about the biology and life history of Hyhopsis longirostris (Hubbs and Walker, 1942; Heins and Clemmer, 1975, 1976) and H. ammophilus (Heins et al., 1980), little is known about evolutionary relationships within the group. Hubbs and Walker (1942) noted similarities in ecology and morphology among members of the H. dorsalis group and other minnows inhabiting clear, sandy streams (cf. H. huccata). Smith- Vaniz (1968) called attention to differences in fin color between the two populations of//, longirostris east and west of the Mobile Bay Drain- age Basin and to the existence of an undescribed species (now H. ammophilus) inhabiting the Mobile Bay Drainage. Coburn (1982) considered the H. dorsalis species group to be mono- phyletic based on two moiphological characters. Mayden (1989) allied the H. dorsalis group with the H. amhlops species group. He removed the species of the H. dorsalis group from Notropis and placed them in a restricted Hyhopsis. Mayden (1989) excluded several species from Hyhopsis, referring them to Erimystax, Macrhyhopsis, Platygohio, and Extrariiis. Hyhopsis sensu Mayden (1989) also includes two additional species formerly placed in Notropis (H. alhorus and H. hifrenatiis). Herein, we include H. huccata as well, a species formerly placed in the monotypic genus Ericymha, based on the sister group relationship between H. huccata and other Hyhopsis sensu Mayden ( 1989). Unfortunately, Mayden (1989) was unable to resolve further the relationships between the H. dorsalis group and other Hyhopsis. Although considerable progress has been made toward resolving the relationship of the H. dorsalis group to other cyprinids, the phylogenetic relationships within the group remain problematic. Suttkus and Boschung (1990) described similarities in overall body shape of H. ammophilus and H. sahinae compared with the more elongate body of H. longirostris. Subsequently, Suttkus (1991) informally allied H. ammophilus with H. rafinesque based on the presence of a caudal spot on the base of the fin in living and freshly preserved specimens and allied these species with H. longirostris based on the presence of red and/or orange fin colors. The purpose of this paper is to test these hypotheses of relationship among members of the H. dorsalis group. RELATIONSHIPS AMONG HYBOPSIS DORSALIS GROUP 3 METHODS AND MATERIALS Fish for protein electrophoresis were collected by seining and were immediately frozen in liquid nitrogen, transported to the laboratory, and stored at -70°C. Liver, skeletal muscle, and brain were dissected and homogenized separately in a 1:1 (v:v) mixture of tissue and 0.01 M Tris- 0.001 M EDTA and 0.001 M mercaptoethanol, pH 6.8. Homogenates were centrifuged at 15,000x g for 10 min at 5°C. Within 72 h, the supernatant fractions were electrophoresed at 5°C on horizontal starch gels composed of 12% hydrolyzed potato starch. Thirty-three presumptive gene loci were visualized by histochemical staining. Enzyme nomenclature follows the recommendations of the International Union of Biochemistry Nomencla- ture Committee (1984) and locus nomenclature follows the recommenda- tions of Buth (1983). Enzymes, loci, tissue sources, and electrophoretic conditions are listed in Table 1. Electromorphs were coded a, b, c, etc.. in order of increasing anodal mobility. Allelic designations are relevant to this study only. Three morphological transformation series were included in the analysis of relationships, as follows. ( 1 ) The H. dorsalis group has a deep anterior notch in the mesethmoid-supraethmoid (coded b), whereas H. buccata, other Hybopsis, and related genera, lack the notch (coded a) (Cobum, 1982; Mayden, 1989). (2) The kinethmoid of the H. dorsalis group is elongate and straight (coded b), whereas in H. buccata and other related minnows, the bone is curved and not elongate (coded a) (Cobum, 1982; Mayden, 1989). (3) Xanthic fin color occurs in H. buccata, H. rafiuesque, H. ammophihis. and H. longirostris (coded a) but not in H. dorsalis or H. sabinae (coded b) (Suttkus, 1991; pers. obs.). Two other morphological characters were considered. Suttkus and Boschung (1991) reported that H. longirostris is more elongate than either H. ammophilus or H. sabinae: H. dorsalis and H. rafinesque are similar to H. ammophilus and H. sabinae (pers. obs.). However, H. buccata also is elongate, and because more exact quantification of body proportions was lacking, we excluded this character. Suttkus (1991) mentioned the presence of a caudal spot in living and freshly preserved H. rafinesque and H. ammophilus: the spot apparently is absent in H. longirostris. Although H. ammophilus and H. rafinesque can be distinguished from H. longirostris on the basis of this feature, caudal spots occur as polymorphisms in//, buccata and //. dorsalis (pers. obs.). Pending a more thorough examination of both the homology of caudal spots and the frequency of their occurrence in different species, we eliminated this character from the analysis. Phylogenetic analyses were performed using H. buccata as the outgroup. The two morphological characters reported by Coburn (1982) and dis- cussed above were used to constrain the analyses to consider only a UNIV. KANSAS MUS. NAT. HIST. OCC. PAP. No. 152 Table 1. Enzymes, loci. International Union of Biochemistry Nomenclature Committee numbers, tissue sources, and electrophoretic conditions for examination of the Hyhopsis dorsalis species group. Tissue Electrophoretic Enzyme lUBNC No. Locus source conditions* Aconitate hydratase 4.2.1.3 M-Acon-A Muscle A Adenosine deaminase 3.5.4.4 Ada-A Muscle B Adenylate kinase 2.7.4.3 Ak-A Muscle E Aspartate amino- 2.6.1.1 M-Aat-A Muscle E transferase Calcium binding Nonspecific Cbp-1 Muscle E proteins Cbp-2 Muscle E Creatine kinase 2.7.3.2 Ck-A Muscle E Ck-B Brain E Dipeptidase 3.4.13.11 Pep-A Muscle F Pep-B Muscle F Fructose bisphosphate 4.1.2.13 Fba-C Brain A aldolase Fumarate hydratase 4.2.1.2 Fum-A Muscle B General protein Nonspecific Gp-1 Muscle E Glucose-6-phosphate 1.1.1.49 G6pdh-B Liver C dehydrogenase Glucose-6-phosphate 5.3.1.9 Gpi-A Muscle A isomerase Gpi-B Muscle A Glyceraldehyde-3-phos- 1.2.1.12 Gapdh-A Muscle C phate dehydrogenase Gapdh-B Brain C Glycerol-3-phosphate 1.1.1.8 G3pdh-A Muscle B dehydrogenase Isocitrate dehydrogenase 1.1.1.42 M-Icdh-A Muscle B L-Lactate dehydrogenase 1.1.1.27 Ldh-A Muscle A Ldh-B Muscle A Ldh-C Liver A Malate dehydrogenase 1.1.1.37 M-Mdh-A Liver B (NAD+-dependent) S-Mdh-A Liver B S-Mdh-B Muscle B Malate dehydrogenase 1. 1 . 1 .40 S-Me-A Muscle B (NADP+-dependent) Phosphoglucomutase 5.4.2.2 Pgm-A Brain G Phosphogluconate 1.1.1.44 Pgdh-A Brain A dehydrogenase Superoxide dismutase 1.15.1.1 Sod-A Liver G Triose phosphate 5.3.1.1 Tpi-A Brain B isomerase Tpi-B' Brain B Tpi-B^ Brain B *A: Tris-citrate-NADP pH 7.0. 8 V/cm 10 h; B: Tris-citrate pH 8.0. 6.5 V/cm 10 h: C: Tris- borate-EDTA-NAD-NADP pH 9.1, 11 V/cm 14 h: D: Borate-NAD pH 8.8, 7 V/cm 1 1 h: E: Histidine-citrate-NADPpH7.0,5 V/cm 11 h; F: LiOH 14 V/cm lOh: G: Poulik-NADP. 10 V/ cm 10 h. RELATIONSHIPS AMONG HYBOPSIS DORSALIS GROUP 5 monophyletic //. c/<9/-5(3//5 group (Coburn, 1982; Mayden, 1989). Given this constraint and the relative lack of resolution between the H. dorsalis group and other Hyhopsis (Mayden. 1989). use of a single outgroup seemed justified; however, we realize that this decision limits the kinds of questions we can address. Two levels of phylogenetic analysis were performed. The first was a "qualitative" analysis using PAUP 3.0 (Swofford, 1990). The locus was considered the transformation series (= data column or "character"), and alleles were considered characters (= cell symbols or "character states"). Species with multiple electromorphs at a locus were coded for all alleles present. Analyses were performed using the "polymorphic" option rather than the "uncertain" option because more than one allele was known to exist. The polymorphic option counts the evolution from one fixed condi- tion to another fixed condition as two steps rather than one step. Loci with more than two electromorphs were analyzed unordered. Both DELTRAN and ACCTRAN optimizations were used. Because of the small number of taxa, the "exhaustive search" option was used; this option optimizes char- acters on all possible tree topologies. The second level of analysis was performed with FREQPARS (algo- rithm in Swofford and Berlocher, 1987; program by David Swofford, Illinois Natural History Survey, Urbana-Champaign. Illinois) using a ma- trix of allele frequencies and one morphological character (xanthic color). It was treated like an electromorph with a frequency of 1.0. The topologies analyzed with FREQPARS were those derived from the PAUP analysis and one additional, contrived topology that treated the eastern and western populations of H. longiwstris as a paraphyletic group. Specimens examined. — All material examined is deposited in the ich- thyological collection of the Museum of Natural History of The University of Kansas and is listed below under three categories — electrophoresis, color pattern, and osteology. All osteological specimens are dry prepara- tions following the technique of Mayden and Wiley ( 1984) unless specified "C&S," in which case they were cleared and stained following the tech- nique of Dingerkus and Uhler (1977). State, drainage, and catalogue num- bers are given for electrophoretic and color pattern specimens; state and catalogue number are given for osteological specimens. Parenthetical num- bers indicate the number of specimens examined for each category. Vouch- ers for the electrophoretic analysis include the remains of actual specimens (number of specimens indicated) and whole preserved specimens. Com- plete locality data are available on request from E. O. Wiley. Electrophoresis: Hybopsis ammophilus. — Alabama: Alabama River Dr.: 22935 (6): Tombigbee River Dr.: 22936 (6), 22937 (9). H. biiccata. — Alabama: Yellow River Dr.: 22927 (8); Florida: Escambia River Dr.: 22928 (6); Mississippi: Pascagoula River Dr.: 22929 (3), 22930 (4). 6 UNIV. KANSAS MUS. NAT. HIST. OCC. PAP. No. 1.^2 H. ilorsalis.— Kansas: Kansas River Dr.: 22931 (13), 22932 (11). H. longirostris. — Alabama: Escatawpa River Dr.: 22938 (32); Perdido River Dr.: 22939 (25). H. rafinesque. — Mississippi: Yazoo River Dr.: 22341 (20). H. sahiiiae. —LoinsiANA: Sabine River Dr.: 22933 (7). 22934 (3). 22322 (9). Color pattern : H. ammophUus. — Alabama: Tombigbee River Dr.: 14340 (13). H. huccata. — Mississippi: Pascagoula River Dr.: 19673 (14). H. dorsaUs. — Nebraska: Platte River Dr.: 4846 (39); Kansas: Kansas River Dr.: 22436(69), 16237(135). H. longirostris. — Alabama: Escambia River Dr.: 17389 (24); Florida: Yellow River Dr.: 17394 (39); Louisiana: Amite River Dr.: 17946 (13); Mississippi: Jordan River Dr.: 16871 (100). H. rafinesque. — Mississippi: Yazoo River Dr.: 22541 (16). H. sabinae. — Louisiana: Sabine River Dr.: 19673 (14). Osteology: H. ammophilus. — Alabama: 14540 (5). H. huccata. — Alabama: 17764(5). H. ciorsalis.— Kansas: 22091 (5). H. longirostris. — Alabama: 17589 (4), 16871 (5). H. rafinesque. — Mississippi: 22541 (5 C&S). H. sabinae. — Louisiana: 6237 (5). RESULTS Presumptive loci and alleles. — Of the 33 presumptive loci examined, 13 were fixed for all species and 20 varied among and/or within species. The following loci were fixed: M- Aeon- A, Ak-A. Cbp-2, Fba-C, Gp-1, G6pdh-B, Gapdh-B, MTcdh-A, Ldh-C, S-Mdh-B. S-Me-A, Tpi-B', and Tpi-B-. Numbers of each genotype for all variable loci are listed in Table 2. The samples representing eastern and western populations of//, longirostris showed fixed differences and were separated in the phylogenetic analyses. Qualitative data analysis. — The genetic data (Table 2) and morpho- logical data were combined to produce a qualitative data matrix (Table 3). PAUP analysis resulted in three equally parsimonious trees (Fig. la-c). These trees support the following hypotheses of relationship: ( 1 ) //. dorsalis is the sister of all remaining members of the group; (2) H. sabinae is the sister of the remaining species; and (3) among the remaining four forms, the two populations of //. longirostris form a sister pair. The relationships among //. longirostris. H. rafinesque, and H. ammophilus were not re- solved. Quantitative data analysis. — The three trees resolved in the PAUP analysis and a tree treating the two populations of H. longirostris as paraphyletic were used in a FREQPARS analysis. These tree topologies RELATIONSHIPS AMONG HYBOPSIS DORSALIS GROUP 7 and the results expressed as path lengths and tree lengths are shown in Figure 1. The analysis favors the hypothesis that H. ammophilus is the sister species of//, longirostris (Fig. la). Character evolution as interpreted by the FREQPARS analysis can be summarized in two ways. The FREQPARS interpretation of allelic and morphological character distribu- tions among species is shown in Table 4. A qualitative representation of the origin of apomorphic allelic and morphological characters is depicted in Figure 2. The eight polymorphic alleles at internodes are carried through at least one speciation event. There are 14 terminal polymorphisms. Ho- moplasies are rare (three loci and one morphological character). Two electromorphs show parallel acquisition (Tpi-A'' and Pgdh"*) and one "re- verses" to the plesiomoiphic mobility (Gpi-B'). In such cases, there are alternate, equally parsimonious interpretations of character evolution. For example, an alternate explanation for the hypothesis that Gpi-B electromorphs a and b are synapomorphic at the intemode between Nodes 11 and 12 is that electromorph Gpi-B'' is synapomorphic at this intemode and that Gpi-B^ is autapomorphic for H. sahinae. Interestingly, //. biiccata showed parallel acquisition of xanthic fin colors with H. ammophilus, H. sahinae, and H. loniiirostris. DISCUSSION The analysis was constrained because ( 1 ) the relationship between the H. dorsalis group and its closest relative is unresolved and (2) only a single outgroup was used in the analysis. Thus, we could not test the monophyly of the group using additional characters derived from our electrophoretic data. However, electromorphs at four loci that might provide a test of the monophyly of the group are shown at the bottom of Figure 2 on the unrooted intemode between //. huccata and the //. dorsalis group. The hypothesis that //. sahinae is more closely related to //. longirostris, H. ammophilus, and //. rafinesque is supported by the greater number of derived electromorphs on this stem relative to other branches on all trees (Figs. 1, 2). Similarly, the hypothesis that //. ammophilus. H. rafinesque, and H. longirostris form a monophyletic group is well supported, as indicated by the second longest branch in all topologies (Fig. 1). Of the three equally parsimonious hypotheses of relationship shown in the PAUP analysis (Fig. la-c), the shortest FREQPARS tree (71.4 steps) supports the hypothesis that //. ammophilus is the sister species of//, longirostris (Fig. la). A sister-group relationship between //. rafinesque and H. ammophilus (Suttkus, 1991 ) would require a tree of 73.4 steps. In addition, the branch length supporting these two taxa is zero (Fig. lb). The tree showing H. rafinesque as the sister species of//, longirostris is slightly shorter than the tree supporting the hypothesis of a sister-group relationship between H. UNIV. KANSAS MUS. NAT. HIST. OCC. PAP. No. 152 C: *_, uo m U"j U^i ir^i >/") tr-, ir-; u-j m rt « o Xi 03 CJ C3 S' cd So ""^ c3 X) « « o Xi c« o c« JO W c o in a; C O 3 D. O n. T3 C cS in 'o 1> D. on in ?> in t3 o rsi (N (^1 ri ri ri n ri ri I^ r«-, rj fN en m r«-, r<", ro r<"i r^, r<~, <-<-, r«-. 0 3 O o 73 X) X- -a -a in r3 C3 X X 13 C3 73 X ~ in 73 T3 r3 r- (N On \o r~ m T3 « £" XI -O ■O 73 rt Xj "O X3 Xi fN (N 73 73 X> X3 n Xi i^ ■* T3 x> •^ 1 < ^ ' O X) •O Xi X) ^ « rt X) X) X) « U-2-0 oxx ri X CO O X T3 C^ X Xi CO 5 X rt CO Xi U CJ Jii Xi CO CO X) O "O O X T3 X X Xi CO CO O ^ CO e^ O C^ Xl X X u O X o X ,X X X X> cox CO OT3 Cj^O X Xi CO X o x: o o e OJ CO X X> X X X) CO X X O X X) o o -o -o U X X; O O X CO c:. ~ vj. CO CO CO CO 1 4— t < ^^ CO < CO 1 < x: s •o Xi Ji U < u u < < oa ;- Q. D. .i ..1. ^ ^ S 5: — "O '0 E '0 < -2 < < < ^ OJ a. D. T3 TD ._ C !U OJ D/J 5j 0 D. s • — D- a. D- a. C/5 h- ^ c CO X RELATIONSHIPS AMONG HYBOPSIS DORSALIS GROUP 11 si 'i^.- "^.'^^v o -o >. ■ — ' (U c 00 r~* r3 Qi ^ ^ .£ c o _2 _o Di [1- 3: a. 5 '-^« aj — ;— ejj C3 3 E