Content uploaded by Naercio A. Menezes
Author content
All content in this area was uploaded by Naercio A. Menezes on Aug 28, 2023
Content may be subject to copyright.
© e Author(s) 2023. Published by Oxford University Press on behalf of e Linnean Society of London. All rights reserved. For permissions, please e-mail: journals.permissions@
oup.com
Received 27 April 2022; revised 30 January 2023; accepted 6 March 2023
Original Article
Revision of the generic classication of pike cichlids using
an integrative phylogenetic approach (Cichlidae: tribe
Geophagini: subtribe Crenicichlina)
Henrique R.Varel la1,2,4,*, Sven O.Kullander3, Naércio A.Menezes1, ClaudioOliveira2,
HernánLópez-Fernández4,
1Museu de Zoologia da Universidade de São Paulo, São Paulo, SP, Brazil
2Laboratório de Biologia e Genética de Peixes, Departamento de Biologia Estrutural e Funcional, Instituto de Biociências, Universidade Estadual
Paulista, IBB/UNESP, Botucatu, SP, Brazil
3Museum of Natural History, Stockholm, Sweden
4Department of Ecology and Evolutionary Biology and Museum of Zoology, University of Michigan, Ann Arbor, MI, USA
*Corresponding author. E-mail: hrvarella@gmail.com
ABSTRACT
Pike cichlids form the largest clade of Neotropical cichlids, with over 100 species presently classied in two genera: Crenicichla (93 species wide-
spread in rivers of South America east of the Andes) and Teleocichla (nine rheophilic Amazonian species). Here, we combined a new dataset of
216 morphological characters with molecular data compiled from published sources, comprising 74 terminal taxa of pike cichlids (68 out of 102
valid species, plus four putative new species), and performed phylogenetic analyses using maximum likelihood, Bayesian inference, and parsi-
mony. Based on a synthesis of our results and previous phylogenies, we propose a new classication in which the clade including all pike cichlids
is elevated to the rank of subtribe (Crenicichlina) and the genus Crenicichla is redened, including three subgenera: Crenicichla (monotypic
with the type species), Batrachops (resurected as subgenus), and Lacustria (new subgenus). Teleocichla is maintained as a valid genus and four
new genera are proposed: Wallaciia, Saxatilia, Hemeraia, and Lugubria. Our results on character mapping support the hypothesis that resource
partitioning in environments with fast-owing water and rocky beds might have played a role in the origin or maintenance of the great diversity
of pike cichlids, resulting in parallel evolution of similar ecomorphs.
Keywords: taxonomy; cichlids; morphology; molecular data; parsimony; maximum likelihood; Bayesian inference; continuous characters;
extended implied weighting; character evolution
INTRODUCTION
With 1732 valid species (Fricke et al. 2022), Cichlidae are the
most species-rich family of freshwater euteleosts (Smith et al.
2008) and a well-known example of exceptional lineage and
phenotype diversication among vertebrates (Near et al. 2013;
Rabosky et al. 2013). Cichlids are also recognized as excellent
models for the investigation of macroevolutionary processes
(Gante and Salzburger 2012), particularly adaptive radiation
(e.g. Kocher 2004; Young et al. 2009; Wagner et al. 2012; López-
Fernández et al. 2013; McGee et al. 2020). However, evolu-
tionary studies have been much more focused on the African
lineages than on the Neotropical cichlid subfamily Cichlinae
(sensu Farias et al. 2000; Smith et al. 2008; López-Fernández et
al. 2010; Friedman et al. 2013; McMahan et al. 2013; Ilves et al.
2018).
e sh called pike cichlids (English; scientic and
aquarium literature in general), jacundás, joanas, micholas,
mataguaros, or cabezas-amargas (Latin America) form a clade
comprising over 100 species (20% of the taxon diversity of
Neotropical cichlids). An updated list with all nominal and
valid species of pike cichlids, with authorities, is presented in
Table 1.
ese species are presently placed in two genera: Crenicichla
Heckel, 1840 and Teleocichla Kullander, 1988. Crenicichla en-
compasses 93 valid species (Kullander 2003; Fricke et al. 2022)
distributed in almost all major drainages of South America east
Zoological Journal of the Linnean Society, 2023, XX, 1–43
hps://doi.org/10.1093/zoolinnean/zlad021
Advance access publication 21 June 2023
Original Article
Downloaded from https://academic.oup.com/zoolinnean/advance-article/doi/10.1093/zoolinnean/zlad021/7204395 by Universidade de Sao Paulo (USP) user on 27 July 2023
2 • Var e l l a et al.
Table 1. Synonym list of species of pike cichlids, with nominal species divided by the previous classication in the genera Crenicichla and
Teleocichla, but with valid species classied into the new classication. Asterisks denote valid species not included in the analyses (check
Material and Methods section for explanation)
Nominal species (previous
genus Crenicichla)
Author(s) of nominal species Valid genus (subgenus) specic epithet
1Crenicichla acutirostris Günther 1862 1 Lugubria acutirostris
2Crenicichla adspersa Heckel 1840 2 Lugubria adspersa **
3Crenicichla funebris Heckel 1840
4Crenicichla saxatilis albopunctata Pellegrin 1904 3 Saxatilia albopunctata ***
5Crenicichla alta Eigenmann 1912 4 Saxatilia alta
6Crenicichla vaillanti Pellegrin 1903
7Crenicichla pterogramma Fowler 1914
8Crenicichla cardiostigma Ploeg 1991
9Crenicichla anamiri Ito and Py-Daniel 2015 5 Wallaciia anamiri *
10 Crenicichla anthurus Cope 1872 6 Saxatilia anthurus
11 Perca brasiliensis Bloch 1792 7 Saxatilia brasiliensis ***
12 Sparusnhoquunda La Cepède 1802
13 Crenicichla menezesi Ploeg 1991
14 Crenicichla britskii Kullander 1982 8 Saxatilia britskii
15 Crenicichla cametana Steindachner 1911 9 Crenicichla (Batrachops) cametana
16 Crenicichla astroblepa Ploeg 1986
17 Crenicichla celidochilus Cascioa 1987 10 Crenicichla (Lacustria) celidochilus
18 Crenicichla chicha Varella, Kullander and Lima 2012 11 Hemeraia chicha
19 Crenicichla brasiliensis fasciata Pellegrin 1904 12 Lugubria cincta *
20 Crenicichla cincta Regan 1905
21 Crenicichla compressiceps Ploeg 1986 13 Wallaciia compressiceps
22 Crenicichla coppenamensis Ploeg 1987 14 Saxatilia coppenamensis ***
23 Crenicichla cyanonotus Cope 1870 15 Crenicichla (Batrachops) cyanonotus *
24 Crenicichla cyclostoma Ploeg 1986 16 Crenicichla (Batrachops) cyclostoma
25 Crenicichla dandara Varella and Ito 2018 17 Lugubria dandara *
26 Crenicichla empheres Lucena 2007 18 Crenicichla (Lacustria) empheres
27 Crenicichla enata Gill 1858 19 Saxatilia enata ***
28 Crenicichla gaucho Lucena and Kullander 1992 20 Crenicichla (Lacustria) gaucho
29 Crenicichla geayi Pellegrin 1903 21 Crenicichla (Batrachops) geayi
30 Crenicichla gillmorlisi Kullander and Lucena 2013 22 Crenicichla (Lacustria) gillmorlisi *
31 Crenicichla hadrostigma Lucena 2007 23 Crenicichla (Lacustria) hadrostigma
32 Crenicichla haroldoi Luengo and Britski 1974 24 Crenicichla (Lacustria) haroldoi
33 Crenicichla heckeli Ploeg 1989 25 Wallaciia heckeli
34 Crenicichla hemera Kullander 1990b 26 Hemeraia hemera
35 Crenicichla guentheri Ploeg 1991
36 Crenicichla hu Piálek, et al. 2010 27 Crenicichla (Lacustria) hu
37 Crenicichla hummelincki Ploeg 1991 28 Saxatilia hummelincki ***
Downloaded from https://academic.oup.com/zoolinnean/advance-article/doi/10.1093/zoolinnean/zlad021/7204395 by Universidade de Sao Paulo (USP) user on 27 July 2023
New generic classication of pike cichlids • 3
Nominal species (previous
genus Crenicichla)
Author(s) of nominal species Valid genus (subgenus) specic epithet
38 Crenicichla igara Lucena and Kullander 1992 29 Crenicichla (Lacustria) igara
39 Crenicichla iguapina Kullander and Lucena 2006 30 Crenicichla (Lacustria) iguapina
40 Crenicichla iguassuensis Haseman 1911 31 Crenicichla (Lacustria) iguassuensis
41 Crenicichla inpa Ploeg 1991 32 Saxatilia inpa
42 Crenicichla isbrueckeri Ploeg 1991 33 Saxatilia isbrueckeri ***
43 Crenicichla jaguarensis Haseman 1911 34 Crenicichla (Lacustria) jaguarensis
44 Crenicichla jegui Ploeg 1986 35 Crenicichla (Batrachops) jegui
45 Crenicichla johanna Heckel 1840 36 Lugubria johanna
46 Cychla fasciata Jardine 1843
47 Crenicichla obtusirostris Günther 1862
48 Crenicichla johanna carsevennensis Pellegrin 1905
49 Crenicichla jupiaensis Britski and Luengo 1968 37 Crenicichla (Lacustria) jupiaensis
50 Crenicichla jurubi Lucena and Kullander 1992 38 Crenicichla (Lacustria) jurubi
51 Cychla labrina Spix and Agassiz 1831 39 Saxatilia labrina
52 Cycla lacustris Castelnau 1855 40 Crenicichla (Lacustria) lacustris
53 Crenicichla dorsocellata Haseman 1911
54 Crenicichla biocellata Ihering 1914
55 Crenicichla lenticulata Heckel 1840 41 Lugubria lenticulata
56 Crenicichla ornata Regan 1905
57 Crenicichla lepidota Heckel 1840 42 Saxatilia lepidota
58 Crenicichla edithae Ploeg 1991
59 Crenicichla lucenai Maos et al. 2014 43 Crenicichla (Lacustria) lucenai **
60 Crenicichla lucius Cope 1870 44 Saxatilia lucius ***
61 Crenicichla lugubris Heckel 1840 45 Lugubria lugubris
62 Cychla rutilans Jardine 1843
63 Crenicichla macrophthalma Heckel 1840 46 Crenicichla (Crenicichla) macrophthalma
64 Crenicichla santaremensis Haseman 1911
65 Crenicichla maculata Kullander and Lucena 2006 47 Crenicichla (Lacustria) maculata **
66 Crenicichla mandelburgeri Kullander 2009 48 Crenicichla (Lacustria) mandelburgeri
67 Crenicichla brasiliensis marmorata Pellegrin 1904 49 Lugubria marmorata **
68 Crenicichla minuano Lucena and Kullander 1992 50 Crenicichla (Lacustria) minuano
69 Crenicichla missioneira Lucena and Kullander 1992 51 Crenicichla (Lacustria) missioneira
70 Crenicichla monicae Kullander and Varella 2015 52 Lugubria monicae *
71 Crenicichla mucuryna Ihering 1914 53 Crenicichla (Lacustria) mucuryna *
72 Crenicichla multispinosa Pellegrin 1903 54 Lugubria multispinosa
73 Crenicichla nickeriensis Ploeg 1987 55 Saxatilia nickeriensis ***
74 Crenicichla notophthalmus Regan 1913 56 Wallaciia notophthalmus
75 Crenicichla pellegrini Ploeg 1991 57 Saxatilia pellegrini ***
Table 1. Continued
Downloaded from https://academic.oup.com/zoolinnean/advance-article/doi/10.1093/zoolinnean/zlad021/7204395 by Universidade de Sao Paulo (USP) user on 27 July 2023
4 • Var e l l a et al.
Nominal species (previous
genus Crenicichla)
Author(s) of nominal species Valid genus (subgenus) specic epithet
76 Crenicichla percna Kullander 1991 58 Lugubria percna **
77 Crenicichla phaiospilus Kullander 1991 59 Lugubria phaiospilus
78 Crenicichla ploegi Varella et al. 2018 60 Saxatilia ploegi
79 Crenicichla prenda Lucena and Kullander 1992 61 Crenicichla (Lacustria) prenda
80 Crenicichla proteus Cope 1872 62 Saxatilia proteus ***
81 Crenicichla proteus argynnis Cope 1872
82 Batrachops nemopterus Fowler 1940
83 Crenicichla nijsseni Ploeg 1991
84 Crenicichla punctata Hensel 1870 63 Crenicichla (Lacustria) punctata
85 Crenicichla polysticta Hensel 1870
86 Crenicichla pydanielae Ploeg 1991 64 Saxatilia pydanielae ***
87 Crenicichla regani Ploeg 1989 65 Wallaciia regani
88 Batrachops reticulatus Heckel 1840 66 Crenicichla (Batrachops) reticulata
89 Crenicichla elegans Steindachner 1881
90 Batrachops punctulatus Regan 1905
91 Crenicichla rosemariae Kullander 1997 67 Lugubria rosemariae *
92 Crenicichla santosi Ploeg 1991 68 Saxatilia santosi ***
93 Sparus saxatilis Linnaeus 1758 69 Saxatilia saxatilis
94 Scarus biocellatus Walbaum 1792
95 Scarus Pavo La Cepède 1802
96 Scarus pavoninus Gray 1854
97 Batrachops scoii Eigenmann 1907 70 Crenicichla (Lacustria) scoii
98 Crenicichla (Batrachops)
multidens
Steindachner 1915
99 Labrus amarus Larrañaga 1923
100 Crenicichla lacustris semifasciata Devincenzi 1939
101 Crenicichla sedentaria Kullander 1986 71 Crenicichla (Batrachops) sedentaria *
102 Crenicichla saxatilis semicincta Steindachner 1892 72 Saxatilia semicincta ***
103 Crenicichla clancularia Ploeg 1991
104 Batrachops semifasciatus Heckel 1840 73 Crenicichla (Batrachops) semifasciata
105 Acharnes chacoensis Holmberg 1891
106 Boggiana ocellata Perugia 1897
107 Crenicichla simoni Haseman 1911
108 Crenicichla sipaliwini Ploeg 1987 74 Saxatilia sipaliwini ***
109 Crenicichla stocki Ploeg 1991 75 Crenicichla (Batrachops) stocki
110 Crenicichla johanna strigata Günther 1862 76 Lugubria strigata
111 Crenicichla sveni Ploeg 1991 77 Saxatilia sveni ***
112 Crenicichla taikyra Cascioa et al. 2013 78 Crenicichla (Lacustria) taikyra
Table 1. Continued
Downloaded from https://academic.oup.com/zoolinnean/advance-article/doi/10.1093/zoolinnean/zlad021/7204395 by Universidade de Sao Paulo (USP) user on 27 July 2023
New generic classication of pike cichlids • 5
of the Andes, from coastal drainages of Venezuela, Trinidad and
Tobago, Guianas and Suriname in northern South America to
the Río de La Plata in Argentina and Uruguay in the southern
edge of its distribution. e nine valid species of Teleocichla are
restricted to clear-water, fast-owing tributaries of the Amazon
River basin (Vare ll a et al. 2016).
e Crenicichla–Teleocichla clade is also ecologically di-
verse [see Říčan et al. (2021a: 2/29) for general information
on ecology]. Recent studies on the paerns of phenotypic di-
versication of Neotropical cichlids using function-driven
ecomorphology (Feilich and López-Fernández 2019) show
that a vast portion of the functional morphospace of the
Nominal species (previous
genus Crenicichla)
Author(s) of nominal species Valid genus (subgenus) specic epithet
113 Crenicichla tapii Piálek et al. 2015 79 Crenicichla (Lacustria) tapii
114 Crenicichla tendybaguassu Lucena and Kullander 1992 80 Crenicichla (Lacustria) tendybaguassu
115 Crenicichla ternetzi Norman 1926 81 Lugubria ternetzi *
116 Crenicichla tesay Cascioa and Almirón 2009 82 Crenicichla (Lacustria) tesay
117 Crenicichla tigrina Ploeg, Jégu and Ferreira 1991 83 Lugubria tigrina
118 Crenicichla tingui Kullander and Lucena 2006 84 Crenicichla (Lacustria) tingui*
119 Crenicichla tuca Piálek et al. 2015 85 Crenicichla (Lacustria) tuca
120 Crenicichla urosema Kullander 1990a 86 Wallaciia urosema
121 Crenicichla virgatula Ploeg 1991 87 Wallaciia virgatula *
122 Crenicichla viata Heckel 1840 88 Crenicichla (Lacustria) viata
123 Crenicichla wallacii Regan 1905 89 Wallaciia wallacii
124 Crenicichla nanus Regan 1913
125 Crenicichla yaha Cascioa, Almirón and Gómez 2006 90 Crenicichla (Lacustria) yaha
126 Crenicichla yjhui Piálek et al. 2019b 91 Crenicichla (Lacustria) yjhui
127 Crenicichla ypo Cascioa et al. 2010 92 Crenicichla (Lacustria) ypo
128 Crenicichla zebrina Montaña, López-Fernández and Taphorn
2008
93 Crenicichla (Lacustria) zebrina *
129 Acharnes niederleini Holmberg 1891 Nomen dubium of the subgenus Crenicichla (Lacustria)
from Río Uruguay (Varella 2011)
130 Cycla conibos Castelnau 1855 Nomen dubium of Crenicichla (Kullander 1986), not
deeply investigated herein
131 Cycla multifasciata Castelnau 1855 Nomen dubium of Crenicichla (Kullander 1986), not
deeply investigated herein
132 Scarus rufescens Walbaum 1792 Indicated as Crenicichla in Fricke et al. (2022), not
deeply investigated herein
133 Sparus subbruneus Walbaum 1792
Nominal species (previous
genus Teleocichla)
Author(s) of nominal species Valid genus (subgenus) specic epithet
134 Teleocichla centisquama Zuanon and Sazima 2002 94 Teleocichla centisquama
135 Teleocichla centrarchus Kullander 1988 95 Teleocichla centrarchus
136 Teleocichla cinderella Kullander 1988 96 Teleocichla cinderella
137 Teleocichla gephyrogramma Kullander 1988 97 Teleocichla gephyrogramma
138 Teleocichla monogramma Kullander 1988 98 Teleocichla monogramma
139 Teleocichla preta Var ella et al. 2016 99 Teleocichla preta
140 Teleocichla prionogenys Kullander 1988 100 Teleocichla prionogenys
141 Teleocichla proselytus Kullander 1988 101 Teleocichla proselytus
142 Teleocichla wajapi Varella and Moreira 2013 102 Teleocichla wajapi
Table 1. Continued
Downloaded from https://academic.oup.com/zoolinnean/advance-article/doi/10.1093/zoolinnean/zlad021/7204395 by Universidade de Sao Paulo (USP) user on 27 July 2023
6 • Var e l l a et al.
Neotropical cichlids is occupied by representatives of the
Crenicichla–Teleocichla clade (Arbour and López-Fernández
2013; Astudillo-Clavijo et al. 2015). Pike cichlids exhibit
ecomorphological characteristics related to feeding biomech-
anics nearly unique among South American Neotropical cichlids
(Arbour and López-Fernández 2013) and the early evolution of
a novel foraging strategy (fast-burst predators; see: Astudillo-
Clavijo et al. 2015) may have allowed them to rapidly diversify
within the predatory niche, preventing other cichlid lineages to
occupy similar roles. Teleocichla species exihibit morphological
specialization toward rheophily (e.g. Kullander 1988; Lujan and
Conway 2015; Va rel la et al. 2016) and are considered in the most
recent assessments of the International Union for Conservation
of Nature (IUCN) to be at some level of extinction threat mainly
due to the impacts of hydroelectric development in Amazonian
rivers.
Early phylogenetic analyses based exclusively on morphology
placed Crenicichla s.l. as the sister-group to Cichla Bloch and
Schneider, 1801 (e.g. Stiassny 1991; Cascioa and Arratia 1993a,
b; Kullander 1998). However, subsequent morphological and
total-evidence analyses showed that Crenicichla, together with
Teleocichla (when included), belongs to the tribe Geophagini
(e.g. Farias et al. 1999; Sparks and Smith 2004; López-Fernández
et al. 2005a, b, 2010, 2012; Smith et al. 2008; McMahan 2013),
albeit with varying placements within the tribe. More recently,
a densely sampled study compiling 19 molecular loci available
from GenBank (Burress et al. 2017), and a phylogenomic ana-
lysis based on hundreds of exons with genus-level sampling of
Neotropical cichlids (Ilves et al. 2018), corroborated the place-
ment of pike cichlids within Geophagini and as sister-group to a
clade composed of Acarichthys Eigenmann, 1912 and Biotoecus
Steindachner, 1875.
While some studies included morphological characters and
used a total-evidence approach (López-Fernández et al. 2005a,
2012), no phenotypic dataset exists for phylogenetic analysis
with comparable taxonomic sampling to studies based on mo-
lecular data. Considering the importance of morphology for
resolving phylogenetic relationships among living and fossil
taxa, and to provide taxonomic diagnoses, the development of
morphological datasets that complement existing molecular
data represents a logical continuation to the recent advances
in phylogenetic understanding of dierent groups of cichlids.
Moreover, a combined phylogeny including morphological data
can provide an independent test of the phylogenies based exclu-
sively on DNA data and can shed light on transformations that
characterize organism evolution.
Pre-phylogenetic taxonomy of pike cichlids
In this section, we provide a summary of the taxonomy of pike
cichlids before the development of the rst phylogenies. For
an extended and commented version, accompanied with an il-
lustrative scheme emphasizing supra-specic classication, see
the Supporting Information, Appendix S1. Since the descrip-
tion of Crenicichla and Batrachops by Heckel (1840) to classify
some elongate, small-scaled South American cichlids, the debate
about the composition, validity, and diagnostic characters of
these genera remained predominant in the literature until 1911
(in chronological order: Günther 1862; Eigenmann and Bray
1894; Perugia 1897; Pellegrin 1903, 1904; Regan 1905, 1913;
Haseman 1911). Aer this early period of taxonomic debate,
the number of species in the genus increased slowly and without
critical analysis on classication (e.g. Ihering 1914; Steindachner
1915; Devincenzi 1939; Fowler 1940; Britski and Luengo 1968;
Luengo and Britski 1974).
Over the last four decades, however, considering the numerous
species that were already included in the genus Crenicichla, sev-
eral studies have aempted to subdivide the genus based on
morphological gaps to facilitate taxonomic comparisons (e.g.
Kullander 1981, 1982, 1988, 1990a, 1991, 1997; Ploeg 1991;
Lucena and Kullander 1992; Kullander and Lucena 2006).
Additionally, during this period, the genus Teleocichla was pro-
posed to comprise small rheophilic species considered closely
related to Crenicichla based on discussion of morphological
characters (Kullander 1988). is period also includes the only
global revision of Crenicichla s.l., performed by Ploeg (1991),
covering 72 valid species, with 15 of them described in that
study. Ploeg’s classication recognized most groups previously
proposed in Kullander’s studies but synonymized Teleocichla
under Crenicichla, considering it a monophyletic subgroup of the
Crenicichla wallacii group. However, Ploeg continued to use in-
formal species groups instead of revising the classication using
formal categories such as genera or subgenera.
Aer Ploeg (1991), various taxonomic revisions were per-
formed at regional scales. ey uncovered the diversity of pike
cichlids in the Uruguay river basin (Lucena and Kullander 1992;
Lucena 2007) and in the coastal drainages of Brazil (Kullander
and Lucena 2006). More recently, a series of papers have ex-
plored more deeply the diversity in the tributaries of the Río
Paraná below the Itaipu dam, including the Río Iguaçu, and
described several species of the Crenicichla lacustris group (e.g.
Varella 2011; Piálek et al. 2012, 2015, 2019b; Cascioa et al.
2013). Given all those contributions, the Crenicichla lacustris
group went from nine valid species in Ploeg (1991) to 32, and
more species are waiting to be formally described (see: Vare ll a
2011). As interest in the taxonomy was accompanied with the
development of phylogenetic hypotheses (based exclusively
on molecular data), the Crenicichla lacustris group became the
best-known group within Crenicichla, both taxonomically and
phylogenetically (see next section).
From a historical perspective, a change in diagnostic charac-
ters used through revisionary taxonomic work can be detected.
Keys for identication in classical papers (e.g. Regan 1905, 1913;
Haseman 1911) repeatedly relied on features of general body
shape and a few characters of the jaws. However, from Kullander
(1981) forward, colour paern features were largely used to diag-
nose species groups. More recently, with the improvement of
many sh collections worldwide and eld documentation, char-
acteristics of coloration that can vary sexually, ontogenetically,
or under dierent stages of preservation have also been explored
and used (e.g. Piálek et al. 2010, 2015; Cascioa et al. 2013;
Varella et al. 2018). Contrastingly, internal morphological char-
acters such as osteology or so tissue and organs (brain, muscles,
etc.) have not been extensively explored within Crenicichla, with
the exception of Kullander (1988). Beyond a few osteological
comparisons (Kullander 1990b, 1991; Ploeg 1991; Varel la et al.
2012, 2016, 2018), mostly limited to infraorbitals, internal mor-
phological analyses were limited to guring and measuring the
Downloaded from https://academic.oup.com/zoolinnean/advance-article/doi/10.1093/zoolinnean/zlad021/7204395 by Universidade de Sao Paulo (USP) user on 27 July 2023
New generic classication of pike cichlids • 7
lower pharyngeal jaw (LPJ), counting vertebrae and gill rakers,
or checking for microbranchiospines on gill arches. Altogether,
osteological characters of pike cichlids have only been par-
tially explored, with existing work based on limited taxon
sampling aimed at taxon diagnosis (e.g. Kullander 1988) or
establishing higher level relationships between Crenicichla and
other Neotropical genera (e.g. Cichocki 1976; Stiassny 1991;
Cascioa and Arratia 1993a; Kullander 1998, López-Fernández
et al. 2005a).
Phylogenetic relationships of pike cichlids
Ploeg (1991) performed the only species-level morpho-
logical phylogenetic study of the pike cichlids available to date.
However, given the unconventional methodology, Ploeg’s re-
sults cannot be formally considered to be a phylogenetic analysis
and are of limited use in clarifying evolutionary relationships
within Crenicichla. First, althought Ploeg proposed 32 morpho-
logical characters, he selected a priori the ‘suitable’ or ‘useful’
characters to reconstruct intergroup relationships (Ploeg 1991:
gs 169–170) and to infer relationships among species within
each group (Ploeg: 1991: 171–175), using only 5–10 of the 32
characters to resolve each group. It is not clear which were the
outgroups used to polarize the character states in any of the re-
constructions, because he used supra-especic outgroups to
polarize species-level reconstructions (e.g. C. lugubris group
as outgroup for the species-level phylogeny of the C. saxatilis
group). Moreover, monophyly of each proposed group was xed
a priori, thus preventing formal tests of clade composition de-
rived from a global species-level analysis. Finally, and most im-
portantly, the data matrices on which his analyses were based
were not provided in the dissertation or elsewhere. Altogether,
the analyses performed by Ploeg (1991) cannot be replicated,
reviewed, or amended, and thus must be taken with caution.
Aer Ploeg (1991), understanding of pike-cichlids’ phyl-
ogeny remained static for almost two decades. e genus-level,
multilocus phylogeny of Neotropical cichlids by López-Fernández
et al. (2010) included eight species of Crenicichla and addressed
only marginally the relationships among pike cichlids. Kullander
et al. (2010) developed the rst true Crenicichla-focused analysis
with a molecular phylogeny based on the mitochondrial gene
cytochrome b (60 sequences representing 46 haplotypes and 21
species). is study was criticized in a subsequent paper (Piálek
et al. 2012: 55) for its insucient taxon sampling (e.g. absence
of species of the Crenicichla lugubris group and inclusion of very
few northern taxa) and genetic coverage (single gene marker),
inappropriate data partition scheme, and for using only Cichla
as outgroup, i.e. did not use any of the closely related geophagini
taxa identied in previous morphological and molecular data
hypotheses as closer relatives of Crenicichla than Cichla (e.g.
Farias 2000, López-Fernandéz et al. 2005a,b; Smith et al. 2008).
However, Kullander et al. (2010) primarily provided phylo-
genetic support for the monophyly of most species groups of
Crenicichla (e.g. C. saxatilis, C. wallacii, and C. reticulata groups)
and of Teleocichla. Although unresolved in a Bayesian analysis,
their parsimony topology resulted in a monophyletic Crenicichla
lacustris group, which included Crenicichla scoii [previously
placed in the Crenicichla reticulata group by Ploeg (1991)] and
Crenicichla viata (so far considered a member of the Crenicichla
lugubris group). Signicantly, Kullander et al. (2010) rst intro-
duced the idea that the Crenicichla missioneira group endemic of
the Río Uruguay basin should be considered a species ock and
provided early evidence of hybridization within Crenicichla.
Piálek et al. (2012) performed multilocus analyses of the
pike cichlids based on three mitochondrial genes [mt-cyb,
mt-nd2, mt-rnr2 (16S)] and one nuclear (rps7), comprising
602 sequences for 169 specimens. at paper focused on the
biogeography of southern lineages placed in the Crenicichla
lacustris group by most previous studies, and the sampling was
performed accordingly: 134 of the 161 terminal samples of
Crenicichla represented these lineages; the remaining lineages of
Crenicichla (and Teleocichla) were poorly represented. However,
this study provided the rst broad-scale phylogenetic hypoth-
esis for the pike cichlids. In agreement with most of Kullander
et al.’s (2010) results, it recovered the previously named species
groups and provided further support for a diverse C. lacustris
group. Within this group, Piálek et al. (2012) corroborated the
existence of a species ock endemic to the Río Uruguay (C.
missioneira complex) and proposed another one for the endemic
species from the tributaries of the middle Río Paraná basin and
Río Iguaçu (C. mandelburgeri complex). Finally, that study re-
inforced the idea that Teleocichla is nested within the Crenicichla
clade, making it paraphyletic, as suggested by Ploeg (1991) but
unresolved in Kullander et al. (2010).
More recently, phylogenetic studies of pike cichlids entered
the ‘genomic era’. Burress et al. (2017) used ultraconserved
elements (UCEs) of 30 species of Crenicichla (28 valid, two
putative new) and three species of Teleocichla (two valid, one
new). Burress et al. (2018) also analysed ddRad sequencing data
(double digest restriction-site associated DNA) covering most
of the taxonomic diversity of the clade (59 species of Crenicichla
and ve of Teleocichla). Ilves et al.’s (2018) exon-based
phylogenomic analysis of the Neotropical genera, which broadly
replaced López-Fernández et al.’s (2010) multilocus phylogeny,
included 12 species of Crenicichla and three of Teleocichla. ese
phylogenies comprised hundreds of loci (hundreds of thou-
sands of base pairs) and greatly strengthened the phylogeny of
the pike cichlids, building a more robust framework for dating
and usage in phylogenetic comparative methods. Although not a
phylogenomic study, Říčan et al. (2021a) used dense taxonomic
and, for some lineages, population-level sampling of pike cich-
lids of two mitochondrial DNA markers (mt-cyb and ND2). is
study focused on species delimitation analyses for assessment
of the diversity of pike cichlids, but also performed a phylogen-
etic analysis. Except for the sister-group relationship between
Teleocichla and Crenicichla and the paraphyly of the Crenicichla
lugubris group, it agrees well with the previous hypothesis and
provides an incipient phylogenetic background for macroevo-
lutionary studies on evolution of pike cichlids outside of the
Crenicichla lacustris group.
Phylogenomic studies agreed with previous molecular ana-
lyses in corroborating the existence of distinct clades among
the pike cichlids and provide increasingly stable hypotheses of
their composition. However, none of these studies addressed
the classication of pike cichlids, and all highlight several re-
maining challenges that must be addressed before a full revi-
sion of the classication of the Crenicichla s.l. can be completed.
Among them: (i) the available phylogenies present either an
Downloaded from https://academic.oup.com/zoolinnean/advance-article/doi/10.1093/zoolinnean/zlad021/7204395 by Universidade de Sao Paulo (USP) user on 27 July 2023
8 • Var e l l a et al.
unbalanced sampling of the lineages, being more complete for
the C. lacustris group, or a good taxonomic sampling but low mo-
lecular coverage. (ii) e studies that included C. macrophthalma
disagree on its phylogenetic placement. In Kullander et al.
(2010), C. macrophtalma is part of a polytomy with C. wallacii,
C. reticulata, and the C. saxatilis groups (Bayesian inference), or
is unresolved (parsimony). In Piálek et al. (2012) and Říčan et
al. (2021a), it is sister to C. lacustris (as Ploeg inferred in 1991).
In the most recent phylogenomic studies, it is either sister to
the C. lacustris group (100% supermatrix of UCEs, Burres et al.
2017), sister to the C. reticulata group (50–95% supermatrix
of UCEs, Burres et al. 2017), or sister to the clade formed by
the C. lacustris and C. reticulata groups (ddDseq, Burress et
al. 2018). (iii) e latest aempts to review the classication
of pike cichlids using morphology were made by Kullander et
al. (2010) but included only characters traditionally used in the
taxonomic literature, and by Varel la et al. (2018) through the
updated diagnosis of the C. saxatilis group. None of the studies
have yet identied putative synapomorphies to support the re-
lationships of C. macrophthalma to its closely related groups,
nor have they proposed diagnostic characters for the various
clades of pike cichlids.
In this paper, we generate a large dataset of morphological
characters and complementary molecular data compiled from
published sources to revisit the phylogeny of the pike cich-
lids through an integrative approach under parsimony, max-
imum likelihood (ML), and Bayesian inference (BI). Our
study comprises one of the most taxonomically comprehensive
phylogenies of the pike cichlids to date and is the rst using a
morphological dataset. We propose a new classication of the
Crenicichla species groups based on a synthesis of our and pre-
vious phylogenies, providing morphological synapomorphies
and diagnosis for each clade. Secondarily, we provide further
support for hypotheses of the relationship of pike cichlids to the
rest of Neotropical cichlids.
MATERIAL AND METHODS
Terminal taxa for character coding
e datasets combining morphological and molecular data
comprise 98 terminal taxa: 24 outgroup and 74 ingroup taxa
(pike cichlids). Our sampling covers 68 valid species within the
Crenicichla–Teleocichla clade (66.7% of its taxonomic diversity),
including 63.4% of the valid species of Crenicichla and all valid
Teleocichla (Table 2). Outgroup sampling included representa-
tives of successive lineages within the Neotropical cichlid clade
following the phylogeny of Ilves et al. (2018) (Table 3).
We codied morphological characters for 90 of 98 terminal
taxa, including 68 ingroup taxa and 22 outgroup taxa. Our taxon
sampling ensures that most of the morphological and ecological
diversity of pike cichlids is covered. is strategy also ensures
tests of species group monophyly by including multi-species
sampling for all clades of pike cichlids. Our approach does not
aim to resolve species-level relationships within various species
complexes that need further taxonomic investigation but are
beyond the scope of our study (e.g. Crenicichla saxatilis group).
Moreover, we analysed Crenicichla hemera and Crenicichla chicha
for the rst time in a phylogenetic context and tested their hy-
pothesized status as a distinct group (sensu Varel la et al. 2012;
contra Ploeg 1991). For each species group, we examined dif-
ferent ecomorphologies by including both lacustrine and
rheophilic species, as well as taxa known to have divergent diets
or foraging behaviours (piscivorous when adult vs. durophagous
vs. generalized feeding behaviour and diet).
We also included most species in the Crenicichla lacustris
group sensu Piálek et al. (2012) and Burress et al. (2018), be-
cause this group is the most ecomorphologically heterogeneous
among Crenicichla. Also, although its monophyly has been re-
covered by all phylogenies with molecular data, the diagnosis of
the group and its various species complexes is still under devel-
opment. We codied all species of the C. missioneira and C. scoi
complexes from the Río Uruguay, and all species from the Río
Iguaçu and upper Río Paraná basins, including some putative
new species. Crenicichla sp. Paraná corresponds to what has been
previously identied as C. niederleinii (Holmberg 1891; see dis-
cussion in: Varella 2011). Crenicichla lacustris, C. iguapina, and
C. punctata represented the species of the coastal Atlantic drain-
ages, and C. mandelburgeri and C. hu represented species from
the middle Río Paraná basin.
Valid species not included in the dataset are indicated with
asterisks in Table 1. Some valid species were not included be-
cause either specimens or molecular data were not available at
the time when we were developing the character coding and
matrix (indicated with *). In other cases, some species were not
included because other, morphologically very similar species
were already in the matrix or because their identication was un-
certain (**). However, most of the excluded species were part
of the Crenicichla saxatilis group (***), of which we purposefully
performed a reduced sampling. e C. saxatilis group contains
34 nominal especies, but only 23 are considered valid (Varell a
et al. 2018; but see: Říčan et al. 2021a). However, the taxonomy
of this clade remains confusing, and the group is a priority for
species-level revision. Pending such revision, we sampled the
group following observations by Varella et al. (2018), which ten-
tatively linked morphological variation within the group with
habitat associations and biogeography, recognizing three sub-
groups. We included Crenicichla saxatilis, Crenicichla lepidota, C.
britskii, and C. labrina as representatives of the deeper-bodied
taxa (sensu Va rella et al. 2018), with low number of E1 scales
and more elongated suborbital marking (stripe-like). Crenicichla
inpa and C. ploegi represent the slender taxa with high number
of E1 scales, short and triangular suborbital marking, and hu-
meral blotch below the lateral line. Finally, Crenicichla alta and C.
anthurus represent the slender species with high E1 scale counts,
short and triangular suborbital marking, and humeral blotch
with centre on the lateral line (i.e. blotch displaced dorsally).
Specimen acquisition and preparation
A list of the specimens gathered for morphological character
coding, from several ichthyological collections worldwide, is
presented in the Supporting Information, File S1. e col-
lection codes of the repository institutions follow Fricke
and Eschmeyer (2022). Size of the specimens is represented
by the standard length (SL), which is measured from the
tip of the upper jaw to the middle of the base of the caudal
n. e comparative study of external morphology was per-
formed directly on specimens preserved in ethanol, except
for some characters of coloration that demanded observation
Downloaded from https://academic.oup.com/zoolinnean/advance-article/doi/10.1093/zoolinnean/zlad021/7204395 by Universidade de Sao Paulo (USP) user on 27 July 2023
New generic classication of pike cichlids • 9
of fresh material in the eld or photos of live specimens (see
Acknowledgments for colleagues who provided photos of spe-
cimens). e osteological study was based mainly on cleared
and stained specimens following the protocol of Taylor and
van Dyke (1985, with adaptations) or on stained specimens
following the protocol of Datovo and Bockmann (2010), in
which bones and cartilage are stained but muscles are not di-
gested. Dry skeletons were additionally used for large species.
e osteological study of Crenicichla mandelburgeri was based
on a computerized tomography scan (CTscan), obtained
at the University of Michigan Museum of Zoology, using a
Nikon XT H 225ST µCT Scanner, and visualized and edited
using DGONFLY soware v.3.6 (Object Research Systems
Inc, under a free academic license granted to HLF). As a re-
sult, this species was not codied for characters related to ten-
dons and ligaments.
Character coding for morphological datasets
Morphological characters proposed in previous phylogenetic
studies of Cichlidae were revised (e.g. Cichocki 1976; Stiassny
1981, 1982, 1987, 1991, 1992; Stiassny and Jensen 1987;
Lippistch 1990, 1993, 1995; Webb 1990; Cascioa and Arratia
1993a, b; Kullander 1998; López-Fernández et al. 2005a, 2012;
Landim 2006; Chakrabarty 2007). Although many new charac-
ters were proposed herein, most of the informative characters for
the ingroup relationships were derived from previous taxonomic
studies, highlighting Kullander (1988, 1990a, 1990c, 1991,
1997, 2006), Ploeg (1991), Lucena and Kullander (1992),
and Varella et al. (2018). Colour-marking terminology follows
mainly Kullander (1986, 1988), Ploeg (1991), López-Fernández
et al. (2005a), and Vare lla et al. (2012, 2018), but we renamed
elements when necessary to describe primary homologies, des-
pite the nomenclature used in previous studies. Osteological
Table 2. Commented taxonomic sampling of the ingroup (Crenicichla–Teleocichla clade), with the total of samples included in the combined
analyses, number of taxa represented with both morphological and molecular data (TE), with morphological data only (Morph), or with
molecular data only (Mol)
Total samples TE (Morph/Mol) Comments
C. macrophthalma 21 (1/2) C. macrophthalma-P treated as additional taxon with molecular data
only; C. macrophthalma-MK represented with molecular and
morphological data.
C. lacustris group sensu
Piálek et al. (2012)
32 20 (29/23) Including all species of the C. missioneira and C. scoi complexes
(Río Uruguay), and all species from the Río Iguaçu and upper Río
Paraná basins. Also including four putative new species sensu Var el la
(2011): Crenicichla sp. nov. Paraná, Crenicichla sp. nov. Paranaíba;
Crenicichla a. iguassuensis BIG LIPS, and Crenicichla a. tesay BIG
LIPS. ese new species and other ve valid species are represented
with morphological characters only, whereas C. ypo, C. ijhui, and C.
taikyra represented with molecular data only.
C. reticulata group sensu
Ploeg (1991), updated
65 (6/5) C. cametana, C. cyclostoma and C. jegui are ecologically dier-
ent sympatric species from rapids of the Río Tocantins basin.
Crenicichla geayi, C. reticulata and C. semifasciata are widespread spe-
cies of the Orinoco River, Amazon basin (including Guianas), and
La Plata basin, respectively.
C. wallacii group sensu
Kullander (1990a) and Ito
and Py-Daniel (2015)
64 (6/4) Covering both geographically widespread species (C. wallacii, C.
regani, and C. notophthalmus) and the localized endemic, rheophilic
C. compressiceps, C. urosema, and C. heckeli; the last two included for
the rst time in phylogenetic analyses, represented with morpho-
logical data only.
C. saxatilis group sensu
Vare ll a et al., 2018)
83 (8/3) Including representatives of slender and deep-bodied taxa following
comparisons from Varella et al. (2018; see text).
C. lugubris group sensu
Ploeg (1991)
95 (8/6) Including four of the 10 species of the C. lugubris group s.s. and four
of seven species from the C. acutirostris group sensu Montaña et
al. (2008). Crenicichla lenticulata Colombia corresponds to Pialek
et al.’s (2012) sequences (voucher from aquarium trade without
detailed locality, not reexamined); treated as additional taxon with
molecular data only.
C. hemera and C. chicha 20 (2/-) Never included in phylogenetic analyses and hypothesized to form a
dierent group (Varella et al. 2012; contra Ploeg 1991)
Teleocichla 96 (9/6) All nine valid species included. T. centisquama, T. prionogenys, and T.
wajapi represented with morphological data only.
Ingroup 74 44 (69/49) 25 ingroup taxa represented with morphology only and ve
represented with molecular data only.
Downloaded from https://academic.oup.com/zoolinnean/advance-article/doi/10.1093/zoolinnean/zlad021/7204395 by Universidade de Sao Paulo (USP) user on 27 July 2023
10 • Var e l l a et al.
nomenclature follows Stiassny (1982) and Kullander (1988,
1998). Nomenclature for individual ossicles of the infraorbital
series follows López-Fernández et al. (2005a: 645, g. 11A).
Most of the aforementioned studies provided good-quality
illustrations of many of the characters used herein, particularly
those of squamation and osteology that are mostly informative
for relationships of pike cichlids among Neotropical cichlids.
Kullander (1988) described and illustrated in detail the oste-
ology of Teleocichla. Taxonomic papers of Kullander (1983,
1986, 1989, 1996) and Kullander and Nijssen (1989) also pro-
vide richly illustrated anatomical descriptions. erefore, illus-
trations in literature were linked to our observations and new
illustrations were restricted to new characters or those that have
not been well-illustrated before.
During character examination, we oen observed problems
concerning the logical basis of character coding, mostly re-
lated to non-mutually exclusive states in multistate characters.
is is particularly common in mixed/composite coding, also
called ‘complex characters’ (see: Kitching et al. 1998; Strong
and Lipscomb 1999), when two or more variables are com-
bined to describe a multistate character. e combination of a
neomorphic variable related to appearance or lack of a structure
(i.e. presence/absence) with one or more transformational vari-
ables (i.e. its colour, size, or other shape changes), or the com-
bination of two or more transformational variables, can lead to
redundance in at least two of the character states and, conse-
quently, to the lack of independence between them. Denition
of neomorphic and transformational variables follows Sereno
(2007).
Herein, we avoided composite coding by applying a re-
ductive coding (e.g. Strong and Lipscomb 1999) and follow
Sereno (2007) in the formulation of a character. If a structure
is considered absent for certain taxa, it is codied with in-
applicable data (-) for characters based on transformational
variables related to this structure, since it is not possible to
infer modications in shape, colour, size, and other aspects
for something that is missing. us, each character statement
is formulated as follows [see Sereno (2007) for nomenclature
of components]: locator(s); variable (and variable quantier, if
necessary): mutually exclusive character states (using brackets
to dene the number of each state). ‘Occurrence’ is used as
the variable for neomorphic characters to dene the pres-
ence or absence of a structure. Example of neomorphic char-
acter statement: [secondary locator] caudal n, [locator] dark
blotch; [variable] occurence: [0] absent; [1] present. Example
of a transformational character statement: [secondary locator]
caudal n, [locator] dark blotch; [variable] shape: [0] ovoid
(vertically elongated); [1] rounded. A list with description, il-
lustrations, and comments on the character statements is pre-
sented in the Supporting Information, Appendix S2. References
of usage of those characters in the literature are provided to-
gether with the matrices in Supporting Information, File S2.
In the main morphological matrix (DiscreteMatrix),
multistate characters were unordered, except the rst 16 char-
acters (numbered 0–15), which are quantitative and present
morphoclinal variation between the extreme states. e ana-
lysis of parsimony performed in TNT allows treating quanti-
tative characters without being ‘discretized’ arbitrarily a priori
(Golobo et al. 2006). us, the variation is normalized as a
one-step character (i.e. normalized with values between 0 and
1) and the optimization is performed directly on the normalized
data instead of on the discretized states. is has not been imple-
mented in programs performing Bayesian inference (BI) or max-
imum likelihood (ML) analysis (but see: Parins-Fukuchi 2018).
To explore the eect of analysing continuous data as such on to
the parsimony analysis, we created another dataset with the rst
16 characters considered as continuous and treated as normalized
data (ContinuousMatrix). In the case of the DiscreteMatrix, the
rst 16 characters were discretized arbitrarily. Morphometric
characters were discretized using the entire range for each taxon
Table 3. Taxonomic sampling of the outgroup, with the total of samples included in the combined analyses, number of taxa represented with
both morphological and molecular data (TE), with morphological data only (Morph) or molecular data only (Mol)
Total samples TE (Morph/Mol) Comments
Tribe Geophagini 17 12 (15/14) Acarichthys heckelii (Müller and Troschel 1849); Apistogramma
taeniata (Günther 1862); Biotodoma wavrini (Gosse 1963);
Biotoecus dicentrarchus Kullander 1989; Biotoecus opercularis
(Steindachner 1875); Crenicara punctulatum (Günther 1863);
Dicrossus lamentosus (Ladiges 1958); Geophagus altions
Heckel 1840; ‘Geophagus’ brasiliensis (Quoy and Gaimard 1824);
‘Geophagus’ steindachneri Eigenmann and Hildebrand in Eigenmann
1910; Gymnogeophagus meridionalis Reis and Malabarba 1988;
Mikrogeophagus ramirezi (Myers and Harry 1948); Satanoperca dae-
mon (Heckel 1840); Satanoperca Lilith Kullander and Ferreira 1988;
and Taeniacara candidi Myers 1935.
Other Neotropical
cichlids
75 (6/6) Heroini: Australoheros minuano Říĉan and Kullander 2008.
Cichlasomatini: Acaronia nassa (Heckel 1840) and Cichlasoma
araguaiense Kullander 1983. Astronotini: Astronotus ocellatus Agas-
siz in Spix and Agassiz 1831. Chaetobranchini: Chaetobranchus
avescens Heckel 1840. Retroculini: Retroculus xinguensis Gosse
1971. Cichlini: Cichla pinima Kullander and Ferreira 2006
Outgroup 24 17 (21/20) Four outgroup taxa represented with morphology only and
three taxa represented with molecular data only.
Downloaded from https://academic.oup.com/zoolinnean/advance-article/doi/10.1093/zoolinnean/zlad021/7204395 by Universidade de Sao Paulo (USP) user on 27 July 2023
New generic classication of pike cichlids • 11
instead of the mean and standard deviation. Meristic data were
discretized based on the modal values for each terminal taxon
as a way to decrease the amount of polymorphic data, which
are treated as unknown by ML and BI soware. All matrices are
presented in the Supporting Information, File S2 and the tables
used for normalization of the continuous data are presented in
the Supporting Information, File S3.
Molecular data
Although this is a morphology-driven analysis, it is unlikely that
a species-level phylogeny of such a diverse and complex group as
the Crenicichla clade could be resolved using morphological char-
acters only. Even the most recent studies on pike cichlids, using
hundreds of molecular loci (Burress et al. 2017, 2018), found
that relationships among the major groups of pike cichlids can be
discordant. Moreover, these and other studies (Piálek et al. 2012,
2019a) also suggest a high degree of ecomorphological conver-
gence in pike cichlids, resulting in high levels of homoplasy that
probably limit the ability of morphology alone to resolve some
relationships. In this context, our strategy was to use available
molecular data from previous studies to complement the mor-
phological matrices. We included morphological data for several
geophagine and other Neotropical taxa as outgroups, because a
secondary objective of this study was to further test the place-
ment of the clade of pike cichlids within the tribe Geophagini
when a large portion of this clade and morphological characters
are included, because previous studies included limited sam-
pling of pike cichlids (López-Fernández et al. 2005a,b, 2010,
2012; Ilves et al. 2018).
A list with all the terminal taxa used in the integrative ana-
lysis with accession codes and references are presented in the
Supporting Information, File S4. We obtained previously pub-
lished sequence data for seven loci: mitochondrial coding genes
cytochrome c oxidase subunit I (mt-co1 or COI), cytochrome b
(mt-cyb or Cytb), NADH dehydrogenase subunit 2 (mt-nd2 or
ND2), NADH dehydrogenase subunit 4 (mt-nd4 or ND4), the
mitochondrial non-coding 16S ribossomal rRNA (mt-rnr2 or
16S), the nuclear, non-coding ribosomal protein S7 intron 1 (rps7
or S7), and the coding nuclear gene recombination activating
protein 2 (rag2). Most sequences were extracted from GenBank
(Clark et al. 2016), except for some mt-co1 data extracted from
e Barcode of Life Data System (BOLD; Ratnasingham and
Hebert (2007)). Gene abbreviations follow the standard no-
menclature of the Zebrash Information Network-ZFIN data-
base (Bradford et al. 2022) and of Genbank.
We prioritized sequences whose vouchers could be con-
rmed by us, including the sequences from Kullander et al.
(2010), López-Fernández et al. (2005a,b, 2010), and almost all
BOLD sequences. We also checked for updates in identication
from previous studies, as some misidentications of species of
Crenicichla in Piálek et al. (2012) that were corrected in Piálek
et al. (2019a).
e molecular dataset comprised 69 terminal taxa, 61 of which
were also codied for morphology. Terminal taxa for which
only molecular data were available include three outgroups:
Chaetobranchus avescens, Geophagus surinamensis, and Mazarunia
mazarunii. Sequences for Crenicichla macrophthalma-P were pub-
lished by Piálek et al. (2012) from the aquarium trade (arguably
from Río Trombetas basin), and we were not able to examine the
voucher, therefore we considered Crenicichla macrophthalma-P as
an additional taxon with molecular data only. Likewise, sequences
from Pialek et al. (2012), from aquarium trade specimens without
dened locality (Colombia), were added as terminal taxa with
molecular data only. Crenicichla macrophthalma-MK corresponds
to a mt-cyb sequence from Kullander et al. (2010) and a mt-co1
sequence from BOLD; in this case, we matched morphological
and molecular data because we conrmed the identication of the
vouchers. Similarly, Crenicichla lenticulata sequences from López-
Fernández et al. (2010) were matched with the morphological
dataset as we could verify the voucher. Crenicichla ypo, Crenicichla
ijhui, and Crenicichla taikyra for which we could not examine spe-
cimens for morphological coding are represented with molecular
data only.
Alignments, translation of coding regions to account for
stop codons and correction of reading frames, when necessary,
were performed in GENEIOUS PRIME 2020.0.5 (Biomaers
Ltd, 2005–2020) using the MUSCLE algorithm (Edgar 2004).
Alignments were inspected by eye and ends were trimmed when
there were gaps in many taxa, to decrease the amount of missing
data and ambiguously aligned sites. Ending gaps (-) were con-
verted to missing data (?) but gaps along the sequences in
mt-rnr2 (16S) and rps7 were maintained. Summary information
on the resulting alignments is provided in Table 4.
Phylogenetic analyses
Matrices with all morphological data discretized a priori
(DiscreteMatrix) were analysed under ML, BI, and parsimony
criteria (Table 5). Matrices with the continuous data treated
as such (ContinuousMatrix) were analysed under parsimony
only. Molecular data in the ‘total evidence’ (TE) approach
combined datasets were the same for DiscreteMatrixTE and
ContinuousMatrixTE.
IQtree (Chernomor 2016) was used to select the best-t
models of nucleotide substitution for both ML and BI analyses.
In the main ML analysis shown in this paper, molecular data were
partitioned by gene, but we explored alternative partitioning
schemes (e.g. coding genes third codon position partitioned
from the rst and second), obtaining very similar results (not
shown). ML analysis was performed using IQtree (Nguyen
et al. 2015) because it implements a Jukes–Cantor model for
morphological data allowing exchange of neighbouring states
only (called ORDERED) that beer ts the rst 16 characters
in our morphological matrix. Moreover, IQtree also applies an
ascertainment bias model (+ASC) that avoids overestimation
of branch lengths due to the absence of constant sites, a typical
aribute of morphological data (Lewis 2001). We divided mor-
phological characters into two partitions for the ML analyses:
rst 16 characters under the ORDERED + ASC model and the
remaining under the Mk + ASC model. Branch support was cal-
culated using ultrafast bootstrap using 1000 replicates.
Bayesian (BI) analysis of the combined dataset was performed
in MrBayes v.3.2.6 (Huelsenbeck and Ronquist 2001; Ronquist
and Huelsenbeck 2003), treating the morphological characters
and the ve molecular markers as separate partitions (Table 5).
Unlike in the parsimony and ML analysis, the rst 16 morpho-
logical characters could not be set up as ordered, thus we treated
Downloaded from https://academic.oup.com/zoolinnean/advance-article/doi/10.1093/zoolinnean/zlad021/7204395 by Universidade de Sao Paulo (USP) user on 27 July 2023
12 • Var e l l a et al.
all morphological characters as unordered using the Lewis
(2001) Mk model. Two independent runs with eight MCMC
chains each were run for 4000000 generations sampling every
4000 generations. e intial 25% of trees from each run were
analysed in TCER v.1.6 (Rambaut et al. 2014), considering
200 eective samples as criterion to determine stationarity.
Node posterior probabilities (estimated in MrBayes) were used
as a measure of branch support.
Parsimony analyses were performed in TNT 1.1 (Golobo
et al. 2003, 2008a; made available by the Willi Hennig Society)
dividing the morphological data into two partitions. In the
DiscreteMatrixTE, the rst partition included the rst 16 char-
acters (numbered 0–15) set as ordered and the second parti-
tion included the remaining 200 characters unordered. In the
ContinuousMatrixTE, the rst partition included the rst 16
characters (characters 0–15) set as continuous and the second
partition with all remaining characters unordered. Ending gaps
of gene alignments were considered as missing data, but within
sequence gaps in mt-rnr2 (16S) and rps7 alignments were
treated as a h state. e main parsimony analyses were per-
formed with all morphological and molecular characters equally
weighted (EW). However, we also explored the eect of the
cladistic congruence between the dierent sources of data, con-
sidering their intrinsic characteristics (i.e. type of data, various
amounts of missing data in each partition) by using extended
implied weighting (XIW; Golobo 2014). For XIW analyses,
we divided the data into nine subsets (continuous and discret-
ized morphological characters plus seven genes) and set the
weighting scheme as follows: to weight each character individu-
ally in the two morphological partitions, to weight-by-average-
homoplasy in the non-coding genes mt-rnr2 (16S) and rps7, and
to weight by codon position in the coding genes (mt-cyb, mt-nd2,
mt-nd4, mt-co1, and rag2). A ‘mild’ constant k = 12 was used in
the XIW analyses as recommended for larger datasets, as it does
Table 4. Summary of information on the morphological and molecular data, with the taxonomic coverage and amount of missing data
Number of taxa Taxonomic coverage
(out/ingroup)
Alignment length Parsimony
informative sites
Missing data—total
(outgroup/ingroup)
mt-rnr2 (16S) 55 (71%/51%) 551 bp 165 (30%) 44% (29%/49%) + 3% gaps
mt-cyb 59 (71%/57%) 1049 bp 504 (48%) 42% (33%/44%)
mt-nd2 33 (13%/41%) 1047 bp 439 (42%) 66% (88%/69%)
mt-nd4 16 (42%/8%) 631 bp 360 (57%) 84% (58%/92%)
mt-co1 28 (58%/19%) 648 bp 273 (42%) 72% (43%/81%)
rag2 22 (63%/9%) 850 bp 54 (6%) 80% (37%/91%)
rps7 48 (63%/45%) 552 bp 115 (21%) 52% (39%/56%) + 3% gaps
Molecular data
(concatenated)
69 (83%/66%) 5328 bp 1910 (36%) 63% (50%/67%) + 0.6%
gaps
Morphological data 90 (88%/93%) 216 characters 100% 9% missing data (12%/7%)
5% inapplicable data
(5%/5%)
0.1% polymorphisms
Table 5. Summary of the phylogenetic analyses performed on the combined datasets. Partitions and best models of nucleotide evolution
and phenotypic data are indicated for ML and BI analyses. Dierent treatments of data and weighting schemes are indicated for Parsimony
analyses; k = constant of concavity; h indicates weight-by-average-homoplasy (for non-coding genes) and h1:2:3 indicates weighting following
homoplasy of gene data partitioned by rst, second and third position of codons (for coding genes)
Maximum
likelihood
(DiscreteMatrixTE)
Bayesian inference
(DiscreteMatrixTE)
Parsimony
(DiscreteMatrixTE)
Parsimony
(DiscreteMatrixTE)
Parsimony
(ContinuousMatrixTE)
Parsimony
(ContinuousMatrixTE)
mt-rnr2
(16S)
TIM2e + I + G4 SYM + I + G4 EW XIW k = 12 + h EW XIW k = 12 + h
mt-co1 TPM2 + F + I + G4 GTR + F + I + G4 EW XIW k = 12 + h1:2:3 EW XIW k = 12 + h1:2:3
mt-nd2 TIM2 + F + I + G4 GTR + F + I + G4 EW XIW k = 12 + h1:2:3 EW XIW k = 12 + h1:2:3
mt-nd4 TPM2u + F + I + G4 GTR + F + I + G4 EW XIW k = 12 + h1:2:3 EW XIW k = 12 + h1:2:3
mt-co1 TPM2 + F + I + G4 GTR + F + I + G4 EW XIW k = 12 + h1:2:3 EW XIW k = 12 + h1:2:3
rag2 K2P + R2 K2P + R2 EW XIW k = 12 + h1:2:3 EW XIW k = 12 + h1:2:3
rps7 K2P + G4 HKY + F + G4 EW XIW k = 12 + h EW XIW k = 12 + h
Morph.
data
part1 (char 0-15) unique partition part1 (char 0-15) part1 (char 0-15) part1 (char 0-15) part1 (char 0-15)
ORDERED + ASC MK model EW ordered ordered XIW k = 12 continuous continuous XIW k = 12
part1 (char 16-215) all unordered and
without
polymorphism
part2 (char 16-215) part2 (char 16-215) part2 (char 16-215) part2 (char 16-215)
MK + ASC EW unordered unordered XIW k = 12 EW unordered unordered XIW k = 12
Downloaded from https://academic.oup.com/zoolinnean/advance-article/doi/10.1093/zoolinnean/zlad021/7204395 by Universidade de Sao Paulo (USP) user on 27 July 2023
New generic classication of pike cichlids • 13
not downweight putative homoplastic characters as strongly as
lower k-values (e.g. default k = 3, Golobo et al. 2008b, 2018).
Searches were performed using both traditional heuristic
searches and TNT’s ‘new technologies’. Traditional heuristic
searches were performed using the tree bisection and reconnec-
tion (TBR) algorithm. us, the traditional parsimony analyses
consisted of a rst TBR run with 10000 replications, saving 100
trees per replication, ensuring that there was no overow of trees
per replication (i.e. trees generated by the replication that could
not be saved because of insucient memory). Additional runs
of TBR branch-swapping on the trees saved in memory until the
number of most parsimonious trees (hereinaer MPT) reaches
stationarity. Parsimony analysis using TNT’s ‘new techonologies’
can be more eective than TBR algorithms for complex datasets
in the search for the global optima (vs. local peaks), by using al-
gorithms to perform cyclic perturbation and search phases. For
this analysis, we used default parameters of sectorial searches
(RSS, CSS, XSS), ratchet iteractions, dri cycles, and tree fusing
runs, and the driven search set to nd trees until reaching the
best score 50 times and stabilizing the consensus ve times with
factor 75. e trees obtained from these searches were also sub-
mied to additional runs of branch swapping. Finally, the most
parsimonious trees (MPT) were summarized into a strict con-
sensus tree. Absolute and relative Bremer’s (1994) support
(ABS and RBS, respectively) based on sampling of suboptimal
trees with 10 additional steps were calculated as measures of
branch support for parsimony topologies. ABS corresponds to
the Bremer’s decay index (number of extra steps needed to col-
lapse a clade). RBS (Golobo and Farris 2001) is also based on
the Bremer support, but takes into account relative amounts of
favourable and contradictory evidence (ratio between presence
and absence of a clade among the sampling of suboptimal trees).
Character reconstructions and synapomorphies
Character reconstructions were performed under parsimony
optimization of both morphological and molecular characters
to identify synapomorphies and character transformation series
for each group using the MP and ML trees in TNT. While we
prioritized proposed morphological synapomorphies based on
the ML tree, alternative reconstructions in the parsimony top-
ology are discussed when pertinent. Character states’ polariza-
tion was based on outgroup comparisons (Nixon and Carpenter
1993), using Retroculus xinguensis as the root, since it has been
recovered as sister-group of all remaining Neotropical cichlids
(e.g. Kullander 1998—Retroculinae) or as part of the sister-
group, together with Cichla (e.g. López-Fernández et al. 2010;
Ilves et al. 2018—Retroculini).
Optimizations of ambiguous morphological transformations
do not follow any a priori decision (i.e. AccTran or DelTran
options). AccTran is much more widely used (a priori) than
DelTran in phylogenetic studies based on phenotypic characters,
certainly inuenced by De Pinna’s (1991) philosophical justi-
cation for preferring reversals over parallelism for complex traits.
However, we follow Agnarsson and Miller’s (2008) arguments
that neither AccTran nor DelTran consistently minimize parallel
gain of complex traits and, therefore, there are no theoretical
grounds for favouring any of them a priori for the optimization
of all ambiguous characters. We performed ad hoc optimization
for each ambiguous transformation, considering interpretation
of the character in previous phylogenetic studies or evidence in
the literature. For example, parallel evolution of traits related to
feeding strategies is well-documented in sympatric assemblages
of pike cichlids, which include several characters of oral and pha-
ryngeal jaws analysed herein (e.g. Piálek et al. 2012, Burress et al.
2017, 2018; Piálek et al. 2019a). Notation of ambiguous trans-
formations is followed by an ‘A’ (AccTran) or a ‘D’ (DelTran),
depending on the option used in the optimization.
RESU LTS
Data heterogeneity and complementarity
We present a taxonomically and anatomically comprehen-
sive original morphological dataset comprising 216 characters
(Supporting Information, Appendix S2) of external morph-
ology (scales, ns, and colour paers) and osteology (axial skel-
eton, neurocranium, suspensorium, branchial and hyoid arches,
and pectoral and pelvic girdles) coded for 90 terminal cichlid
taxa (22 outgroup taxa and 68 pike cichlid species). Missing
data in this morphological dataset are mostly restricted to eight
taxa for which molecular data were available but not coded for
morphology. Due to the way in which characters were coded (i.e.
reductive coding preferred to composite coding), 5% of the data
are inapplicable.
Resulting alignments derived from published sources of mo-
lecular data (seven genes, 5328bp; Supporting Information, File
S5) varied greatly in relation to the amount of missing data and
taxon coverage per locus (Table 4). Largely used in studies of
Neotropical cichlids and of pike cichlids, the markers mt-rnr2
(16S) mt-cyb, and rps7 included the broadest sampling of both
outgroup and ingroup included in our morphological dataset,
with missing taxa ranging between 42% and 52%. ND2 (mt-
nd2), used in Piálek et al. (2012, 2019a), included a comprehen-
sive sampling of pike cihlids but only two outgroup taxa (one
species of Satanoperca and one species of Astronotus species),
representing 41% of species of pike cichlids, but only 13% of
outgroup taxa in our dataset. e molecular marker recombin-
ation activating protein 2 (rag2) was used in studies focused on
Neotropical cichlids but with low sampling of pike cichlids, re-
sulting in good representation of outgroups but only 8–9% of
pike cichlid sampling. Data of mt-co1, while not broadly used in
the phylogenetic studies of Neotropical cichlids, were available
for 28 taxa, with more outgroup (58% of taxa, 43% missing data)
than ingroup (19% of taxa, 81% missing data) representatives.
Main phylogenetic hypothesis and multiple alternatives
We used the the ML tree resulting from the combined dataset
(hereaer ML tree; –ln = –58115.602; Fig. 1, le; Fig. 2, le)
as the reference for topological descriptions and taxonomic de-
cisions throughout the paper. ML analysis provided a fully re-
solved topology (96 nodes) combining model-based treatment
of the morphological characters (i.e. rst 16 characters ordered
and application of ascertainment bias correction model) and
molecular data (i.e. partitioning and modelling). e ML ana-
lyses also recovered most of the relationships among outgroup
taxa based on recent comprehensive phylogenies of Neotropical
cichlids (e.g. López-Fernández et al. 2010; Ilves et al. 2018), as
Downloaded from https://academic.oup.com/zoolinnean/advance-article/doi/10.1093/zoolinnean/zlad021/7204395 by Universidade de Sao Paulo (USP) user on 27 July 2023
14 • Var e l l a et al.
well as relationships among the major groups of pike cichlids.
e consensus topology obtained through BI (BI tree; 90 nodes;
Supporting Information, Appendix S3, Fig. S1) is very similar to
the ML tree with a notable exception discussed below.
Unlike the ML and BI topologies, parsimony analyses of the
combined datasets resulted in dierent arrangements depending
on the treatment of the data or parameters used (Table 6). e
consensus tree based on the analysis of all morphological char-
acters discretized a priori (DiscreteMatrixTE) under equal
weighting (EW) scheme (Fig. 2, right), although less resolved
(82 nodes), agreed well with the ML tree, showing similar re-
lationships between the major groups of pike cichlids and re-
covering all but one of the subgroups. e consensus tree based
on the ContinuousMatrixTE under EW scheme (Supporting
Information, Appendix S3, Fig. S2) resulted in 91 nodes and is
similar to the topology from the analysis of DiscreteMatrixTE
under EW but grouping Crenicichla hemera and C . chicha as sister-
group of all the remaining pike cichlids. Parsimony analyses of
both DiscreteMatrixTE and ContinuousMatrixTE under ex-
tended implied weighting (XIW) resulted in well-resolved con-
sensus topologies (91 nodes each; see Supporting Information,
Appendix S3, Figs S3, S4). Ingroup relationships in the XIW/
DiscreteMatrixTE consensus topology were very similar to
those in the ML tree, except for nding C. macrophthalma nested
within the C. reticulata group or, interpreting alternatively,
nding a sister-group relationship between C. macrophthalma
and the C. reticulata group, while excluding C. jegui from the
group. e XIW/ContinuousMatrixTE topology resulted in
the same arrangement of C. macrophthalma and the C. reticulata
group as the EW/DiscreteMatrixTE analysis, but also recovered
C. hemera–C. chicha as sister to all remaining pike cichlids (i.e.
similar to the EW/ContinuousMatrixTE analysis). Other
minor topological dierences among analyses are commented
as needed in the following sections. Parsimony-based char-
acter reconstructions, lists of synapomorphies, and character
transformation series throughout the topology are given in the
Supporting Information, File S6 for both trees illustrated in
Figure 2 along with further comparisons and comments.
Analyses using morphological datasets exclusively resulted
in topologies with various degrees of resolution and agreement
with hypotheses generated from the combined datasets and pre-
vious studies and were not considered used for taxonomic deci-
sions. Parsimony analysis of the DiscreteMatrix, with the rst 16
characters ordered and analysed under EW, results in a poorly
resolved consensus topology with only 32 nodes but shows
slightly beer resolution (41 nodes) when all characters are un-
ordered. Parsimony analysis of the ContinuousMatrix under EW
resulted in a well-resolved consensus tree (86 nodes). is laer
parsimony tree and the ML tree based exclusively on the mor-
phological data (–ln = –7915.699) are somewhat similar in their
resolution of many of the main groups of pike cichlids. However,
there is evidence of long-branch araction in some parts of the
topologies and the relationships between the groups diverge
from the hypotheses generated with both the combined datasets
and previous studies.
DISCUSSION
An integrative phylogenetic approach for the classication of
pike cichlids
e use of morphological and molecular data in combination is
one kind of ‘total-evidence’ approach common in sh studies in
the last two decades. e classication of Neotropical cichlids
was greatly improved through this method (e.g. Farias et al. 2000;
López-Fernández et al. 2005a; Říĉan et al. 2008; Musilová et al.
2009) because it provides dierent sources of evidence for phylo-
genetic reconstructions and can strengthen hypotheses of rela-
tionships by including taxa represented only by one type of data
or the other. Importantly, including morphological characters in
phylogenetic studies allows identication of synapomorphies and
Tribe Geophagini
Closest
Geophagini taxa
Dicrossus filamentosus
A
caronia nassa
Geophagus brasiliensis
Taeniacara candidi
Geophagus steindachneri
Biotoecus dicentrarchus
Geophagus surinamensis
Biotoecus opercularis
Cichlasoma araguaiense
Mikrogeophagus ramirezi
Geophagus altifrons
Biotodoma wavrini
A
carichthys heckelii
Satanoperca daemon
Retroculus xinguensis
A
ustraloheros minuano
Crenicara punctulatum
A
pistogramma taeniata
Cichla pinima
A
stronotus ocellatus
Mazarunia Mazarunii
Chaetobranchus flavescens
Satanoperca lilith
Gy. meridionalis
SUBTRIBE CRENICICHLINA
“pike cichlids”
6 genera
SUBTRIBE CRENICICHLINA
“pike cichlids”
6 genera
Cichlasoma araguaiense
Mazarunia Mazarunii
A
caronia nassa
Retroculus xinguensis
A
pistogramma taeniata
Mikrogeophagus ramirezi
Crenicara punctulatum
A
ustraloheros minuano
Biotoecus dicentrarchus
Taeniacara candidi
Chaetobranchus flavescens
Gy. meridionalis
Geophagus surinamensis
A
stronotus ocellatus
Geophagus steindachneri
Biotoecus opercularis
Geophagus brasiliensis
A
carichthys heckelii
Satanoperca lilith
Dicrossus filamentosus
Geophagus altifrons
Cichla pinima
Satanoperca daemon
Biotodoma wavrini
100
100
100
98
100
100
98
100
100
100
100
100
113
115
124
103
110
120
119
116
106
108
114
100
112
99
109
118
104
117
121
105
102
111
101
107
100
107
108
101
100
116
106
102
114
110
115
103
105
100
117
100
99
104
26
111
100
8
41
5
12
100
69
10
25
5
13
112
100
113
13
109
8
100
ML
analysis
DiscreteMatrixTE
ML
tree (96 nodes)
Parsimony
Equal Weighting
DiscreteMatrixTE
Strict consensus of
384 MPTs
(82 nodes)
98
100
81
98
87
86
93
92
77
93
11
120
Figure 1. Relationships of the pike cichlids among Neotropical cichlids according to the ML tree (le), –ln = –58115.602, and the strict
consensus of 384 most parsimonious trees (right), with 12233 steps. Both analyses were based on 216 discrete morphological characters and
5328bp distributed along seven genes (mt-rnr2 [16S], mt-cyb, mt-nd2, mt-nd4, mt-co1, rag2, and rps7). Parsimony analysis performed with all
characters equally weighted. Clade numbers shown above the branches of the ML and parsimony trees. Bootstrap values (in percentage) below
the branches of the ML tree and relative Bremer support (in percentage) below the branches of the parsimony tree.
Downloaded from https://academic.oup.com/zoolinnean/advance-article/doi/10.1093/zoolinnean/zlad021/7204395 by Universidade de Sao Paulo (USP) user on 27 July 2023
New generic classication of pike cichlids • 15
generation of phenotypic diagnoses of clades. More recently, this
integrative approach has largely been replaced by high-throughput
sequencing datasets (exon-based, UCEs, or ddDseq), which
yield hundreds to thousands of loci with thousands of informative
characters. In this new scenario, morphological characters have
lile or no inuence on the phylogenetic reconstructions due to
the huge amount of molecular data available but are still essential
for diagnosing clades and for incorporating taxa for which no mo-
lecular data are available, particularly fossils.
In this study, we have used an integrative approach to analyse a
taxonomically comprehensive sampling of pike cichlids, including
the most extensive morphological matrix available for any cichlid
GENUS
Saxatilia
SUBGENUS
Crenicichla
( Batrachops)
SUBGENUS
Crenicichla
( Lacustria)
GENUS
Wallaciia
GENUS
Teleocichla
GENUS
Lugubria
GENUS
Hemeraia
*
GENUS
Crenicichla
18
7
1
65
1
62
1
63
1
53
136
138
1
4
8
1
3
2
1
3
1
1
24
133
191
126
175
1
7
3
182
155
180
192
181
1
7
6
1
4
6
189
1
2
8
13
7
168
158
1
85
16
7
147
135
14
3
1
77
188
15
2
1
7
8
154
1
4
4
157
12
9
1
4
2
18
4
190
193
1
7
9
1
7
2
186
141
1
25
150
1
7
0
12
7
130
1
4
0
15
1
1
2
2
16
4
1
39
1
60
171
161
134
1
4
9
183
169
1
4
5
156
17
4
159
194
1
2
3
166
CLADE A
CLADE A
96
GENUS
Crenicichla
98
85
1
00
1
00
1
00
89
90
1
00
96
99
1
00
94
100
1
00
96
1
00
100
99
99
98
95
1
00
99
1
00
83
67
100
1
00
98
92
98
1
00
71
73
73
94
78
95
90
74
79
91
95
73
96
89
100
88
81
91
66
57
68
82
86
90
90
67
63
69
28
84
77
99
87
85
98
87
87
86
90
88
1
74
1
1
9
1
2
1
133
172
1
4
6
1
20
1
77
1
4
8
152
1
35
15
4
1
3
4
1
26
14
3
156
1
39
1
6
9
1
4
0
1
7
3
158
17
5
171
1
7
8
145
168
1
60
123
12
7
11
8
1
70
141
166
1
4
7
1
7
6
1
55
1
6
4
1
36
1
4
9
125
1
57
153
137
1
7
9
150
1
2
4
1
44
1
6
1
1
22
138
165
1
80
1
62
159
1
4
2
16
3
1
32
1
3
1
15
1
128
167
130
129
ML analysis
DiscreteMatrixTE
ML tree (96 nodes)
Parsimony Equal Weighting
DiscreteMatrixTE
Strict consensus of 384
MPTs
(82 nodes)
22
4
3
33
83
2
5
1
7
20
20
100
1
7
17
20
8
6
7
12
22
25
33
63
80
20
1
7
29
50
2
9
20
20
63
60
20
100
40
1
7
1
7
21
33
1
7
50
20
33
4
0
1
7
17
9
1
3
1
7
4
0
33
44
100
1
7
50
14
50
1
7
100
50
43
33
21
20
20
20
stri
g
at
a
lenticulat
a
i
g
ar
a
m
i
ss
i
one
i
ra
Paran
a
c
in
de
r
e
ll
a
scott
i
i
monogramma
multis
p
inos
a
l
ab
rin
a
em
ph
ere
s
p
rend
a
h
u
u
r
ose
m
a
h
ec
k
e
l
i
p
r
i
onogeny
s
i
guassuens
is
noto
p
hthalmu
s
iguassuensis
BL
macrophthalma
P
gauc
ho
j
o
h
ann
a
lacustris
reticulat
a
lu
g
ubri
s
ce
n
t
r
a
r
chus
j
urub
i
c
hi
c
h
a
haroldo
i
alt
a
wa
j
ap
i
tuca
p
haiospilu
s
tapi
i
acut
ir
ost
ri
s
geay
i
P
a
r
a
n
a
i
ba
a
n
t
h
u
r
us
cyclostom
a
semi
f
asciat
a
y
p
o
l
enticu
l
ata Co
l
om
b
i
a
h
a
d
rostigm
a
gep
h
yrogramm
a
y
ah
a
w
allac
i
i
vi
ttata
came
t
an
a
cent
i
s
q
uam
a
re
g
an
i
p
unctat
a
macro
p
hthalma M
K
inp
a
j
upiaensi
s
minuan
o
y
j
h
u
i
tigrin
a
iguapin
a
j
egu
i
mandelbur
g
er
i
pl
oeg
i
ce
li
doc
hil
us
sa
x
at
ili
s
b
ri
ts
ki
i
l
epi
d
ot
a
compressiceps
p
ret
a
tend
y
baguass
u
tesa
y
p
rose
l
ytu
s
tesayB
L
taikyr
a
j
a
g
uarensi
s
h
e
m
e
r
a
acut
i
rostr
is
he
m
e
r
a
P
a
r
a
n
a
t
uca
alta
v
ittata
wa
ll
aci
i
macro
ph
t
h
a
l
ma MK
t
a
p
i
i
l
e
p
i
d
ot
a
mu
l
tis
p
inosa
l
enticu
l
at
a
notophthalmus
s
tai
ky
r
a
y
ah
a
iguapin
a
j
egu
i
u
r
ose
m
a
le
n
t
i
culata
Colo
m
b
i
a
h
ec
k
e
li
macro
ph
t
h
a
l
ma
P
c
in
de
r
e
ll
a
j
ohann
a
minuan
o
c
y
c
l
ostoma
pl
oegi
y
jhu
i
t
esa
y
i
g
ar
a
c
e
l
i
d
oc
h
i
l
u
s
hu
wa
j
ap
i
mi
ss
i
o
n
e
ir
a
a
n
thu
r
us
gea
yi
emp
h
ere
s
p
ren
da
i
guassuens
is
h
a
d
rostigm
a
s
axati
l
i
s
tesayBL
l
a
b
rin
a
g
auc
ho
iguassuensisB
L
p
rionogen
y
s
c
entrarc
h
u
s
c
ent
i
squam
a
monogramma
j
aguarensi
s
p
re
t
a
c
ompress
i
cep
s
se
mi
fasc
i
ata
h
aro
ld
o
i
c
ame
t
an
a
y
po
j
upiaensi
s
i
np
a
ti
g
rin
a
l
ugu
b
ri
s
c
h
ic
ha
ph
aiospi
l
us
gep
h
yrogramm
a
p
rose
ly
tu
s
stri
g
at
a
regan
i
j
urubi
scotti
i
p
unc
t
a
ta
l
acust
ri
s
Paranai
ba
m
an
d
e
lb
ur
g
eri
t
end
y
baguassu
b
ri
tsk
ii
r
et
i
culata
Figure 2. Relationships among species of pike cichlids according to the maximum likelihood tree (le), –ln = –58115.602, and the strict
consensus of 384 most parsimonious trees (right), with 12233 steps. Both analyses were based on 216 discrete morphological characters and
5328bp distributed along seven genes (mt-rnr2 [16S], mt-cyb, mt-nd2, mt-nd4, mt-co1, rag2, and rps7). Parsimony analysis performed with all
characters equally weighted. Clade numbers shown above the branches of the ML and parsimony trees. Bootstrap values (in percentage) below
the branches of the ML tree and relative Bremer support (RBS, in percentage) below the branches of the parsimony tree. Asterisk indicates the
type species of Crenicichla (C. macrophthtalma), unique representative of the subgenus Crenicichla (Crenicichla) proposed herein. Coloured
boxes illustrate the composition of each clade as per the nomenclature used throughout the paper; discordant branches in relation to clade
composition among analyses are highlighted in red.
Downloaded from https://academic.oup.com/zoolinnean/advance-article/doi/10.1093/zoolinnean/zlad021/7204395 by Universidade de Sao Paulo (USP) user on 27 July 2023
16 • Var e l l a et al.
Table 6. Summary of the results obtained from the dierent analyses performed in this study. Node number and branch support values (in parentheses) are provided for important clades of each
topology, using ML tree as the main hypothesis for comparisons. BS = bootstrap (percentage); PP = posterior probability (percentage); ABS = absolute Bremer support (integers numbers in the
case of EW analysis and decimal in XIW analysis)
DiscreteMatrixTE ContinuousMatrixTE
Ml tree BI tree Pars EW Pars XIW Pars EW Pars XIW
Clades 96 nodes* 90 nodes 82 nodes 91 nodes 91 nodes 91 nodes
Geophagini including pike
cichlids
111 (BS 100) 109 (PP 100) 108 (ABS 8/RBS 8) 109 (ABS 0.6/RBS 25) 109 (ABS 8/RBS 6) 109 (ABS 0.7/RBS 78)
Closest geophagini taxa 116 (BS 98) 115 (PP 100) 111 (ABS 8/RBS 8) 113 (ABS 0.6/RBS 55) 112 (ABS 8/RBS 6) 113 (ABS 0.6/RBS 55)
Clade of pike cichlids
(subtribe Crenicichlina)
124 (BS 100) 123 (PP 100) 120 (ABS > 72/RBS 100) 124 (ABS 1.9/RBS 84) 122 (ABS > 60/RBS 100) 125 (ABS 1.8/RBS 84)
Clade A 132 (BS 96) 130 (PP 100) 128 (ABS 1/RBS 17) 134 (ABS 0.1/RBS 25) 130—excluding Hemeraia (ABS
2/RBS 35)
133—excluding Hemeraia
(ABS 0.2/RBS 32)
Clade Wallaciia + Teleocichla 184 (BS 99) 180 (PP 100) 168 (ABS 9/RBS 50) 179 (ABS 0.1/RBS 25) 177 (ABS 8/RBS 80) 179 (ABS 0.2/RBS 32)
Wallaciia 183 (BS 71) 179 (PP 100) Not recovered: alt. Wallaciia
167 (RBS 29)
178 (ABS 0.04/RBS 16) Not recovered: alt. Wallaciia 176
(RBS 17)
178 (ABS 0.1/RBS 22)
Teleocichla 191 (BS 100) 186 (PP 100) 176 (ABS > 72/RBS 100) 186 (ABS 0.6/RBS 68) 186 (ABS > 60/RBS 100) 186 (ABS 0.6/RBS 71)
Clade Hemeraia/Saxatilia/
Lugubria
131 (BS 96) 129 (PP 99) 127 (ABS 1/RBS 17) 133 (ABS 0.2/RBS 64) Not recovered (Hemeraia ex-
cluded)
Not recovered (Hemeraia
excluded)
Hemeraia 148 (BS 98) 146 (PP 100) 139 (ABS 5/RBS 63) 145 (ABS 0.1/RBS 71) 144 (ABS 3/RBS 70) 145 (ABS 0.2/RBS 84)
Clade Saxatilia + Hemeraia 141 (BS 79) 140 (PP 83) polytomy Not recovered Not recovered Not recovered
Saxatilia 140 (BS 100) 139 (PP 100) 134 (ABS 3/RBS 43) 140 (ABS 0.1/RBS 71) 137 (ABS 3/RBS 44) 139 (ABS 0.2/RBS 98)
Lugubria 130 (BS 100) 128 (PP 100) 126 (ABS 5/RBS 21) 131 (ABS 0.3/RBS 93) 128 (ABS 5/RBS 22) 131 (ABS 0.3/RBS 75)
Crenicichla 123 (BS 99) 122(PP 100) 119 (ABS 3/RBS 43) 123 (ABS 0.2/RBS 29) 120 (ABS 2/RBS 24) 123 (ABS 0.2/RBS 32)
subgenus Crenicichla
(Crenicichla)
122 (BS 100) 121 (PP 100) 118 (ABS 5/RBS 63) 120 (ABS 0.1/RBS 89) 119 (ABS 3/RBS 70) 120 (ABS 0.1/RBS 87)
Clade subgenera
Batrachops + Lacustria
128 (BS 91) polytomy 124 (ABS 2/RBS 9) Not recovered 126 (ABS 1/RBS 4) Not recovered
subgenus Crenicichla
(Batrachops)
133 (BS 88) 131 (PP 64) 129 (ABS 3/RBS 20) 149—exluding C. jegui (ABS
0.2/RBS 29)
131 (ABS 1/RBS 5) 149—exluding C. jegui
(ABS 0.2/RBS 39)
subgenus Crenicichla
(Lacustria)
127 (BS 90) 126 (PP 91) 123 (ABS 3/RBS 20) 128 (ABS 0.2/RBS 29) 125 (ABS 1/RBS 5) 129 (ABS 0.2/RBS 5)
C. missioneira complex 153 (BS 100) 152 (PP 100) 145 (ABS 2/RBS 20) 154 (ABS 0.2/RBS 78) 152 (ABS 4/RBS 37) 154 (ABS 0.2/RBS 75)
C. mandelburgeri complex 163 (BS 84) 160 (PP 58) 150 (ABS 1/RBS 17) 160 (ABS 0.05/RBS 48) 159 (ABS 1/RBS 1) 160 (ABS 0.1/RBS 48)
Clade C. scoii complex 179 (BS 100) 175 (PP 100) 165 (ABS 8/RBS 100) 174 (ABS 0.2/RBS 100) 174 (ABS 7/RBS 100) 174 (ABS 0.2/RBS 100)
Downloaded from https://academic.oup.com/zoolinnean/advance-article/doi/10.1093/zoolinnean/zlad021/7204395 by Universidade de Sao Paulo (USP) user on 27 July 2023
New generic classication of pike cichlids • 17
group, therefore creating a long-needed phylogenetic framework to
revise the taxonomy. e goal of the study, rather than presenting
competing hypotheses to those generated by recent phylogenomic
studies, is to provide a relatively stable generic and subgeneric clas-
sication of pike cichlids based on morphological diagnosis of
clades supported by both molecular and morphological data. To
that end, we reviewed the relationships among species of pike cich-
lids from previous studies in comparison to ours (compiled in Fig.
3) and identied the most concordant elements across topologies
as the basis to propose a new classication of the clade and provide
diagnostic characters for each group of pike cichlids.
Monophyly and phylogenetic position of pike cichlids (tribe
Geophagini: subtribe Crenicichlina)
All our analyses recovered a monophyletic group comprising all
the species formerly classied in Crenicichla and Teleocichla (ML
Node 124; see also Table 6). None of our resulting hypotheses
based on the combined datasets have encountered a separate
Teleocichla (i.e. as sister-group of all remaining species of pike
cichlids), which would support the concept of a large genus
Crenicichla. us, since Teleocichla is monophyletic and still con-
sidered a valid generic name, the genus Crenicichla as formerly
dened should be considered paraphyletic.
All resulting hypotheses also support the placement of the
clade of pike cichlids within the tribe Geophagini, as the most
recent studies on Neotropical cichlids have demonstrated
(Farias et al. 2000; Sparks and Smith 2004; López-Fernández et
al. 2005a,b, 2010, 2012; Smith et al. 2008; Ilves et al. 2018), ra-
ther than as sister-group relationship to Cichla (Stiassny 1981;
Kullander 1998). erefore, we refer to the clade containing all
pike cichlids as the subtribe Crenicichlina.
e subtribe Crenicichlina shows high support in all ana-
lyses (ML-BS 100%, BI-PP 100%, parsimony ABS > 72/
RBS 100%; Table 6) and a large number of synapomorphies
eleocichla
T
T
all 9 valid s
pp
.
Crenicichla
(
Crenicichla
)
2
t
e
rmin
a
l t
a
x
a
Lu
g
ubria
9
terminal tax
a
8 valid s
p
ecie
s
Crenicichla
(
Lacustria
)
32
terminal tax
a
2
8 valid s
p
ecies
He
m
e
r
a
i
a
2
terminal taxa/valid s
pp
.
W
allaciia
W
W
6 terminal taxa/valid s
pp
.
Crenicichla
(
Batrachops
)
6 terminal taxa/valid s
pp.
S
axatilia
8 terminal taxa
/
valid spp
.
T
eleocichla
T
T
5 valid s
pp
.
W
allaciia
W
W
1
4 valid s
pp.
“C. lugubris group”
8 valid species
1
9 valid s
pp
.
Crenicichla
(
Crenicichla
)
1
valid spp
.
C
renicichla
(
Lacustria
)
2
8 valid s
p
ecie
s
Crenicichla
(
Batrachop
s
)
8 valid s
pp.
CLADEA
CLADE A
CLADEA
CLADE A
CLADEA
T
eleocichla
T
T
5 valid s
p
ecie
s
Lu
g
ubria
1
3 terminal taxa/valid spp.
,
i
ncludin
g
C. jegui
a
n
d
i
C. vittata
L
acus
tri
a
9
terminal taxa
8
valid spp.
“
W
allaciia 1”
W
W
C. regan
i
, C. wallaci
i
, C. noto
p
hthalmu
s
Batracho
p
s
10
terminal tax
a
8
valid spp., includin
g
C. scotti
i
S
axatilia
2
7 terminal taxa/24 valid s
pp
.,
i
ncluding C. hemer
a
“
W
allaciia 2”
W
W
C. vir
g
atula, C. urosema
C. com
p
ressice
p
s, C. heckei
i
T
eleocichla
T
T
2
terminal taxa/valid s
pp.
W
allaciia
W
W
4 t
e
rmin
a
l t
a
x
a
3
valid spp
.
S
axatili
a
4 t
e
rmin
a
l t
a
x
a*
2
valid s
pp.
Crenicichla
(
Batrachops
)
2
terminal taxa/valid spp
.
Crenicichla
(
Lacustria
)
1
2 terminal taxa/9 valid s
pp.
C
renicichl
a
(
C
renicichla
)
C
renicichl
a
(
C
. macro
p
hthalma
)
T
eleocichla
T
T
2
terminal taxa/valid s
pp.
Lu
g
ubri
a
5
terminal tax
a
4 valid spp
.
W
allaciia
W
W
3
terminal tax
a
2
valid s
pp.
S
axatili
a
2
terminal taxa*/2 valid spp
.
Crenicichla
(
Batrachops
)
5 terminal tax
a
4 valid s
pp.
Crenicichla
(
Lacustria
)
~23 valid s
p
ecies*
GENUS
Crenicichla
GENUS
Crenicichla
GENUS
Crenicichla
GENUS
Crenicichla
GENUS
Crenicichla
GENUS
Crenicichla
GENUS Crenicichla
C. macrophthalma
not included
CLADEA
C
renicichl
a
(
C
renicichla
)
Crenicichla
(
Lacustria
)
1
0 terminal taxa/valid s
pp
.
Crenicichla (Batrachops)
2
terminal taxa/valid s
pp.
T
eleocichla
T
T
3
terminal tax
a
2
valid s
pp.
Lugubri
a
9
terminal tax
a
8 valid s
pp.
W
allaciia
W
W
3 terminal taxa
/
valid s
pp.
S
axatili
a
5 terminal taxa/valid s
pp.
C
renicichl
a
(C
r
e
ni
c
i
c
hl
a
)
Crenicichla (Lacustria)
1
0 terminal taxa/valid s
pp.
Crenicichla
(
Batrachops
)
2
terminal taxa/valid s
pp
.
T
eleocichla
T
T
3
terminal taxa
2
valid s
pp
.
Lugubria
9
terminal taxa
8
valid s
pp
.
W
allaciia
W
W
3
terminal taxa/valid s
pp
.
S
axatilia
5
terminal taxa/valid s
pp
.
C
renicichl
a
(
C
renicichl
a
)
C
renicichl
a
(
C
renicichla
)
Crenicichla (Lacustria
)
2
5 terminal taxa/valid s
pp.
Crenicichla
(
Batrachops
)
6 terminal taxa/valid spp.
T
eleocichla
T
T
5 terminal tax
a
4 valid s
pp.
Lugubri
a
12
t
e
rmin
a
l t
a
x
a
9 valid s
pp.
W
allaciia
W
W
3 terminal taxa/valid s
pp
.
S
axatili
a
12
t
e
rmin
a
l t
a
x
a
1
1
valid s
pp
.
Crenicichla
(
Lacustria
)
2
valid s
pp.
C
. minuano
a
n
d
C. hadrosti
g
m
a
Crenicichla
(
Batrachops
)
- onl
y
C. gea
yi
T
eleocichla
T
T
3 terminal taxa/valid s
pp
.
Lu
g
ubri
a
4 terminal taxa
/
valid spp.
W
allaciia
W
W
3
terminal taxa
2
valid sp
p
S
axatilia
2
terminal taxa/valid s
pp
.
AB
C
DE
F
G
H
I
“C. acutirostris group”
4 valid species
Figure 3. Synthesis of the phylogenetic relationships among species of pike cichlids according to the summary of our hypotheses (ML
and parsimony trees) and the hypotheses from previous papers. Colour of triangles illustrates the composition of each clade as per the
nomenclature used in this paper and the size of the triangles illustrates the number of terminal taxa included in each analysis. Incongruent
branches aecting the composition of the main clades of pike cichlids are highlighted in red. A, summary of ML tree of this paper; B, Říčan et
al. (2021a): BI tree using BEAST, based on mt-cyb and mt-nd2 sequences (2163 aligned base pairs); C, Ilves et al. (2018): species tree obtained
with ASTL-II (summary coalescence method) for 415 exons (471448bp); D, E, Burress et al. (2017): ML trees based on supermatrices
of 80% completeness (427 UCEs/245084bp) and of 95% completeness (248 UCEs/160654bp), respectively; F, Burress et al. (2018):
ML tree based on 25128 variable sites from the ddDseq study; G, Piálek et al. (2012): BI tree based on four loci (3190bp); H, Kullander
et al. (2010): majority-rule consensus tree from a parsimony analysis of mt-cyb (1137bp); I, Ploeg (1991): phylogenetic diagram from the
interpretation of gures 169–175 based on implicit parsimony argumentation without explicit method. Subgenus Crenicichla (Crenicichla)
as proposed herein is represented only by C. macrophthalma, type species of Crenicichla. Coloured boxes illustrate the composition of each
clade as per the nomenclature used throughout the paper and important discordant branches in relation to the composition of these clades are
highlighted in red.
Downloaded from https://academic.oup.com/zoolinnean/advance-article/doi/10.1093/zoolinnean/zlad021/7204395 by Universidade de Sao Paulo (USP) user on 27 July 2023
18 • Var e l l a et al.
(63 synapomorphies, 40 of them non-ambiguous, shown in
Table 7). Character optimizations on the ML tree and on the
parsimony tree are largely congruent and provide equivalent
synapomorphies for the clade. Of the 40 unambiguous syn-
apomorphies of the ML tree, 33 are likewise unambiguous and
one ambiguous in the parsimony tree optimization. Of 23 am-
biguous synapomorphies from the optimization on the ML tree,
eight are congruent with unambiguous synapomorphies and 10
with ambiguous synapomorphies on the parsimony tree.
Table 7. List of synapomorphies of the clade of pike cichlids (subtribe Crenicichlina; Node 124) based on the optimisation of the characters
on the ML tree (Figs 1A, 2A)
Character transformation Apomorphic condition
Char. 0: 0=>1 Number of scales in the E1 series increasing to 31–38.
Char. 1: 0=>1 Anterior branch of lateral line with 20–31 scales.
Char. 3: 1=>3 Dorsal n with 20-22 spines.
Char. 15: 0=>1 Frontal bones laterally compressed (shorter interorbital distance in relation to the neurocranial length).
Char. 16: 0=>2 More than four columns of scales on the postorbital are of the head.
Char. 31: 0=>1 Lateral line scales longer than the adjacente scales situated ventrally.
Char. 36: 1=>0 Absence of a ‘large interpelvic scale’ on the area between pelvic-n insertions.
Char. 47: 0=>2 Pelvic n rounded, with the second ray longest.
Char. 50: 0=>1 Pectoral n symmetrical, with median rays longest.
Char. 52: 1=>2 Caudal-n blotch at the same level or slightly dorsally to the posterior branch of the lateral line and dis-
placed more posteriorly.
Char. 63: 0=>1 Preorbital dark stripe present.
Char. 91: 0=>1* Lower jaw distincly prognathous.
Char. 105: 1=>0* Infraorbitals 4 and 5 autogenous.
Char. 108: 0=>1 Infraorbital 7 (dermosphenotic) curved, with dorsal opening directed posteriorly.
Char. 111: 0=>1 Absence of supraneural.
Char. 115: 0=>1 Absence of vertebral hypapophysis.
Char. 120: 0=>1 Absence of parhypurapophysis on the last caudal vertebrae.
Char. 123: 0=>1 Presence of ventral postexapophysis on the posteriormost precaudal vertebrae.
Char. 139: 0=>1 Insertion of the posterior palatovomerine ligament on the anteromedial portion of the ectopterygoid or
on the articulation between ecto- and endopterygoid (vs. on the posteromedial side of the palatine).
Char. 141: 0=>1 Dermal splint of the palatine reduced, not reaching the rostral edge of ectopterygoid.
Char. 142: 1=>0 Maxillary process of palatine dorsoventrally compressed (vs. not compressed, approximately cylin-
drical).
Char. 143: 0=>1 Absence of a posteroventral laminar expansion of the palatine.
Char. 145: 0=>1 Presence of a medial laminar expansion on the posterior portion of the palatine.
Char. 148: 0=>1 Metapterygoid narrow, stick-like, with anterodorsal laminar expansion very reduced.
Char. 156: 0=>2 Proximal and distal extrascapulas well-separated.
Char. 157: 0=>1 Distal postcleithrum with a spinuous process directed anteriorly.
Char. 171: 0=>1 Wide precomissural bridge of prootic.
Char. 173: 0=>1 Nasal pore connected to the canal through a long non-ossied tube.
Char. 174: 1=>0* Rostral pore of the nasal on the postlabial margin of snout.
Char. 179: 0=>1 Pharyngobranchial 1 cartilaginous.
Char. 191: 2=>1* External face of the lower rst gill arch with well-ossied, tubercular or triangular, gill rakers.
Char. 192: 1=>0* Presence of teeth on the gill rakers of the external face of ceratobranchial 1.
Char. 193: 0=>1 Absence of interarcual cartilage.
Char. 195: 2=>3 Anteroventral laminar expansion of the epibranchial 2 very reduced without cartilage on its ventral
border.
Char. 200: 1=>0* Presence of unicuspid teeth on the gill rakers of the external face of ceratobranchial 4.
Char. 204: 0=>1 Presence of gill rakers laterally to the lower pharyngeal jaw.
Char. 210: 0=>2 Lateral wings of the urohyal distinctly wider than the depth of medial crest.
Char. 212: 0=>1 Glossohyal relatively long and only slitghly compressed dorsoventrally (vs. distinctly compressed).
Char. 213: 0=>1 Anterior tips of the basipterygia parallel, diverging (vs. converging).
Char. 214: 0=>1 Anterior tip of the basipterygium not surpassing the cleithrum (vs. extending anteriorly to the
cleithrum).
Asterisks represent reversals to plesiomorphic condition of Neotropical cichlids.
Downloaded from https://academic.oup.com/zoolinnean/advance-article/doi/10.1093/zoolinnean/zlad021/7204395 by Universidade de Sao Paulo (USP) user on 27 July 2023
New generic classication of pike cichlids • 19
Apomorphic conditions of characters 111, 115, 120, 139 to
145 (palatine conguration), 173, 179, 193, 195, 213, 214, de-
rived from Kullander’s (1988) characters 1–10, are recovered as
synapomorphies of Crenicichlina. Also, apomorphic conditions
related to the palatine resulted as synapomorphies of pike cich-
lids in López-Fernández et al. (2005a, 2012: character 90 state
2 and character 92 state 4), corresponding to characters 141
and 145 herein. Finally, our results agree with several synapo-
morphies listed for Crenicichla in Kullander (1998), even though
that study recovered it as sister-group to Cichla viz. characters
204 (Kullander’s character 18 state 1), 193 (Kullander’s char-
acter 22 state 1), 141 (Kullander’s character 50 state 1or 2), 140
(Kullander’s character 58 state 1), 111 (Kullander’s 66 state 1),
and 115 (Kullander’s character 77 state 0).
Tribe Geophagini is a well-supported clade in the model-
based analyses (ML-BS 100%; BI-PP 100%) but received
relatively lile support in the parsimony analysis based on
EW/DiscreteMatrixTE (ABS 8/RBS 8%; Table 6). Based on
optimizations on the ML tree, Geophagini is supported by 28
molecular and 17 morphological synapomorphies (Table 8). We
conservatively consider all morphological synapomorphies as
tentative because the sister-group relationship between the tribe
Geophagini and Chaetobranchus avescens (Chatobranchini)
in our dataset is based only on molecular data and so the mor-
phological synapomorphies for Geophagini are all ambiguous.
Optimisation on the parsimony trees resulted in 27 mor-
phological synapomorphies (15 of them unambiguous) for
Geophagini, most of them congruent with the synapomorphies
for the ML tree. Characters 8, 9, 23, 39, 112, 163, 174, 191,
192, and 198 are recovered as unambiguous synapomorphies
and 55, 91, 114, 125, 149, 181, and 200 as ambiguous synapo-
morphies through the optimisation on the parsimony trees.
On the other hand, ve of the synapomorphies from the opti-
misation on the parsimony trees are not recovered as synapo-
morphies of Geophagini, but as synapomorphies of the clade
Geophagini + Chaetobrachus avescens on the ML tree (Node
112). From these, one character deserves aention (char. 187,
with transformation 0=>1 A): presence of a eshy pad associ-
ated with the epibranchials 1 and/or 2. is corresponds to the
so portion of a complex structure generally called ‘epibranchial
lobe’ of geophagines, which is extremely variable in relation to
its development and the association between so pad and the
cartilage or bone of the anterior epibranchials (which can also
be variably developed). is eshy pad is absent in Ch. avescens
and, most probably, its presence would be optimised as another
synapomorphy of Geophagini in the ML tree if morphological
data of Ch. avescens were included.
Among the aforementioned apomorphic conditions, a re-
duced number of concavitites on the pharyngeal plate of the
pharyngobranchial 4 (character 198, state 2) was already
pointed out by Kullander (1998: character 21, state 4) and
López-Fernández et al. (2005a: character 113, state 1) as syn-
apomorphy of the tribe Geophagini. Kullander’s (1998) charac-
ters 4 (state 1), 17 (state 2, partially), and 63 (state 2) were also
recovered as synapomorphies of Geophagini (characters 181,
192, and 125 herein).
e main hypothesis (ML tree) and the BI tree (both model-
based hypotheses) agrees with the most recent studies on
Neotropical cichlids (e.g. López-Fernández et al. 2010; Ilves et
al. 2018) in recovering the sister-group relationship between
pike cichlids (subtribe Crenicichlina) and a clade formed by
Acarichthys and Biotoecus. e subsequent sister-groups are the
Apistogrammines and the Guianacarines sensu Ilves et al. (2018),
forming Node 116 (ML tree) and Node 115 (BI tree). Parsimony
Table 8. List of synapomorphies of the tribe Geophagini (Node 111) based on the optimisation of the characters on the ML tree (Fig. 1, le)
Character transformation Apomorphic condition
Char. 8: 0=>2 D ree or four vertebrae contained within the caudal peduncle.
Char. 9: 1=>0 A Caudal n with more than three procurrent rays.
Char. 23: 0=>2 A Absence or up to three scales on a small posteriormost portion of the interopercular area.
Char. 39: 1=>0 A Absence of scales on inter-radial membranes of the basal portion of the anal n.
Char. 55: 1=>0 A Absence of a surrounding light ring of caudal-n dark blotch.
Char. 91: 1=>0 D Jaws isognathous or lower jaw slightly prognathous.
Char. 112: 0=>1 A Two cartilaginous or ossied supraneurals.
Char. 114: 2=>1 A Centrum of the two anteriomost precaudal vertebrae presenting rostro-caudal compression.
Char. 125: 0=>1 D Dorsal apex of the articulating process of the premaxilla not distinct from its ascending process.
Char. 149: 0=>1 D Absence of hyomandicular-metapterygoid suture.
Char. 163: 1=>0 A Caudal opening of the posterior myodome very reduced or absent.
Char. 174: 0=>1 D Rostral pore of the nasal canal displaced posteriorly from the margin of the postlabial snout (vs.
on the postlabial margin of snout).
Char. 181: 1=>0 A Uncinate process of the epibranchial 1 longer than its anterior arm.
Char. 191: 1=>2 D Gill rakers of the external face of the lower rst gill arch ossied only basally or not ossied,
papilliform.
Char. 192: 0=>1 A Absence of teeth on the gill rakers of the external face of the ceratobranchial 1.
Char. 198: 1=>2 A One concavity on the frayed zone of the caudal edge of the pharyngeal tooth plate of the
pharyngobranchial 4.
Char. 200: 0=>1 A Absence of unicuspid teeth on the gill rakers of the external face of the ceratobranchial 4.
Downloaded from https://academic.oup.com/zoolinnean/advance-article/doi/10.1093/zoolinnean/zlad021/7204395 by Universidade de Sao Paulo (USP) user on 27 July 2023
20 • Var e l l a et al.
topologies also recover a large group including pike cichlids,
Biotoecus, Acarichthys, Apistogrammines, and Guianacarines, but
relationships among them are either unresolved or incongruent
with the model-based topologies.
Overview of the relationships within subtribe Crenicichlina
Our main phylogenetic hypothesis, based on the ML tree (Fig.
2, le) recovers the major species groups proposed in the taxo-
nomic and phylogenetic studies of pike cichlids (subtribe
Crenicichlina), and divides them into two main clades (Nodes
132 and 123). Node 133, hereaer Clade A, includes species for-
merly placed in Teleocichla (orange) and the Crenicichla wallacii
(red), Crenicichla lugubris (yellow), and Crenicichla saxatilis
(light blue) groups. It also includes the Crenicichla chicha–C.
hemera group (green clade), which was not identied as distinct
in previously proposed groupings. Node 123 corresponds to
Crenicichla s.s., comprising C. macrophthalma (type species of the
genus Crenicichla) and species formerly placed in the Crenicichla
reticulata (purple) and Crenicichla lacustris (lilac) groups.
Both clades have high support in the ML and BI trees (BS
96–99%; PP 99.6%; Table 6). In the consensus tree based on
the parsimony-EW/DiscreteMatrixTE analysis (Fig. 2, right),
Clade A shows low support (ABS 1/RBS 17%) and the genus
Crenicichla s.s. moderate support (ABS 3/RBS 43%; Table 6).
e dichotomy between Clade A and Crenicichla s.s. is also found
in the topologies from XIW/DiscreteMatrixTE (Supporting
Information, Appendix S3, Fig. S3). Topologies from the par-
simony analyses of the combined ContinuousMatrixTE under
EW and XIW schemes (Supporting Information, Appendix S3,
Figs S2, S4) also show the two main clades, but the group C.
hemera–C. chicha is not part of Clade A and instead is recovered
as sister-group of all remaining pike cichlids. e dichotomy be-
tween Clade A and Crenicichla s.s. was also found in most pre-
vious phylogenetic analyses using dierent sources of molecular
data (Piálek et al. 2012; Burress et al. 2017, 2018; Ilves et al.
2018) as illustrated in Figure 3. Relationships among taxa within
subclades (coloured boxes on the topologies illustrated in Figure
2) are not as stable and can slightly vary between multiple top-
ologies obtained in our study and in previous studies. ese dis-
agreements in relationships within species groups are common
in all previous phylogenetic studies of pike cichlids and resolving
them will require species-level analyses beyond the scope of this
study.
Taxonomic section
A revised classication of the subtribe Crenicichlina must start
with resolving a nomenclatural inconvenience inadvertently
caused by a choice made by Eigenmann and Bray (1894) over a
century ago when they selected Crenicichla macrophthalma (in-
stead of C. lepidota) and Batrachops reticulatus as the type species
of the genera proposed by Heckel (1840). Crenicichla lepidota
has been clearly resolved as part of the C. saxatilis group since
Kullander (1982); with 23 valid species (Varella et al. 2018),
the C. saxatilis group has been recovered as monophyletic in all
phylogenetic studies. In contrast, the closest phylogenetic rela-
tionships of C. macrophthalma remain controversial and the spe-
cies has not been placed in any of the proposed species groups of
pike cichlids due to its distinctive morphology that includes very
large eyes and almost uniform body coloration, lacking other-
wise useful colour paern elements and characters related to
sexual dimorphism.
Our hypothesis based on the ML and parsimony trees shown
in Figure 2 shows C. macrophthalma (grey clade) as sister-group
to the clade formed by the C. reticulata (purple) and C. lacustris
(lilac) groups. Our BI tree indicates a polytomy among the three
groups and parsimony analyses of the combined datasets sug-
gested a sister-group relationship between C. macrophthalma
and the C. reticulata group, but also resulted in a clade formed by
C. macrophthalma, and the C. lacustris and C. reticulata groups. In
combination with recent molecular phylogenies of pike cichlids
(Piálek et al. 2012; Burress et al. 2017, 2018; Říčan et al. 2021a),
it is clear that C. macrophthalma forms a clade with the C. lacustris
and C. reticulata groups, even though the relationships among
the three are still unclear.
In this context, our rationale for a revised classication of
pike cichlids consisted of the following arguments: (i) As a
guiding principle, we maintain that a classication based on
less-inclusive, valid taxonomic ranks and standardized diag-
noses is more eective in fostering further taxonomic and bio-
logical studies of pike cichlids than informal and sometimes
conictingly dened species groups. (ii) Keeping well-supported
monophyletic groups as informal ‘species groups’ of Crenicichla
postpones an inevitable decision on whether to synonymize the
valid and strongly monophyletic genus Teleocichla. (iii) Similar
to the case of Teleocichla, other species groups of pike cichlids
are strongly monophyletic and, by virtue of being part of Clade
A, not directly related to C. macrophthalma, while sharing a
similar phylogenetic hierarchy with it and other species groups
within the clade Crenicichla s.s.. As such, either all pike cichlid
taxa should be considered Crenicichla based on one of the two
clades, thus synonymizing the monophyletic Teleocichla and
failing to recognize the monophyly of several other subclades
within Clade A, or the genus Crenicichla should be limited to the
clade Crenicichla s.s. thus recognizing the phylogenetic hierarchy
of Clade A, Teleocichla and its other monophyletic subclades.
We nd that the laer alternative provides a taxonomically
clearer and biologically more informative classication. Cichlids
are increasingly used as models for evolutionary studies and the
dierent clades of pike cichlids delimit signicant evolutionary,
ecological, and biogeographic units that are otherwise hidden by
keeping all 102 valid species as Crenicichla s.l..
erefore, we redene Crenicichla s.s. as a monophyletic
genus including C. macrophthalma, and the C. lacustris and C.
reticulata groups based on widely concordant phylogenetic re-
lationships between our study and previous analyses (see com-
parisons in Fig. 3). We also reclassify the monophyletic species
groups within Clade A by elevating them to separate genera, thus
preserving the genus-level validity of Teleocichla and avoiding
the establishment of a paraphyletic Crenicichla. We refrain from
extending our phylogenetic classication to within the newly
dened genera until relationships among their species are fur-
ther claried. Based on these considerations, we propose the
following revised classication of pike cichlids (Node 124),
following our main phylogenetic hypothesis (ML tree: Fig. 2,
le). New generic and subgeneric names were adapted from the
oldest nominal species in each group.
Downloaded from https://academic.oup.com/zoolinnean/advance-article/doi/10.1093/zoolinnean/zlad021/7204395 by Universidade de Sao Paulo (USP) user on 27 July 2023
New generic classication of pike cichlids • 21
Subtribe Crenicichlina
Node 132—Clade A
Node 184
Node 183 (red clade)—Wallaciia new genus
Node 191 (orange)—Teleocichla Kullander 1988
Node 131
Node 130 (yellow)—Lugubria new genus
Node 141
Node 148 (green)—Hemeraia new genus
Node 140 (light blue)—Saxatilia new genus
Node 123—Crenicichla Heckel 1840
Node 122 (grey)—Crenicichla (Crenicichla) new subgenus
Node 128
Node 133 (purple)—subgenus Crenicichla
(Batrachops) Heckel 1840
Node 127 (lilac)—Crenicichla (Lacustria)
new subgenus
As discussed in detail through the next sections of this paper, the
subgenus Lacustria of Crenicichla is restricted to Southern drain-
ages of South America (southern and south-east Atlantic coastal
drainages, and Río de La Plata basin) but shows great ecological
and morphological diversity. Indeed, this group has been the
focus of studies on evolution, mainly related of the species
ocks in the Río Uruguay and Río Parana basins (e.g. Pialek et
al. 2019a; Burress et al. 2022) and is acquiring a very prominent
status as such model. In contrast, the subgenus Crenicichla (i.e.
Crenicichla macrophthalma) is Amazonian and almost all species
of the subgenus Batrachops of Crenicichla are distributed in the
Amazon, Orinoco, and Guianas basins. e only exception is
Crenicichla (Batrachops) semifasciata, which is restricted to the
Río de La Plata basin. Members of these subgenera, Crenicichla
and Batrachops, have morphologies very dierent from the sub-
genus Lacustris but, considering that the position of Crenicichla
macrophthalma is still uncertain, it makes sense to consider the
genus Crenicichla, as proposed herein, as a unit for further studies
on biological, biogeographical, and evolutionary aspects.
Teleocichla is formed by small rheophilic species restricted
to clear-water tributaries of the Amazon basin, in contrast with
other dwarf pike-cichlids of Wallaciia, with most species widely
distributed in tributaries in the Amazon basin, Guianas, and
Orinoco. e genus Lugubria comprises easily diagnosable, large
species distributed in the Amazon basin, Guianas, and Orinoco,
which undergo drastic colour-changing during ontogeny, mostly
related with sexual maturity. Species of Saxatilia are relatively
conserved in general shape and, presumably, in function, and
occur in most rivers of South America east of the Andes.
A specic-level taxonomic revision of all genera is beyond
the objectives of this paper. However, the long-needed nomen-
clatural acts and taxonomic reorganization we performed are a
useful backbone for subsequent taxonomic studies inside each of
the genera and subgenera. It will also help in technical studies, as
inventories of ichthyofauna, ocial assessments about status of
conservation of the species, and curation of collections.
A rened classication reduces the universe of uncertainty
for the identication of the jacundás and improves the notion of
general biodiversity of South American rivers. It is common to
nd published checklists and museums’ collections with several
‘Crenicichla sp.’, because it is already implicit that it is too dicult
to identify one or few species in a universe of 102 valid species.
ese many lots of unidentied species usually include more
than one species belonging to dierent groups that are easily
distinguishable from each other. For example, it is common to
nd in collections of Southern rivers of South America, a mix
of species of both the C. lacustris group (=subgenus Lacustria of
Crenicichla) and C. saxatilis group (=genus Saxatilia). In collec-
tions of the Amazon basin, more groups are usually mixed, such
as C. reticulata group (=subgenus Batrachops of Crenicichla),
C. wallacii group (=Wallaciia), and C. lugubria (=Lugubria).
Species of Teleocichla, which has been considered a valid genus
since its description in 1988, are normally correctly identied at
the genus level and set apart from the rest of pike cichlids. We
hope that a rened classication, with a standardized diagnosis
for each genus and subgenus, will certainly help (and probably
foster) identication of material of jacundás at species level in
checklists, other technical studies, and museums.
To contribute to subsequent taxonomic work and facilitate the
use of our proposed classication, Table 1 lists all nominal and
valid species of pike cichlids and classies them by genus (and
subgenus) as proposed herein. Below, we oer diagnoses and an
approximate range of distribution for each of these groups.
Node 123—genus Crenicichla Heckel 1840
Crenicichla Heckel, 1840: 416 (type species Crenicichla
macrophthalma Heckel, 1840, designated subsequently by
Eigenmann and Bray 1894: 620).
As mentioned above, the genus Crenicichla as redened
herein presents high node support in the ML and BI trees,
and moderate support in the main parsimony analysis (EW/
DiscreteMatrixTE) (Table 6). e clade is supported by few
unambiguous synapomorphies that correspond to homoplastic
characters (Table 9) but is supported by 32 molecular trans-
formations. e same morphological synapomorphies are found
from the optimization on the EW/DiscreteMatrixTE pasimony
tree.
As redened herein, Crenicichla comprises 43 valid species div-
ided into three subgenera: Crenicichla (one species), Batrachops
(nine species), and Lacustria (33 species); see Table 1. e genus
Crenicichla shows great morphological and ecological diversity,
which challenges the formulation of a good diagnosis for the
group. us, we make use of a combination of several character-
istics to distinguish the species of Crenicichla from the other pike
cichlids, organized by morphological groups.
Diagnosis: Al l species of Crenicichla, except C. macrophthalma, can
be distinguished from other pike cichlids by sexually dimorphic
characters that include orange or reddish marking on the lat-
eral abdomen of mature females. Apart from C. macrophthalma,
which is known to lack such dimorphism, other species for
which there is no information on the coloration of mature fe-
males probably show this feature (see the distribution of states of
character 89). In most species of Crenicichla, female marking on
the lateral abdomen is accompanied by one or more blotches on
the dorsal n; in some taxa, these blotches are eventually modi-
ed into a horizontal dark bar. W hile this characteristic is shared
Downloaded from https://academic.oup.com/zoolinnean/advance-article/doi/10.1093/zoolinnean/zlad021/7204395 by Universidade de Sao Paulo (USP) user on 27 July 2023
22 • Var e l l a et al.
with some species of Wallaciia, it is absent in the remaining spe-
cies of pike cichlids.
All species of Crenicichla are distinguished from most species
of Saxatilia and Lugubria by the absence of a broad reddish or
purplish pigmentation on the ventral part of the abdomen of
gravid females. All species of Crenicichla are additionally distin-
guished from Saxatilia by the absence of a humeral blotch and of
sexual dimorphism consisting of scaered light spots on mature
males—Crenicichla (Batrachops) jegui has light dots scaered on
the anks and head but their presence is not sexually dimorphic.
All species of Crenicichla are distinguished from Lugubria by
having fewer than 79 scales in the E1 row, except Crenicichla
(Lacustria) viata with 79–93 scales (vs. 79–123 scales in the
species of Lugubria).
All species of Crenicichla have all post-lachrymal infraorbitals
autogenous vs. infraorbitals 4 and 5 fused, forming a median pore
in Teleocichla and Hemeraia. Exceptions are the unique specimen
of C. (La.) igara examined herein, which shows infraorbitals 4
and 5 partially fused, and C. (La.) jupiaensis, with only four in-
stead of ve infraorbitals but without signal of co-ossication
(see: Varell a et al. 2018). Species of Crenicichla are additionally
distinguished from Hemeraia by having most of the ank scales
ctenoid [i.e. a combination of paerns 1 and B2 in all Crenicichla
except C. (C.) macrophthalma, which shows a combination
of paerns 0 and B1] vs. most of the ank scales cycloid (i.e.
combination of paern 3 and B3). Species of Crenicichla are
additionally distinguished from Teleocichla by having rounded
pelvic n with the second ray longest (vs. pointed pelvic n with
the third ray longest), by having a stick-like pharyngobranchial
1 (vs. globular pharyngobranchial 1), by showing the cong-
uration of the urohyal similar to other pike cichlids, i.e. lateral
wings wider than the depth of the medial crest (vs. lateral wings
wide but medial crest rudimentary or absent), and by having the
symmetrical medial processes of the basipterygia diverging an-
teriorly (vs. running very close, not diverging anteriorly).
All species of Crenicichla are also distinguished from all spe-
cies of Wallaciia, except W. h eckel i , by the absence of serrations
on the posterior margin of the supracleithrum. Additionally,
Crenicichla comprises medium-sized species (max. SL
95–294mm) with relatively small eyes (orbital diameter 4.6–
10.9% of SL, with minimum between 4.6–7.5%) and clearly
showing a negative allometry of eye size (eyes decreasing pro-
portionally in size with ontogeny), with the exception of C.
macrophthalma, a medium-sized species (max. SL 200 mm)
with large eyes in small to large specimens (orbital diameter 9.4–
10.6% of SL), with less marked negative allometry in eye size.
us, Crenicichla can be distinguished from Wallaciia, which
comprises only small-sized species (max. SL 52–85mm) with
large eyes (orbital diameter 7.8–12.6% of SL, with minimum
between 7.8–8.6%).
Node 122—Crenicichla (Crenicichla) Heckel 1840—type sub-
genus of Crenicichla.
Type species: Crenicichla macrophthalma Heckel, 1840.
Nominal species: Crenicichla macrophthalma and C. santaremensis
(synonym of C. macrophthalma).
e subgenus is recovered in all analyses performed herein
with high support in ML and BI trees (BS 100% and PP 100%)
and moderate support in the parsimony analysis of EW/
DiscreteMatrixTE (ABS 5/RBS 63%; Table 6). e subgenus
is diagnosed by 50 molecular transformations and 19 morpho-
logical synapomorphies (Table 10). e same morphological
synapomorphies are found from the optimization on the parsi-
monious tree from EW/DiscreteMatrixTE, except for characters
98 and 168 (only in the ML optimization).
Table 9. List of synapomorphies of the genus Crenicichla (Node 123) based on the optimization of the characters on the ML tree (Fig. 2A, le)
Character
transformation
Apomorphic condition Observations
Char. 44: 1=>0 Absence of sexual dimorphism related
to the elongation of the posteriormost
rays of anal n in males.
Sexually dimorphic males with elongated anal ns represent a
common condition of Neotropical cichlids and pike cichlids. ere
are repeated transformations to the absence of this kind of sexual
dimorphism within the subtribe Crenicichlina, also being recovered
as a synapomorphy of Lugubria.
Char. 197: 0=>2 Dorsomedial expansion of the lateral arm
of the epibranchial 4 very reduced, as a
narrow crest, or totally absent.
is character is highly homoplastic, with several transformations
in dierent directions within the subtribe Crenicichlina and with
several secondary transformations to the state 1 within the genus
Crenicichla.
Ambiguous
Char. 8: 2=>3 D Number of the last caudal vertebrae con-
tained within the caudal peduncle increasing
to ve or more + halfcentrum.
It represents a common modication on pike cichlids and may be
related to the elongation of the caudal peduncle, with convergent
transformations corresponding to synapomorphies of Teleocichla
and Lugubria.
Char. 150: 0=>2 A Posterior margin of preopercle regularly
serrated, with short serrations.
e character is very homoplastic, with several transformations
inside the subtribe Crenicichlina. However, it has been largely used
as diagnostic characters for subgroups of the former Crenicichla in
taxonomic papers. According to the optimization performed herein,
transformations from a smooth preopercle to a regularly serrated
preopercle are also found as a synapomorphy of Saxatilia.
Downloaded from https://academic.oup.com/zoolinnean/advance-article/doi/10.1093/zoolinnean/zlad021/7204395 by Universidade de Sao Paulo (USP) user on 27 July 2023
New generic classication of pike cichlids • 23
Apart from some characters related to external morph-
ology, such as squamation (characters 23, 24, 25, 26, 27, and
40) and colour paerns (51, 71, and 80), autapomorphies
come from highly homoplastic characters inside the subtribe
Crenicichlina and are of lile value in formulating a diagnosis
for the species.
Table 10. List of synapomorphies of the subgenus Crenicichla of Crenicichla (Node 122) based on the optimization of the characters on the ML tree
Character transformation Apomorphic condition Observations
Char. 13: 1=>2 Ascending process of the premaxilla much
longer than its dentigerous arm.
A reversal to the plesiomorphic condition inside the subtribe
Crenicichlina. Another reversal 1=>2 considered a
synapomorphy of Teleocichla.
Char. 23: 2=>1 Scales on interopercular area distributed only
on the posterior half.
Char. 24: 1=>0 Scales on the anterior portion of the cheek
(below eye) ctenoid.
Char. 25: 1=>0 Scales on dorsum ctenoid on the entire area
(paern 0).
A convergent transformation occurs only as autapomorphy
of C. viata among the pike cichlids.
Char. 26: 2=>1 Paern B1 of distribution of cycloid/ctenoid
scales on the ventral portion of the body.
Char. 27: 1=>0 Scales ctenoid on the the entire scaled area of
the caudal n except for a narrow marginal
area covered with cycloid scales.
Char. 32: 1=>0 Scales of the midlateral area of body ovoid,
with long vertical axis
A reversal to the plesiomorphic condition inside the subtribe
Crenicichlina. Other reversals also occur in less inclusive
groups of Saxatilia and as autapomorphies of species within
the subgenus Crenicichla (Lacustria).
Char. 33: 0=>1 Presence of scales on the basal portion of the
pectoral n.
Few species of pike cichldis occasionally show scales on
the basal portion of the n, embedded in skin. Convergent
transformations of this character occur as synapomorphy
of Lugubria and as autapomorphies of two species of pike
cichlids.
Char. 40: 0=>1 Scales covering covering the entire area of the
caudal n, except by a narrow distal margin.
Convergent transformations occur as synapomorphy of the
nodes 149 and 138 and as autapomorphy of Crenicichla (La.)
viata.
Char. 51: 1=>0 Absence of a dark blotch on the caudal n. Inside the subtribe Crenicichlina, convergent transform-
ations occur only as autapomorphies of Lugubria phaiospillus,
Lugubria johanna, and Teleocichla preta.
Char. 71: 0=>1 Presente of a dark blotch on pectoral axila. Convergences occur as synapomorphy of Hemeraia, of the
C. missioneira complex within the subgenus Crenicichla
(Lacustris), and of less inclusive groups within Saxatilia and
Lugubria.
Char. 80: 0=>1 Presence of conspicuous dark spots around
the pores of the lateral line scales on ank.
Convergences occur as a synapomorphy of a group wihin
the subgenus Crenicichla (Lacustria), an autapomorphy of C.
(La.) jaguarensis and a group of Wallaciia.
Char. 98: 0=>1 Eyes visible in ventral view (eyes positioned
more laterally in the head)
Char. 109: 1=>0 Posterior of the lachrymal without conexion
with the anterior border of the adjacente
infraorbital.
Char. 114: 0=>1 Two centra of the anteriormost precaudal
vertebrae rostro-caudally compressed
One of the sevral reversals to the plesiomorphic condition
inside the subtribe Crenicichlina.
Char. 117: 0=>1 Last basapophysis situated on the last
abdominal vertebra.
Char. 168: 1=>0 Presence of anterodorsal laminar
expansion of basisphenoid.
Char. 175: 1=>2 Ossied canal of the nasal long and nearly
straight.
Char. 183: 1=>0 Uncinate process and anterior arm of the
epibranchial 1 with similar width.
Downloaded from https://academic.oup.com/zoolinnean/advance-article/doi/10.1093/zoolinnean/zlad021/7204395 by Universidade de Sao Paulo (USP) user on 27 July 2023
24 • Var e l l a et al.
Character 98 (state 1: eyes visible in ventral view) was op-
timized as one of the apomorphic conditions of the subgenus,
with convergences also optimized as synapomorphies of
Saxatilia and Wallaciia. It was dicult to decisively infer whether
the visibility of the eyes (in ventral view) resulted from two vari-
ables instead of one: size of the eye and its relative position in the
head. In Saxatilia, the eyes are situated more laterally in the head
(wide interorbital space). Crenicichla (C.) macrophthalma and
Wallaciia are recognized to have larger eyes, which occupy a large
portion of the upper lateral portion of the head because of their
size and because the interorbital space is narrow. Further evalu-
ation of this character based on ontogenetic information or even
a reinterpretation may be important for subsequent analyses.
Diagnosis: Crenicichla (Crenicichla) macrophthalma is a medium-
sized species of pike cichlid (max. SL 200 mm) that can be
distinguished from all other pike cichlids by the following com-
bination of characters: large eyes (orbital diameter 9.4–10.6%
of SL); body almost entirely covered by ctenoid instead of cyc-
loid scales (scales ctenoid on cheek, ank squamation following
paerns 0 and B1, and scales ctenoid covering the caudal n al-
most entirely); and the absence of dark markings below the eye
(suborbital marking) and on caudal blotch. Among pike cich-
lids, only C. (La.) viata shows a similar squamation paern but
is readily distinguished from C. (C.) macrophthalma by having
more scales in the E1 series (79–93 vs. 65–68) and by having
a suborbital marking, a conspicuous dark midlateral band on
ank, and a caudal blotch.
Additional comparisons between groups: e monotypic subgenus
Crenicichla is distinguished from almost all species of the sub-
genera Batrachops and Lacustria by the absence of sexual di-
morphism expressed by mature females with orange or reddish
pigmentation on the lateral abodomen and by dark blotches
on the dorsal n of mature females. e laer character also
distinguishes the subgenus Crenicichla from most species of
Wallaciia. Crenicichla (Crenicichla) macrophthalma is distin-
guished from most species of Saxatilia and Lugubria by the
absence of a reddish or purplish broad pigmentation on the
ventral abdomen (belly) of mature females. e subgenus
Crenicichla also diers from Saxatilia by the absence (vs. pres-
ence) of a humeral blotch, from Lugubria by having fewer scales
in the E1 series (65–68 vs. 88–123), from Wallaciia (except W.
heckeli) by the absence of serrations on the posterior margin of
supracleithrum, and from Teleocichla and Hemeraia by having all
post-lachrymal infraorbitals autogenous vs. infraorbitals 4 and
5 co-ossied, forming a median pore. e subgenus also diers
from Hemeraia by having regularly serrated instead of smooth
preopercle and from Teleocichla by having a pelvic n with
rounded margin and second ray longest instead of pointed with
third ray longest.
Distribution: Crenicichla macrophthalma is known from tribu-
taries (Río Negro, Río Trombetas, Río Tapajós and Río Xingu)
and the main channel of the Amazonas River.
Node 133—Crenicichla (Batrachops) Heckel 1840, subgenus of
Crenicichla
Batrachops Heckel 1840: 432 (type species Batrachops reticulatus
Heckel 1840, designated subsequently by Eigenmann and Bray
1894: 620).
Boggiania Perugia 1897: 148 (type species Boggiania ocellata
Perugia 1897 = Batrachops reticulata), as junior synonym.
e node representing the subgenus Batrachops has relatively
good support in the ML tree (BS 88%), moderate in the BI tree
(PP 64%), and low support in the parsimony analysis of EW/
DiscreteMatrixTE (ABS 3/RBS 20%; Table 6). e clade is sup-
ported by eight morphological synapomorphies (no molecular
synapomorphies), being ve unambiguous and three ambiguous
(Table 11)—all of them congruent with the optimization
Table 11. List of synapomorphies of the subgenus Batrachops of Crenicichla ((Node 133) based on the optimization of the characters on the
ML tree
Character transformation Apomorphic condition Observations
Char. 48: 0=>1 Lateral portion of the pelvic n with skin thickening. Convergences occur in other groups of pike cichlids.
Condition apparently related to rheophilic behaviour.
Char. 155: 0=>1 Medial process (ascendent) of the proximal
extrascapula about twice longer than the distal process.
Char. 162: 0=>2 Lateral crest of epioccipitals on the eppiocitals only.
Char. 163: 0=>1 Caudal opening of the posterior myodome absent. Convergent transformations optimized as
autapomorphies of several species.
Char. 211: 1=>0 Posterior margin of the glossohyal convex. Convergent transformation as synapomorphy of
Teleocichla.
Ambiguous
Char. 14: 1=>2 A Horizontal portion of the preopercle longer than the
vertical portion.
Optimized as a synapomorphy, but with a reversal
on the node 150, which comprises four of the six ter-
minal taxa of the group.
Char. 129: 0=>1 A Decreasing to two or three rows of teeth in the
symphyseal region of the premaxilla.
Convergences as synapomorphies of the C. (La.)
scoii complex (node 179) and as autapomorphy of
C. (La.) jupiaensis.
Char. 150: 2=>1 A Posterior margin of the preopercle with few weak
serrations irregularly distributed.
Downloaded from https://academic.oup.com/zoolinnean/advance-article/doi/10.1093/zoolinnean/zlad021/7204395 by Universidade de Sao Paulo (USP) user on 27 July 2023
New generic classication of pike cichlids • 25
performed on the tree obtained from the parsimony analysis of
EW/DiscreteMatrixTE.
None of them corresponds to characteristics used in previous
papers to diagnose the C. reticulata group, except for character
150 state 1, which distinguishes Crenicichla and Batrachops
sensu Heckel (1840). e lack of unambiguous apomorphies
in Batrachops is probably due to the inclusion of Crenicichla
(Batrachops) jegui, with a morphology that diverges from the re-
maining species of the subgenus in several aspects (see Remarks).
Indeed, C. (B.) jegui was previously placed in the C. lugubris
group by Ploeg (1991) and has been placed among the species
of the C. reticulata group based on recent molecular phylogenies
(Burress et al. 2017, 2018). We recovered monophyly of the C.
reticulata group, including C. jegui (i.e. subgenus Batrachops),
in the ML and BI trees, as well as in the parsimony analysis
of EW/DiscreteMatrixTE and EW/ContinuousMatrixTE.
Parsimony analyses of XIW excluded C. jegui from the subgenus
Batrachops and placed it instead as sister to a clade of Crenicichla
macrophthalma + remaining Batrachops. Node 149, which in the
ML tree represents the subgenus Batrachops, excluding C. jegui,
show 11 synapomorphies, most of them congruent with char-
acteristics used in previous diagnoses of the C. reticulata group
and useful to partially diagnose the subgenus Batrachops herein
(char. 40: 0=>1 A; char. 41: 0=>2 A; char. 52: 2=>3; char. 73:
0=>1; char. 78: 0=>1; char. 131: 0=>2; char. 176: 0=>2; char.
206: 0=>1).
Diagnosis: Species in the subgenus Batrachops, except C. (B.)
jegui, are distinguished from all remaining pike cichlids by
the reticulate colour paern on the anks emerging from the
dark pigmentation on the base of the scales. However, the
conspicuosness of the resulting horizontal stripes forming this
paern can vary across species and through ontogeny. Species
in the subgenus Batrachops, except C. (B.) jegui, are also distin-
guished from all remaining pike cichlids by expressing the dark
blotch on the caudal n displaced posteriorly and dorsally in
relation to the base of the caudal n and the posterior branch
of the lateral line, and from all remaining species of the subtripe
Crenicichlina, except those in the subgenus Lacustria, by the
orange or reddish marking of the lateral abdomen in sexually
mature females. All species of the subgenus Batrachops, except
C. (B.) jegui and C. (B.) cyclostoma, are distinguished from all
other pike cichlids, except some species of Lugubria and C.
(La.) viata, by having almost the entire caudal n densely
covered by scales (vs. caudal n only partially covered by scales
and scales arranged in single or two series along the inter-radial
membranes).
All species in the subgenus Batrachops are additionally dis-
tinguished from the subgenus Crenicichla by the dark vertical
bars expressed as a series of blotches along the midlateral area
in adults, by the presence (vs. absence) of a caudal blotch, and
by the cycloid (vs. ctenoid) scales on the cheek. Batrachops is
additionally distinguished from the subgenus Lacustria by the
absence of a dark suborbital marking. Crenicichla (B.) jegui has
a well-dened, uniformly pigmented suborbital stripe running
obliquely from the ventral margin of the orbit to the corner of
the preopercle, a unique condition of the species among the sub-
genus, presumably not homologous with the suborbital marking
formed by dark dots of the subgenus Lacustria.
All species of the subgenus Batrachops are additionally dis-
tinguished from Saxatilia by the absence of a humeral blotch,
from Lugubria by 55–75 (vs. 88–123) scales in the E1 series, and
from Wallaciia by their larger body size (max. SL 96–216mm vs.
52–85mm in Wallaciia) and by the absence of serrations on the
posterior border of the suplacleithrum (except W. heck e l i, with
smooth supracleithrum). All species of the subgenus Batrachops
are also distinguished from Teleocichla and Hemeraia by having
autogenous post-lachrymal infraorbitals (vs. infraorbitals 4
and 5 coossied, forming a median pore); they are further dis-
tinguished from Teleocichla by the rounded pelvic n with the
second ray longest (vs. pelvic po inted with third ray longest), and
by prognathous or isognathous jaws (vs. hypognathous jaws).
Further distinguished from Hemeraia by having most scales on
ank ctenoid (combination of paerns 1 and 3 on the dorsal
portion and paern B2 on the ventral portion of the body) vs.
having most scales on ank cycloid (combination of paerns 3
and B3; see characters 25 and 26 and Supporting Information,
Appendix S2, Fig. S2).
Remarks: Most species of the subgenus Batrachops correspond
to the previously proposed C. reticulata group, traditionally char-
acterized by a cylindrical body, depressed head, short snout, and
large gape, all of which have been correlated with the occupation
of benthic habitats (e.g. Kullander 1986; Ploeg 1991). However,
in rapids of the Río Tocantins, the sympatric species C. (B.) jegui
and C. (B.) cyclostoma represent strong deviations from the typ-
ical Batrachops morphology. Crenicichla (B.) jegui is a reophilic,
sedentary (boom sier) species with very depressed body and
head, eyes dorsally displaced on the head, and distinctly prog-
nathous jaws. e reticulate colour paern widespread among
Batrachops species is replaced by a coarse paern of light blotches
and vermiculations on the entire dorsal portion of the body,
suggesting cryptic coloration in fast-owing, turbulent waters.
Contrastingly, C. (B.) cyclostoma lives among dark rocks in mod-
erate- to fast-owing waters of the Río Tocantins and has laterally
compressed body and head with a darker overall coloration. e
two species also dier in the feeding apparatus, suggesting dis-
tinctive diets and foraging strategies. All species of the subgenus
Batrachops have oral jaw teeth in the outer row rmly xed, while
teeth in the inner rows are slightly movable, a character that helps
to distinguish Batrachops from the subgenus Crenicichla, the
genera Saxatilia, Lugubria, Wallaciia, and Hemeraia, and from
most species of the subgenus Lacustria. However, while outer
row teeth in most species of Batrachops are unicuspid and mod-
erately robust, C. (B.) jegui has caniniform teeth (long and thin),
and the teeth of C. (B.) cyclostoma are unusually large and thick,
suggesting pronounced dierences in diet among the three lin-
eages. Similarly, the lower pharyngeal jaw (LPJ) diers, with
the common condition in the subgenus being a robust and wide
LPJ bearing papilliform teeth on its medioposterior portion,
whereas in C. (B.) jegui the LPJ is longer and thinner, bearing
only cuspidated teeth, and in C. (B.) cyclostoma the LPJ is stout
and strongly sutured, bearing strong molariform teeth. All this
suggests that C. (B.) jegui may be a specialized, probably am-
bush, predator and C. (B.) cyclostoma is probably durophagous,
foraging on rocky substrates similarly to Teleocichla preta (Varel la
et al. 2016). As a result of these ecomorphologically unique ari-
butes, nding universally diagnostic characters shared between
Downloaded from https://academic.oup.com/zoolinnean/advance-article/doi/10.1093/zoolinnean/zlad021/7204395 by Universidade de Sao Paulo (USP) user on 27 July 2023
26 • Var e l l a et al.
C. (B.) jegui, C. (B.) cyclostoma, and the remaining species of the
subgenus Batrachops remains challenging.
Distribution: Amazon basin (main channel and tributaries), Río
Orinoco, and Essequibo river basins. Crenicichla (B.) semifasciata
is the only species found in the Río Paraguai and in the lower
and middle portions of the Río Paraná. Recent reports from
the upper Río Paraná basin probably result from anthropogenic
introductions (Roa-Fuentes et al. 2015).
Node 127—Crenicichla (Lacustria), new subgenus of Crenicichla
urn:lsid:zoobank.org:act:D8B740B7-EF56-437E-8DCA-
530741D4D775
Type species: Cycla lacustris Castelnau, 1855.
e subgenus Lacustria is recovered in all analyses performed
herein with high support in the ML and BI trees (BS 100% and
PP 91.1%), but low support in the parsimony analysis of EW/
DiscreteMatrix (ABS 3/RBS 20%; Table 6). It is also recovered
in the recent phylogenies based on molecular data exclusively
(Piálek et al. 2012; Burress et al. 2017, 2018). Six morphological
synapomorphies (four unambiguous and two ambiguous) and
37 molecular unambiguous transformations are optimized for
the Node 127 in the ML tree (Table 12)—all of them coincident
with the optimization on the consensus tree based on parsimony
analysis of the EW/DiscreteMatrixTE.
Diagnosis: Species of the subgenus Lacustria are distinguished
from all species of pike cichlids by the presence of a dark sub-
orbital paern formed by a variable number of conspicuous
dark spots (punctulations), more or less scaered over the
cheek. Species of Lugubria, Saxatilia, and Hemeraia also have a
suborbital marking, as well as C. (B.) jegui, but this is interpreted
as another paern, not formed by spots but uniformly pigmented
(see characters 57–60). Species of the subgenus Lacustria are fur-
ther distinguished from most species of Lugubria by the presence
of 41–75 scales in the E1 row (vs. 88–123), with the exception
of C. (La.) viata (79–93 scales), and by the absence of a dark
blotch behind the pectoral n (vs. blotch appearing in adults of
most Lugubria species). From Saxatilia, all species of Lacustria are
distinguished by the absence of a humeral blotch and by sexually
dimorphic females with an orange or reddish marking laterally on
the abdomen instead of a broad reddish or purplish pigmentation
on the ventral portion of the abdomen of Saxatilia.
Lacustria is distinguished from the subgenus Batrachops, ex-
cept C. (B.) cyclostoma, by the laterally compressed body (vs.
nearly cylindrical or depressed) and from Batrachops, except C.
(B.) jegui, by the absence of a reticulate colour paern on the
ank resulting from the dark pigmentation on the base of each
ank scale. Lacustria diers from the subgenus Crenicichla by
the presence of cycloid (vs. ctenoid) scales on the cheek, dorsal
head, and area anterior to the dorsal n, and on the chest; by the
presence of a caudal blotch and by sexually dimorphic females
showing orange or reddish marking on the lateral abdomen
and, occasionally, one or more dark blotches on dorsal n (vs.
absence of these sexually dimorphic features related to color-
ation). Species of Lacustria are additionally distinguished from
Wallaciia by their size (max. SL 95–294mm vs. 52–85mm) and
by the smooth posterior margin of the supracleithrum (vs. pos-
terior margin with serrations in all Wallaciia, except W. h e ckeli ).
e subgenus is additionally distinguished from Teleocichla
and Hemeraia by having all the post-lachrymal infraorbitals au-
togenous (vs. infraorbitals 4 and 5 co-ossied forming a median
pore).
Table 12. List of synapomorphies of the subgenus Lacustria of Crenicichla (Node 127) based on the optimization of the characters on the ML
tree
Character transformation Apomorphic condition Observations
Char. 59: 0=>1 Presence of dark suborbital punctulations. A unique transformation of the Node 127.
Char. 85: 0=>1 Sexually dimorphic females showing dark one or more
dark blotches on the dorsal n.
Convergent transformations optimized as synapo-
morphy of a less inclusive group of Wallaciia (Node
186) and as autapomorphy of C. (B.) cyclostoma,
and a reversal occurs in C. (La.) viaa.
Char. 124: 0=>1 Posterior border of the alveolar process of the
premaxilla curved, bulbous.
Convergent transformations optimized as
autapomorphies of several species of Teleocichla and
Hemeraia chicha.
Char. 202: 1=>0 Microbranchiospines regularly distributed on the
external (lateral) face of the second to fourth
branchial arches.
A reversal to the plesiomorphic state inside the
subtribe Crenicichlina.
Ambiguous
Char. 66: 0=>1 A Post-temporal dark marking present in adults. is character is very homoplastic inside the
subtribe Crenicichlina. Optimized also as syn-
apomorphies of the Node 132 (Clade Hemeraia-
Lugubria-Saxatilia) and of Teleocichla, and with two
reversals inside the subgenus Crenicichla (Lacustria).
Char. 177: 1=>2 A Posterior foramen of the nasal unique or divided,
diplaced anterior and with the posterior portion of the
canal ossied, resulting in the openings placed at
the middle of the nasal canal.
is is a very homoplastic character and the state
2 represents a secondary modication of the
synapomorphic condition of pike cichlids (state 1).
Downloaded from https://academic.oup.com/zoolinnean/advance-article/doi/10.1093/zoolinnean/zlad021/7204395 by Universidade de Sao Paulo (USP) user on 27 July 2023
New generic classication of pike cichlids • 27
Distribution: Atlantic coastal rivers in Brazil (from the Río
Buranhem to the Laguna dos Patos system) and in the Río
Paraná, Río Uruguay, and Río Paraguay basins (Río de La Plata
superbasin). A recently discovered, undescribed species expands
the distribution of the subgenus to the Río São Francisco basin
in north-eastern Brazil (H. Varella, pers. obs.).
Remarks: Our results agree with all previous studies based on
molecular data, which place Crenicichla viata among the spe-
cies of the subgenus Lacustria, former C. lacustris group (Fig. 3).
is is the only species of the subgenus widely distributed in the
three major rivers of the La Plata superbasin and falls in dierent
positions of the subgenus depending on the analysis performed.
Before those recent studies, C. (La.) viata was considered a
member of the C. lugubris group (e.g. Kullander 1991, 1997;
Ploeg 1991; Fig. 3I), mainly by being a medium to large species
with numerous small scales on the anks.
Less inclusive groups in the subgenus Lacustria: Node 153 cor-
responds with what has been called the C. missioneira group or
complex since Lucena and Kullander 1992 or Uruguay River
species ock (URSF, Burress et al. 2017, 2018; Fig. 3D−F).
is group is recovered in all analyses performed herein and is
well-supported in the ML and BI topologies (BS 100% and PP
100%) but has low support in the parsimony analysis of EW/
DiscreteMatrixTE (ABS 2/RBS 20%; Table 6). Besides the spe-
cies included as terminal taxa in the analyses performed herein,
the C. missioneira complex also includes Acharches niederleinii,
considered a nomen dubium of the Río Uruguay (Río Chapecó)
that could correspond to C. (La.) missioneira, C. (La.) minuano,
or C. (La.) hadrostigma (Varella 2011; Steinhauser 2019). All
species are endemic to the middle and upper Río Uruguay basin.
e C. missioneira complex can be characterized by 10 mor-
phological synapomorphies (seven of them unambiguous) and
91 unambiguous molecular transformations. Among the mor-
phological synapomorphies, four are useful to diagnose the
group among the species of the subgenus Lacustria and were
used in the previous characterization of the group by Lucena
and Kullander (1992): char. 12: 2=>1: lachrymal as long as deep
(approximately square) vs. lachrymal longer than deep; char. 60:
1=>2 D: dark suborbital marking composed of one or few spots
restricted to the posterior portion of the cheek; char. 71: 0=>1:
presence of a dark blotch on the pectoral axila; and char. 150:
2=>0 A: posterior margin of the preopercle smooth, with no
serrations. Additionally, most species of the C. (La.) missioneira
complex show sexually dimorphic males with the dark blotch of
the caudal n fragmented, accompanying the paern of irregular
small blotches scaered on the caudal peduncle.
e C. missioneira complex is the sister-group of all remaining
species of Lacustria, encompassing several species from the
Atlantic coastal drainages of Brazil, the Río Paraná, Río Paraguay,
and the Río Uruguay basins. e C. scoii complex (sensu
Lucena and Kullander 1992) is recovered in all analyses per-
formed herein (Node 179). It includes the three remaining valid
species of the subgenus distributed in the Río Uruguay basin:
C. (La.) scoii, C. (La.) gaucho, and C. (La.) prenda. e clade
is well-supported in the ML and BI analyses (BS 100% and PP
100%), and in the parsimony analysis of EW/DiscreteMatrixTE
(ABS 3/RBS 100%; Table 6). Twelve synapomorphies (nine of
them unambiguous) and 39 unambiguous molecular transform-
ations characterize the group. Species of this group have been
traditionally dierentiated from other species in the subgenus by
having their body less compressed laterally, a wider interorbital
area, short snout, and a blunt, wide mouth. Some of these char-
acteristics were optimized as synapomorphies for the group viz.
char. 13: 1=>0: ascending process and dentigerous arm of the
premaxilla with similar lengths; char. 15: 1=>0: frontal bones
less compressed laterally (interorbital distance 26.3–50.9% of
the neurocranium length); and char. 175: 1=>0: nasal canal
short and curved; and char. 176: 0=>2: laminar ossication
of the nasal canal well-developed both medially and laterally.
Other synapomorphies of the C. missioneira group are related to
their teeth, arranged in fewer series but more rmly xed onto
the jaws (char. 129: 0=>1 A; char. 130: 0=>1; and char. 131:
0=>2). ese characters have been used to distinguishing this
group from many species of Lacustria (e.g. Lucena and Kullander
1992).
e C. mandelburgeri complex (Piálek et al. 2012), or Paraná
River species ock (Burress et al. 2018), is recovered as mono-
phyletic (Node 163) with high support in the ML tree (BS
98%), but moderate support in the BI consensus tree (PP
58.3%), and low support in the parsimony analysis based on
EW/DiscreteMatrixTE (ABS 3/RBS 17; Table 6). Although this
group is characterized by one synapomorphy only, it is exclusive
for the group: char. 86: 0=>1: dark blotches on the dorsal n of
sexually dimorphic females modied into a unique horizontal
dark band. e synapomorphic condition and the composition
of the group agree with Piálek et al. (2012), with the exception
of the inclusion of C. (La.) jaguarensis. A group formed with all
endemic species of the Río Paraná basin is only found in the par-
simony analyses under equal weighting (EW/DiscreteMatrixTE
and EW/ContinuousMatrixTE), but not in the ML and BI
analyses, as well as in the other parsimony analyses performed
herein. is is due to to the discordant position of Crenicichla
(La.) sp. Paraná within the subgenus. Because of its distribution
in the region of the Itaipu reservoir (transition between upper
and middle Río Paraná), further reassessment of the relationsips
of the species is necessary to beer understand the biogeography
of the subgenus Lacustria.
Hypertrophied lips in the subgenus Lacustria: Hypertrophied lips,
such as those observed in C. (La.) tendybaguassu and C. (La.)
tuca, have evolved exclusively within the subgenus Lacustria
of Crenicichla. Similar phenotypes are found in phylogenetic-
ally distant lineages in African lakes (e.g. Colombo et al. 2013;
Baumgarten et al. 2015) and Central America (e.g. Elmer et
al. 2010; Manousaki et al. 2013) but are not found in other
pike cichlids or Geophagini taxa. e labial turgescence in the
geophagine Gymnogeophagus labiatus is not considered hom-
ologous as it does not result in the formation of median lobes
(Lucena and Kullander 1992).
According to our results, the occurrence of hypertrophied
lips is not restricted to C. tendybaguassu and C. tuca, but also
observed in two other morphologically dened, putative new
taxa belonging to the subgenus Lacustria in the Río Iguaçu
basin. ree pairs of big-lipped/typical taxa of pike cichlids
in the Río Iguaçu basin, matching in relation to overall shape
and coloration, were included in the combined phylogenetic
Downloaded from https://academic.oup.com/zoolinnean/advance-article/doi/10.1093/zoolinnean/zlad021/7204395 by Universidade de Sao Paulo (USP) user on 27 July 2023
28 • Var e l l a et al.
analysis. Crenicichla iguassuensis BL and C. tesay BL are big-
lipped forms corresponding to C. iguassuensis and C. tesay, re-
spectively. e third pair is C. tapii (typical) and C. tuca (big
lips). Crenicichla iguassuensis BL and C. tesay BL were repre-
sented in the combined datasets only with morphological data
and the remaining were represented with both molecular and
morphological data.
Only Crenicichla tuca and C. tapii are recovered as sister-taxa
in our main hypothesis (ML tree), in the BI tree, and in some of
the parsimony analyses. e other big-lipped forms do not group
with their typical-lipped correspondents. Piálek et al. (2012)
also suggested the existence of several lineages of big-lipped pike
cichlids in the Río Iguaçu basin. More recently, and with a more
complete sampling, Rícan et al. (2021a) found several dierent
big-lipped forms within both C. missioneira and C. mandelburgeri
complexes, but this study was inconclusive regarding the species
delimitation within these clades.
According to the optimization of morphological char-
acters on the ML tree (Fig. 4), hypertrophy with the for-
mation of ventral and dorsal lobes (Fig. 4: char. 97) is
convergent in Crenicichla tendybaguassu, C. tuca, C. tesay BL,
and C. iguassuensis BL (char. 97). e development of the
lobes can vary largely intraspecically and during ontogeny
(Supporting Information, Appendix S2, Fig. S22), and it is
accompanied by the development of a mental process (Fig.
4: char. 97) on the dentary, but this occurs convergently in C.
jupiaensis and C. (Batrachops) cyclostoma, which have strong
oral jaws and symphyseal teeth. Osteological modications co-
died herein, correlated with narrowing and downturning of
the mouth (characters 13, 91, 128, 142, and 159) and incre-
ment of biting force (characters 127, 131, 132, and 133), are
not exclusive of big-lipped taxa, but occur in several species
that deviate from the typical pelagic-predatory morphology of
pike cichlids. Narrowing of the mouth is more pronounced in
big-lipped forms of the Río Iguaçu basin when compared with
their typical-lipped correspondents. e dierence is more
evidente in the pair C. iguassuensis (typical Crenicichla pelagic
predator, with wide mouth gape and prognathous jaws) and C.
iguassuensis BL (crevice feeder, with narrow mouth gape and
isognathous jaws).
Considering our comparisons between big-lipped and
typical-lipped correspondents, features on the LPJ (directly
correlated to food processing, e.g. Burress et al. 2017), are
more conservative than those on the oral jaws (correlated to
foraging). LPJ of big-lipped and typical-lipped correspond-
ents are similar in shape and robustness (character 205) and
type of teeth on the medioposterior portion (character 206),
but the LPJ of big-lipped taxa shows denser dentition [as al-
ready pointed out by Lucena and Kullander (1992) for C.
tendybaguassu in the context of the C. missioneira complex].
Even in C. iguassuensis vs. C. iguassuensis BL pair, with pro-
nounced differences on morphology of oral jaws, the LPJ of
the big-lipped form has denser dentition but is very similar
in shape and types of teeth (only unicuspid and bicuspid
teeth, no papilliform or molariform teeth, which suggests
preference for durophagy). The only clearly specialized
molluscivorous species in the Río Iguaçu basin is what we
identify as C. yaha (sensu Varella 2011; but for incongruences
on species delimitations between authors, see also: Piálek et
al. 2015 ), with much enlarged medioposterior teeth on the
LPJ, and this is the only species that does not have a big-
lipped correspondent.
Figure 4. Right. Diversity of ecomorphologies in the assemblage of the subgenus Lacustria of Crenicichla in the Río Iguaçu basin, illustrating
the three pairs of typical- vs. big-lipped taxa that match in coloration and body shape: A, Crenicichla iguassuensis, NUP 6710, 97.6mm SL; B,
Crenicichla iguassuensis BL (big lips), MHNCI 7057, 96.7mm SL; C, Crenicichla tesay, NUP 3751, 100.8mm SL; D, Crenicichla tesay BL (big
lips), MHNCI 7729, 97.5mm SL; E, Crenicichla tapii, NUP 1795C, 101.5mm SL; F, Crenicichla tuca (big lips), MHNCI10378, 133.2mm SL.
Le. Mapping of characters related to narrowing and downturning of the mouth and oral jaws (foraging) and of characters related to the shape
and dentition of the lower pharyngeal jaw (LPJ; food processing) on the ML tree, in the context of the Crenicichla mandelburgeri complex
(Node 163). LPJ in dorsal view of: A, C. iguassuensis, NUP 6694, 106.9mm SL; B, C. iguassuensis BL, NUP 1795, 125.8mm SL; C, C. tesay,
NUP 3751, 97.2mm SL; D, C. tesay BL, NUP 591, 152.3mm SL; E, C. tapii, NUP 1788, 134.6mm SL; F, C. tuca, NUP 1795C, 128.0mm SL;
G, C. yaha, NUP2966C, 92.6mm SL.
Downloaded from https://academic.oup.com/zoolinnean/advance-article/doi/10.1093/zoolinnean/zlad021/7204395 by Universidade de Sao Paulo (USP) user on 27 July 2023
New generic classication of pike cichlids • 29
Node 183—Wallaciia, new genus
urn:lsid:zoobank.org:act:9780B331-2587-460D-AE49-
F589D167CEA9
Type species: Crenicichla wallacii Regan, 1905.
e genus Wallaciia is recovered in the ML and BI ana-
lyses, as well as in the parsimony analysis, under extended
implied weighting (XIW/DiscreteMatrixTE and XIW/
ContinuousMatrixTE). However, it is not recovered in the par-
simony analyses under equal weighiting (EW/DiscreteMatrix
and EW/ContinuousMatrixTE). We hypothesize that the dis-
cordance is a result of long-branch araction caused by conver-
gence in characters associated with specializations to rheophily
(see Remarks). Such convergence clustered W. h eckel i and W.
urosema, two species that inabit rapids and were represented by
morphological characters only, with Teleocichla. Although the
node corresponding to the genus Wallaciia was moderately sup-
ported in the ML tree (BS 71%; Table 6), it shows high support
in the BI tree (PP 100%) and is recovered as monophyletic in all
previous molecular studies (Kullander et al. 2010; Piálek et al.
2012; Burress et al. 2017, 2018; Fig. 3). Moreover, three of the
ve unambiguous morphological synapomorphies optimized
for the group (Table 13) are congruent with characters used to
diagnose the C. wallacii group in previous studies and also to
diagnose the genus Wallaciia as proposed herein.
Diagnosis: Wallaciia includes small-sized species (max. SL
52–85mm) that dier from all pike cichlids by the following
combination of traits: large eyes (orbital diameter 7.8–12.6%
of SL); presence of serrations on the posterior margins of the
supracleithrum; posterior margin of preopercle with prominent,
spine-like serrations; pectoral n with 13–14 rays (vs. 15–18; in
modal values); reduced predorsal squamation (not extending on
to the NFL0; see char. 28); and pelvic n rounded with second
ray longest and all post-lachrymal infraorbitals autogenous (vs.
infraorbitals 4 and 5 co-ossied, forming a median opening).
Wallaciia heckeli deviates from the typical Wallaciia morphology
by having some characters convergently shared with Teleocichla,
another small-bodied genus. For example, W. heck el i has smooth
preopercle and suprachleitrum, jaws not prognathous, and the
paern of xation of oral teeth and conguration of the lower
pharyngeal jaw is dierent from the typical Wallaciia species.
Also, it can be distinguished from other Wallaciia species by
having smaller eyes, infraorbitals co-ossied forming a median
opening (vs. 4 and 5 autogenous), and pelvic n pointed, with
third ray longest (vs. rounded). Besides, Wallaciia diers from
Teleocichla by lacking several other osteological modications as-
sociated to life in rapids and diagnostic of the later (see below).
Species of Wallaciia are additionally distinguished from
Saxatilia by the absence of a humeral blotch and suborbital
marking, from Lugubria by having fewer scales in the E1 series
(52–64 vs. 88–123 scales) and fewer vertebrae (31–34 vs.
38–42), from Hemeraia by having most scales on the anks
ctenoid (combination of paerns 1 and 2 on dorsal portion
and paern B2 on ventral portion of the ank) vs. most ank
scales cycloid (combination of paerns 3 and B3). Wallaciia
are additionally distinguished from the subgenus Crenicichla
of Crenicichla, by cycloid (vs. ctenoid) scales on the cheek and
on the chest. Wallaciia diers from the subgenus Batrachops of
Crenicichla by the absence of a reticulate colour paern on the
ank (vs. reticulate paern on anks formed with the dark pig-
mentation on the base of the scales), and from the subgenus
Lacustria of Crenicichla by the absence of a suborbital marking
(vs. presence of dark puntulations more or less scaered on the
cheek forming dierent paerns of suborbitals markings).
Distribution: As proposed here, Wallaciia includes eight spe-
cies that are distributed in the Amazon basin, Río Orinoco, and
Essequibo river basins. Species of Wallaciia are present in almost
all major rivers of the Amazon basin from the Río Madeira to the
Table 13. List of synapomorphies of the genus Wallaciia (Node 183) based on the optimization of the characters on the ML tree
Character transformation Apomorphic condition Observations
Char. 6: 2=>1 Pectoral ns with 13–14 rays. Inside the subtribe Crenicichlina, there is only one con-
vergent transformation as autapomorphy of Teleocichla
centisquama.
Char. 150: 0=>3 Posterior margin of preopercle strongly serrated,
with spine-like proeminent serrations.
Inside the subtribe Crenicichlina, another conver-
gent transformation is optimized as synapomorphy of
Lugubria and a reversal occurs in W. h ec ke li .
Char. 153: 0=>1 Presence of serrations on the posterior margin of
supracleithrum.
Exclusive synapomorphy of Wallaciia inside the subtribe
Crenicichlina (occurring as autapomorphy of Dicrossus
lamentosus in the outgroup), with a reversal in W. h ec ke li .
Char. 183: 1=>0 Uncinate process and anterior arm of the
epibranchial 1 with similar width (from uncinate
process wider).
is character is very homoplastic along the entire top-
ology and also includes a reversal in W. heck eli .
Ambiguous
Char. 98: 0 => 1 A Eyes visible in ventral view (eyes positioned more
laterally in the head)
Convergent transformations occurring as synapo-
morphies of the subgenus Crenicichla (Crenicichla) and
of Saxatilia.
Char. 209: 1=>0 A Spine of the anterior portion of the urohyal
directed posteriorly or dorsally.
As optimized herein, it represents one of the many
reversals to the plesiomorphic condition that occurred
inside the subtribe Crenicichlina.
Downloaded from https://academic.oup.com/zoolinnean/advance-article/doi/10.1093/zoolinnean/zlad021/7204395 by Universidade de Sao Paulo (USP) user on 27 July 2023
30 • Var e l l a et al.
Río Tocantins, but are not found in the Western portion of the
Amazon basin, i.e. tributaries and main channel of the Amazon
basin west of Santo Antonônio do Içá.
Remarks: is is the rst phylogenetic study of pike cichlids that
includes Wallaciia heckeli and W. urosema, two rheophilic species
that inhabit rapids of the Río Trombetas and Río Tapajós basins,
respectively. In the two parsimony analyses of the combined
datasets under the equal weighting scheme, these species cluster
with Teleocichla, as successive sister-groups. is is probaby a re-
sult of long-branch araction of some convergences shared by
W. hec k eli and Teleocichla, resulting in synapomorphies of an ex-
panded Teleocichla (including W. h e ckeli and W. urosema) or of
the group Teleocichla + W. he c k eli in the aforementioned parsi-
mony trees. Many of these convergent characters are correlated
with specializations for rheophilic habitats and specialized ben-
thic diets (e.g. Kullander 1988; Varella et al. 2016), including
reduced cheek squamation, dierences in number of precaudal
and caudal vertebrae, acute mouth with narrow gape, ridged
vomer (linked with a downturned mouth), and modications
on the format and teeth of the lower pharyngeal jaw (char. 19:
0=>2; char. 113: 0=>1; char. 128: 0=>1; char. 159: 1=>2 A;
char. 205: 2=>3; char. 206: 0=>1). Interestingly, these characters
are shared between Teleocichla and three species of Wallaciia in-
stead of two—the third being W. compresiceps, also a rheophilic
species that inhabits rapids in the Río Tocantins. Further study
is needed to help clarify the relationships of rheophilic Wallaciia
species, because W. compressiceps is the only taxon in our matrix
represented by both molecular and morphological data, whereas
the others are represented by morphology only, which is probaby
the reason why W. compressiceps grouped with Wallaciia instead
of Teleocichla, as did the two species with only morphological
data.
Node 191—genus Teleocichla Kullander 1988
Teleocichla Kullander, 1988: 196 (type species Teleocichla
centrarchus Kullander 1988)
Crenicichla non Heckel (1840)—Ploeg (1991: 122): synonym
not followed by subsequent authors.
Teleocichla monophyly is always recovered with high sup-
port (ML-BS 100%; BI-PP 100%; parsimony-ABS > 72/RBS
100%; Table 6) and characterized by 39 unambiguous molecular
transformations and 32 morphological synapomorphies (19 un-
ambiguous and 13 ambiguous; Table 14). Especially useful to
diagnose the genus are characters 19, 47, 48, 73, and 75 com-
bined, 91, 105, 113, and 174. For simplicity, Table 14 only pre-
sents six of the 13 ambiguous synapomorphies optimized on the
ML tree. ese characters are not only diagnostic of the genus
(char. 75) but represent repeated modications apparently asso-
ciated with rheophyly (chars. 13, 128, 159, and 206) and body
size reduction (char. 169) in the context of pike cichlids.
Most synapomorphies of Teleocichla agree with the putative
synapomorphies proposed or discussed as related to rheophyly
by Kullander (1988) in his diagnosis and anatomical descrip-
tion of the genus. We oer a simplied diagnosis that follows the
standard for other groups proposed herein.
Diagnosis: Teleocichla comprises small-sized species (max. SL
48–90mm, with exception of T. preta with max. 121.3mm SL)
that are distinguished from all other pike cichlids, except species
of Hemeraia and Crenicichla (La.) igara, by having infraorbitals 4
and 5 co-ossied, forming a median pore (vs. infraorbitals 4 and
5 autogenous). Teleocichla can be further distinguished from all
pike cichlids by the combination of the following apomorphic
conditions: pelvic n long and pointed, with third ray longest
(almost reaching the genital papilla) and with a skin thickening
on the lateral portion; mouth hypognathous; cheek squamation
restricted to its posterior portion; rostral pore of the nasal canal
displaced posteriorly from the postlabial margin of the snout;
and two or more caudal than precaudal vertebrae. Teleocichla
also presents several apomorphic conditions of internal morph-
ology that are useful to diagnose the genus (Table 14).
Species of Teleocichla are additionally distinguished from all
species of the sister-group Wallaciia, except W. h ec ke li , by the ab-
sence of serrations on the posterior margin of the supracleithrum.
Teleocichla is further distinguished from Hemeraia, Saxatilia,
subgenus Lacustria of Crenicichla, and many species of Lugubria
by the absence of suborbital markings. Teleocichla also diers
from Saxatilia by the absence of a humeral blotch, from Lugubria
by having fewer scales in the E1 series (53–79 scales, except
T. centisquama with 85–86 vs. 88–123 scales), from subgenus
Batrachops of Crenicichla, except C. (B.) jegui, by the absence of
a reticulate colour paern on the ank (vs. reticulate colour pat-
tern on the ank formed by the dark pigmentation on the base of
individual scales), and from Crenicichla (Crenicichla) by having
all cycloid (vs. ctenoid) scales on cheek, anterodorsal portion of
the body, and on the chest.
Distribution: e nine species of Teleocichla are apparently re-
stricted to the rapids of clear-water tributaries of the Amazon
basin, with species distributed in the Río Tapajós, Río Xingu,
Río Tocantins-Araguaia, Río Tocantins, and Río Jari basins.
Remarks: In this study, a clade including all species of Teleocichla
and Wallaciia as sisters to each other is recovered in all ana-
lyses combining morphological and molecular data, with good
support in the ML and BI trees (BS 99% and PP 100%), and
moderate support in the parsimony analysis based on EW/
DiscreteMatrixTE (A BS 10/RBS 53%; Table 6). is grouping is
also recovered in the ML analysis of the molecular data only, and
by the parsimony analysis of the ContinuousMatrix (i.e. morph-
ology only) under the equal weighting scheme. is grouping
agrees with the phylogenetic inferences made by Ploeg (1991;
Fig. 3I), based exclusively on morphological data, and was also
recovered in some of the topologies presented by Burress et
al. 2017 using UCEs (Fig. 3E). Other analyses in Burress et al.
(2017; Fig. 3D), and other previous studies (Burress et al. 2018;
Ilves et al. 2018; Fig. 3C, F) based on phylogenomic datasets,
disagree with our hypotheses and nd Teleocichla as sister-group
of the remaing groups of clade A.
On the ML tree, 14 morphological transformations are op-
timized as synapomorphies of this group, seven of them un-
ambiguous. Many of these synapomorphies are modications
related to reduction in body size, including char. 10: 1=>0:
maximum body size < 100 mm SL, with a reversal only in T.
preta that reaches 125mm SL. e others are the reduction of
the squamation on the anterodorsal portion of the body, or pre-
dorsal area (char. 28: 0=>1); a reduction of the supraoccipital
Downloaded from https://academic.oup.com/zoolinnean/advance-article/doi/10.1093/zoolinnean/zlad021/7204395 by Universidade de Sao Paulo (USP) user on 27 July 2023
New generic classication of pike cichlids • 31
Table 14. List of synapomorphies of the genus Teleocichla (Node 191) based on the optimization of the characters on the ML tree
Character transformation Apomorphic condition Observations
Char. 8: 2=>3 Five or more vertebrae + halfcentrum con-
tained witin the caudal peduncle.
Condition related to the elongation of caudal peduncle, with
several convergent transformations occuring inside the subtribe
Crenicichlina.
Char. 19: 0=>2 External area corresponding to the adductor
mandibulae muscles covered by scales only
on a small dorsoposterior portion.
is condition is generally related to rheophyly and convergent
transformations are abundant inside the subtribe Crenicichlina,
mainly in the clade Teleocichla + Wallaciia and in the genus
Crenicichla.
Char. 47: 2=>1 Pelvic pointed medially, with third ray long-
est.
Exclusive synapomorphy without reversal.
Char. 48: 0=>1 Presence of skin thickening on the
lateral portion of the pelvic n.
Condition generally related to rheophyly, with convergent
transformations occurring as synapomorphies of Crenicichla
(Batrachops) and less inclusive groups of Crenicichla (Lacustria),
as well as autapomorphy of Lugubria phaiospilus.
Char. 73: 0=>1 Dark vertical vertical bars expressed as a
series of blotches along the midlateral area
in adults.
Convergences occur as synapomorphy of a less inclusive group
in Crenicichla (Batrachops) and as synapomorphies of two small
groups and autapomorphies of some species of Crenicichla
(Lacustria).
Char. 91: 1=>2 Lower jaw hypognathous—upper jaw ex-
tending more anteriorly than the
lower jaw.
e transformation from a distinctly prognathous lower jaw to a
hypognathous lower jaw (1=>2) occurs only as synapomorphy
of Teleocichla. However, transformations from isognathous or
slightly prognathous to a hypognathous lower jaw (0=>2) occur
three times in Crenicichla (Lacustria).
Char. 105: 0=>1 Infraorbitals 4 and 5 fused, forming a
median pore. Convergent transformations
occurring as synapomorphy of Hemeraia
and as autapomorphy of C. (La.) igara.
Char. 113: 0=>1 Two to six more caudal vertebrae than
precaudal vertebrae (vs. more precaudal
vertebrae, equal numbers or only one
more caudal vertebra).
Inside the subtribe Crenicichlina, convergent transformations
occur as autapomorphies of W. heck eli , C. (La.) hadrostigma and
C. (La.) jupiaensis.
Char. 131: 0=>2 Heterogeneous xation of teeth onto oral
jaws: teeth of the outer row rmly xed and
teeth of inner rows slightly mobile.
is character is already mentioned in other groups, since it is
also optimized as synapomorphies of Crenicichla (Batrachops)
and of C. (Lac.) scoii complex. Other convergences occur in-
side the subgenus Crenicichla (Lacustria).
Char. 138: 1=>0 Dorsal process of the anguloarticular taller
than the posterior border of the alveolar
arm or approximately equal in height.
Inside the subtribe Crenicichlina, other three convergent trans-
formations occur in the genus Crenicichla.
Char. 142: 0=>1 Maxillary process of the palatine
approximately cylindrical.
Reversals occurs in Teleocichla centisquama, T. proselytus and
T. gephyrogramma, and convergences occur repeatedly inside
Crenicichla (Lacustria).
Char. 157: 1=>0 Absence of an anteriorly directed spinuous
process on the distal postcleithrum.
is condition represents a reversal to the plesiomorphic con-
dition inside the subtribe Crenicichlina. See comment on the
codication in the Supporting Information, Appendix S2[/
scolor].
Char. 164: 0=>1 Absence of a metapterygoid-vomer suture. Synapomorphy exclusive considering only the pike cichlids, but
with convergences considering the outgroup taxa.
Char. 174: 0=>1 Rostral pore of the nasal canal displaced
posteriorly from the postlabial margin
of the snout.
A reversal to the plesiomorphic condition inside the pike
cichlids. Convergent transformations occurring only as
autapomorphies of C. (La.) tesay and C. (La.) jupiaensis.
Char. 178: 0=>1 Triangular, pointed posterior margin of the
vomerine lateral process.
Convergence optimized as autapomorphy of Wallaciia
compressiceps and reversals found in the node 193 (T.
centrarchus + T. preta) and in T. monogramma.
Char. 208: 0=>1 Absence of spine on the anterior
portion of the urohyal.
Convergent transformation occurs only as autapomorphy of C.
(La.) hadrostigma.
Char. 210: 2=>3 Lateral wings of the urohyal distinctly
wider than the depth of its medial crest
or medial crest rudimentary or absent.
Convergent transformation occurring only as autapomorphy
of C. cametana. is represents a secondary modication of the
synapomorphic condition in the subtribe Crenicichlina (i.e.
medial crest reduced, lower than the width of the lateral wings).
Downloaded from https://academic.oup.com/zoolinnean/advance-article/doi/10.1093/zoolinnean/zlad021/7204395 by Universidade de Sao Paulo (USP) user on 27 July 2023
32 • Var e l l a et al.
medial crest and of the paired lateral crests of the neurocranium
(char. 160: 2=>3 A; char. 162: 0=>2 A); opening of the NFL3
along the canal connecting NLF2 and NLF4, lacking an acces-
sory canal (char. 172: 0=>1); and phar yngobranchial 1 diminute,
globular (char. 180: 0=>3).
Node 148—Hemeraia, new genus
urn:lsid:zoobank.org:act:6F581429-473C-480E-9D6D-
6B4898DF1A7C
Type species: Crenicichla hemera Kullander, 1991
Hemeraia, comprising only H. hemera and H. chicha, is re-
covered in all analyses performed herein, although in dierent
positions within the subtribe Crenicichlina (see Remarks):
as sister-group of the clade Saxatilia (ML and BI trees);
in a trichotomy with Saxatilia and Lugubria (Parsimony/
EW/DiscreteMatrixTE); as sister-group of the the clade
Saxatilia + Lugubria (Parsimony/XIW/DiscreteMatrixTE);
and as sister-group of all remaining pike cichlids (parsimony
analysis of ContinuousMatrixTE under EW and XIW schemes).
As this is the rst phylogenetic study that included both H.
hemera and H. chicha, there are no previously published phylo-
genetic hypotheses for comparisons.
Nevertheless, monophyly of the genus is well supported in
the ML and BI trees (BS 98% and PP 100%), and has moderate
support in the parsimony tree based on EW/DiscreteMatrixTE
(ABS 5/RBS 63%; Table 6). e two known species in the
genus were represented only by morphological data, but none-
theless the clade was diagnosable by six unambiguous and one
ambiguous synapomorphies (Table 15). Synapomorphic condi-
tions of characters 25, 26, and 105 help to diagnose the group
among pike cichlids and coincide with diagnostic characters
in previous taxonomic papers that included the two species
(Varella et al. 2012, 2018).
Character transformation Apomorphic condition Observations
Char. 211: 1=>0 Posterior margin of the glossohyal
straight or concave.
A reversal to the plesiomorphic condition inside the subtribe
Crenicichlina. Reversal occurring in T. preta.
Char. 215: 0=>1 Symmetrical medial processes of the
basipterygia running very close, not
diverging anteriorly.
Exclusive synapomorphy, without any reversal.
Highlighted ambiguous synapomorphies (6 of 13)
Char. 13: 1=>2 A Ascending process of the premaxilla
much longer than the dentigerous arm.
Highly homoplastic character in the context pike cichldis, but
with only one reversal inside Teleocichla as autapomorphy of T.
preta.
Char. 75: 0=>1 A Dark bars expressed as a series of
blotches, with a displacement of the
blotches in the midlateral area, resulting
in a ‘zig-zag’ paern.
Reversals occur in T. proselytus, T. centrarchus and T. cinderella.
Char. 128: 0=>1 D Symmetrical anterior portions of the
ascending processes of the right and le
premaxillae separated but with reduced
space.
is character is highly homoplastic and related to modica-
tions of the jaws in rheophilic species of dierent groups of
pike cichlids (resulting in acute mouth with narrow gape),
being optimized as synapomorphies of less inclusive groups
and/or autapomorphies of species within Hemeraia, Wallaciia,
Crenicichla (Batrachops), and Crenicichla (Lacustria).
Char. 159: 1=>2 A Anterior margin of vomer convex,
ridged.
e character is very homoplastic and condition 2 is somewhat
correlated to modications of rheophyly found in pike cichlids
(linked to a downturned mouth). is condition is optimized as
synapomorphies of less inclusive groups and autapomorphies of
species wihin Wallaciia and within the subgenera Crenicichla
(Lacustria) and Crenicichla (Batrachops).
Char. 169: 1=>0 D Vomerine sha and parasphecnoid with
straight connection instead of a sutured
connection.
Condition related to decreased size of the species, a reversal is
found in the relatively large species T. preta and convergences
are optimized as a synapomorphy of one less inclusive group
and as autapomorphy of one species within Wallaciia.
Char. 206: 0=>1 D Teeth of the medioposterior portion of
the lower pharyngeal jaw (LPJ) moder-
ately molarized (papilliform).
Modications in dentition of the LPJ are generally associ-
ated with foraging and diet specialization of pike cichlids
living in fast owing waters with rocky substrate. A second-
ary modication to enlarged molariform teeth is optimized
as synapomorphy of Node 193 (T. preta and T. centrarchus)
and autapomorphy of T. wajapi. Several transformations from
only unicuspid to papilliform or molariform teeth on LPJ
occur within Wallaciia, Crenicichla (Lacustria), and Crenicichla
(Batrachops).
Table 14. Continued
Downloaded from https://academic.oup.com/zoolinnean/advance-article/doi/10.1093/zoolinnean/zlad021/7204395 by Universidade de Sao Paulo (USP) user on 27 July 2023
New generic classication of pike cichlids • 33
Diagnosis: Hemeraia is distinguished from other pike cichlids
by the combination of: reduced distribution of ctenoid scales
(vs. cycloid) on anks (paern 3 and B3; see characters 25, 26
and Supporting Information, Appendix S2, Fig. S2); a smooth
preopercle (vs. serrated); infraorbitals 4 and 5 co-ossied, forming
a median pore (vs. autogenous); and presence of a suborbital
marking. Hemeraia is further distinguished from Lugubria by
having fewer scales in the E1 series (58–75 vs. 89–123). Hemera
is further distinguished from Saxatilia by the absence of a humeral
blotch that appears in early ontogenetic stages (only adults of H.
hemera shows a dark blotch just posterior to the pectoral n that
resembles a humeral blotch but appears only late in ontogeny).
Hemeraia is further distinguished from all species of Teleocichla,
which also have infraorbitals 4 and 5 co-ossied and (most of
them) a smooth preopercle, by having typical pike-cichlid morph-
ology without the various characters related to rheophyly found
in Teleocichla, including the pelvic pointed with third ray longest,
almost reaching the genital papilla, mouth downturned, reduced
cheek squation, and some degree of molarization of LPJ teeth.
Distribution: Hemeraia hemera is distributed in the Río Aripuanã
basin and H. chicha in the Río Juruena, Río Tapajós basin.
Remarks: e position of species of Hemeraia species within the
subtribe Crenicichlina is controversial in both previous taxonomic
studies and in our present work. Ploeg (1991) allocated Crenicichla
guentheri (junior synonym of C. hemera) among the species of the
C. saxatilis group, which herein corresponds to Saxatilia. Va rella
et al. (2012), in the description of Crenicichla chicha and based
mainly on external morphology, considered Hemeraia hemera and
H. chicha as sister-taxa by sharing features such as reduced ctenoid
squamation on anks, smooth preopercle, and infraorbitals 4 and
5 co-ossied. However, their placement in any of the major groups
of pike cichlids remained unclear because these species share
some characters (mostly related to meristics) with the C. saxatilis
(herein Saxatilia) group and others with the C. lugubris (herein
Lugubria) group (related to ontogenetic change of coloration).
e discordance of the position of Hemeraia between our dif-
ferent analyses may be partially due to the lack of molecular data,
together with the use of the rst 16 characters as continuous data.
e clade gathering Hemeraia, Saxatilia, and Lugubria, i.e. node
131 in ML tree (Fig. 2 le) and node 129 in the parsimony tree
based on EW/DiscreteMatrixTE (Fig. 2 right), is supported by
ve and six unambiguous synapomorphies, respectively. Among
these, characters 4 (number of dorsal-n rays) and 5 (number
of anal-n rays) are multistate quantitative characters treated
as continuous in the ContinuousMatrixTE. As a side-eect of
treating additive characters as continuous, these characters are
downweighted in the analysis and the resulting shis are dif-
ferent from the discretizations of the states performed a priori.
is side-eect would have been of lile impact if molecular data
of these species were available to be included in the analyses.
Node 140—Saxatilia, new genus
urn:lsid:zoobank.org:act:59D81458-AA56-426D-BC9E-
73781A861282
Type species: Perca saxatilis Linnaeus, 1758
Saxatilia is recovered as monophyletic with high sup-
port in the ML and BI trees (BS100%, PP 100%), and mod-
erate to low support in the parsimony analysis based on EW/
DiscreteMatrixTE (ABS 3/RBS 43%; Table 6). is genus com-
prises 23 valid species and corresponds almost entirely to the C.
saxatilis group, considered monophyletic in all previous studies
since Ploeg (1991; see Fig. 3). irty-one unambiguous mo-
lecular transformations and 11 morphological synapomorphies
(eight of them unambiguous) are optimized for the group
(Table 16).
Table 15. List of synapomorphies of the genus Hemeraia (Node 148) based on the optimization of the characters on the ML tree
Character transformation Apomorphic condition Observations
Char. 25: 1=>2 Reduction of the distribution of ctenoid (vs. cycloid)
scales on the dorsal portion of the body (paern 3).
Exclusive synapomorphy.
Char. 26: 2=>3 Reduction of the distribution of ctenoid (vs. cycloid)
scales on the ventral portion of the body (paern B3).
Exclusive synapomorphy.
Char. 46: 0=>1 Sexually dimorphic males with elongation of the median
rays of the caudal n, resulting in a lanceolate shape.
Convergent transformations occur in two less inclu-
sive groups within Wallaciia and Teleocichla.
Char. 71: 0=>1 Presence of a dark blotch on pectoral axilla. Convergent transformations occurring as synapo-
morphies of Crenicichla (Crenicichla), the C. (La.)
missioneira complex and of two less inclusive groups
withing Saxatilia and Lugubria.
Char. 105: 0=>1 Infraorbitals 4 and 5 fused, forming a median pore. Convergent transformations occurring as synapo-
morphy of Teleocichla and as autapomorphy of C.
(La.) igara.
Char. 127: 0=>1 Posterior border of the alveolar process of the
premaxilla curved, bulbous.
Highly homoplastic character, with several con-
vergent transformations inside the subtribe
Crenicichlina.
Ambiguous
Char. 87: 1 => 0 A Absence of sexual dimorphism of a colour paern
on the dorsal n comprising a well-dened dark margin
and a light submarginal area.
Representing a reversal of the apomorphic condition
found in the Node 131 (presence of this kind of sex-
ual dimorphism).
Downloaded from https://academic.oup.com/zoolinnean/advance-article/doi/10.1093/zoolinnean/zlad021/7204395 by Universidade de Sao Paulo (USP) user on 27 July 2023
34 • Var e l l a et al.
Diagnosis: Species of Saxatilia can be diagnosed from all other
pike cichlids by the presence of humeral blotch, which is evi-
dent from juvenile stages. Most species of Lugubria also have a
dark blotch posterior to the pectoral n that resembles a humeral
blotch but appears only later in ontogeny. Species of Saxatilia are
additionally distinguished from all pike cichlids by the combin-
ation of 31–71 scales in the E1 series, suborbital marking present,
serrated preopercle, slender lachrymal, and sexual dimorphism
in which males usually have light, small botches scaered on
anks and females show a round belly with a broad reddish or
purplish pigmentation.
Saxatilia also diers from Hemeraia by having most scales on
the ank ctenoid (combination of paerns 1 or 2 on the dorsal
area and paern B2 on the ventral area; see characters 25–26
and Supporting Information, Appendix S2, Fig. S2) vs. cycloid
(combination of paerns 3 and B3; see characters 25–26 and
Supporting Information, Appendix S2, Fig. S2). From Lugubria,
Saxatilia diers by the number of vertebrae (31–36 vs. 38–42
in Lugubria), from Wallaciia by the absence (vs. presence) of
serrations on the posterior margin of supracleithrum, and from
Teleocichla by all post-lachrymal infraorbitals autogenous (vs.
infraorbitals 4 and 5 co-ossied), pelvic n short and rounded
with second ray longest (vs. pointed with third ray longest almost
reaching genital papilla), and by the absence of osteological
modications related to rheophyly. Saxatilia is distinguished
from the subgenus Crenicichla by the presence (vs. absence) of a
dark blotch on the caudal n and by cycloid (vs. ctenoid) scales
on cheek and chest. From all species of the subgenus Batrachops
of Crenicichla, except C. (B.) jegui, Saxatilia diers by the ab-
sence of a reticulate colour paern on the ank (vs. having a re-
ticulate paern formed by the pigmentation on the base of each
scale); and from the subgenus Lacustria of Crenicichla by a sub-
orbital marking uniformly pigmented instead of formed by dark
punctulations more or less scaered on the suborbital region.
Distribution: Saxatilia is the most widespread genus of pike cich-
lids, occurring in all major rivers of South America, including the
Río Orinoco basin, the Amazon basin, all major drainages of the
La Plata basin, the coastal drainages of Trinidad and Tobago,
the Guianas, and Atlantic coastal drainages from north-eastern
Brazil to the Lagoa dos Patos and Lagoa Mirim systems in southern
Brazil and Uruguay. Recent collections from areas inuenced by
hydroelectric dams in coastal drainages in south-eastern Brazil (e.g.
Río Paraíba do Sul) have shown the presence of Saxatilia (prelim-
inarily identied as S. britskii), but it is unclear if these records rep-
resent native distribution or recent human introductions.
Table 16. List of synapomorphies of the genus Saxatilia (Node 140) based on the optimization of the characters on the ML tree
Character transformation Apomorphic condition Observations
Char. 3: 3=>2 Reduction in the number of dorsal-n spines to
17–19 spines.
Convergent transformations occurring as autapomorphies
of Teleocichla centisquama and Wallaciia heckeli.
Char. 12: 2=>3 Narrow lachrymal, much longer than deep. Convergences optimized as synapomorphy of Crenicichla
(Batrachops) and as autapomophy of Wallaciia wallacii.
Char. 13: 1=>0 Ascending process of premaxilla with similar
length or slightly shorter than the
dentigeous arm.
Highly homoplastic character inside the subtribe
Crenicichlina, with several convergent transformations
occurring within Lugubria and Crenicihla.
Char. 15: 1=>0 Frontal bones comparatively less compressed
(vs. compressed), as expressed by the ratio
inteorbital distance/neurocranium length.
is represents one of the many reversals to the
plesiomorphic condition occurring inside the subtribe
Crenicichlina.
Char. 68: 0=>1 Presence of a dark humeral blotch. Exclusive synapomorphy, without reversals.
Char. 84: 0=>1 Sexually dimorphic males showing scaered
light dots on anks.
Reversals occur as autapomorphies of Saxatilia inpa, S.
britskii, and S. labrina. e only convergence occurs as
autapomorphy of Lugubria multispinosa.
Char. 150: 0 => 2 Posterior margin of preopercle regularly serrated,
with short serrations.
Highly homoplastic character, with several transform-
ations in several directions occurring within the subtribe
Crenicichlina.
Char. 197: 0 => 1 Dorsomedial expansion of the lateral arm of the
epibranchial 4 very reduced, as a narrow crest, or
totally absent.
Highly homoplastic character, with several transformations
in both directions within the subtribe Crenicichlina.
Ambiguous
Char. 58: 0=>1 A Suborbital dark blotch semicircular or triangular,
restricted to the area close to the ventral margin
of the orbit.
ere is reversal occurring in the node 143, which in-
cludes Saxatilia species with suborbital marking ex-
tending caudoventrally, and convergences occurring
as autapomorphie of Hemeraia hemera and Lugubria
acutirostris.
Char. 98: 0 => 1 A Eyes visible in ventral view (eyes positioned
more laterally in the head)
Convergent transformations occurring as synapomorphies
of the subgenus Crenicichla (Crenicichla) and of Wallaciia.
Char. 202: 1 => 0 A Microbranchiospines regularly distributed
on the external (lateral) face of the second to
fourth branchial arches.
Highly homoplastic character with several transformations
in both directions (states 0 and 1) inside the subtribe
Crenicichlina.
Downloaded from https://academic.oup.com/zoolinnean/advance-article/doi/10.1093/zoolinnean/zlad021/7204395 by Universidade de Sao Paulo (USP) user on 27 July 2023
New generic classication of pike cichlids • 35
Node 130—Lugubria, new genus
urn:lsid:zoobank.org:act:D758B01C-2A66-4402-B63D-
A20FDA199C96
Type species: Crenicichla lugubris Heckel, 1840.
Lugubria is recovered with high support in the ML and
BI trees (BS100%, PP 100%) but low support in the par-
simony analysis based on EW/DiscreteMatrixTE (ABS
6/RBS 25%; Table 6). The genus comprises 16 valid
species and corresponds to the Crenicichla lugubris group
of Kullander (1991, 1997; excluding Crenicichla vittata)
and Ploeg (1991; excluding C. vittata and C. jegui; Fig. 3I).
This group has been recovered by previous studies based on
phylogenomics and good taxonomic coverage (Burress et al.
2017, 2018; Fig. 3D–F). Fourteen unambiguous molecular
transformations and 15 morphological synapomorphies
(nine of them unambiguous) are optimized for the genus
(Table 17).
Table 17. List of synapomorphies of the genus Lugubria (Node 130) based on the optimization of the characters on the ML tree
Character transformation Apomorphic condition Observations
Char. 0: 1=>2 Number of scales in the E1 row increasing to
79–123.
Convergences are optimized as autapomorphy of
Crenicichla (La.) viata and Teleocichla centisquama.
Char. 3: 3=>4 Number of dorsal n spines increasing
to 23–25.
Convergence occurs only in a subgroup of Crenicichla
(Batrachops) and a reversal occur as autapomorphy of
Lugubria johanna.
Char. 7: 2=>3 Number of vertebrae increasing to 39–44. Convergences optimized as autapomorphies of C (B.)
jegui and C. (La.) viata.
Char. 18: 1=>2 Number of scale rows between the anterior and
posterior branches of the lateral line increasing
to 4–5 rows
Convergences occur as autapomorphies of Hemeraia chi-
cha, S. ploegi, and C (La.) viata and a reversal occurs in
the node 136 inside Lugubria.
Char. 33: 0=>1 Basal portion of the pectoral n with scales on
inter-radial membranes, embedded in skin.
A reversal occurs in the node 136 and convergences occur
as synapomorphy of the subgenus Crenicichla (Crenicichla)
and as autapomorphies of C. (B.) reticulata and Wallaciia
urosema (only one specimen observed with these scales).
Char. 43: 1=>0 Sexually dimorphic males not showing
accentuated elongation of the posteriormost
dorsal-n rays.
e lack of sexual dimorphism related to elongation
of unpaired ns occur repeatedly inside the subtribe
Crenicichlina, resulting in a high degree of homoplasy for
the characters 43 and 44.
Char. 44: 1=>0 Sexually dimorphic males not showing
accentuated elongation of the posteriormost
anal-n rays.
Char. 150: 0=>3 Posterior margin of preopercle strongly serrated,
with spine-like proeminent serrations.
Inside the subtribe Crenicichlina, another convergent
transformation is optimized as synapomorphy of Lugubria
and a reversal occurs in Wallaciia heckeli.
Char. 177: 1=>2 Posterior foramen of the nasal unique or divided,
diplaced anterior and with the posterior
portion of the canal ossied, resulting in the
openings placed at the middle of the nasal canal.
Very homoplastic character. State 2 represents a second-
ary modication of the synapomorphic condition of pike
cichlids (state 1).
Char. 197: 0 => 2 Dorsomedial expansion of the lateral arm of the
epibranchial 4 very reduced, as a narrow crest, or
totally absent.
Highly homoplastic character. Several transformations
in dierent directions within the subtribe Crenicichlina
and with several secondary transformations to the state 1
within the genus Crenicichla.
Ambiguous
Char. 8: 2=>3 Number of caudal vertebrae contained
within the caudal peduncle increasing to ve or
more vertebrae + halfcentrum.
Convergences already listed in the table of synapo-
morphies for Teleocichla.
Char. 17: 1=>2 A Number of scale rows on the area bwteen the
last spine of dorsal n and the anterior branch
of lateral line.
Convergent transformations occur as autapomorphies of
C (B.) jegui and C. (La.) viata. A secondary transform-
ation (2=>3) occurs inside Lugubria (Node 135).
Char. 32: 1=> 2 A Scales on the midlateral area of body ovoid, with
long axis horizontal.
Convergent transformations occur as autapomorphies of
T. centisquama and T. cinderella. A reversal to state 1 occurs
only in Lu. lugubris.
Char. 175: 1=>2 Nasal canal long and nearly straight. Convergences occur repeatedly within the genus
Crenicichla.
Char. 183: 1=> 0 A Uncinate process and anterior arm of the
epibranchial 1 with similar width (from uncinate
process wider).
is character is very homoplastic along the entire top-
ology and also includes a reversal in Lu. tigrina and in the
Node 137.
Downloaded from https://academic.oup.com/zoolinnean/advance-article/doi/10.1093/zoolinnean/zlad021/7204395 by Universidade de Sao Paulo (USP) user on 27 July 2023
36 • Var e l l a et al.
Diagnosis: Lugubria includes species with relatively large body
(max. SL 225–300mm) distinguished from all other pike cich-
lids by the combination of reduced size of the scales in the whole
body reected in the high number in the E1 row (88–123 scales;
see Remarks), the high count of so rays in the dorsal and anal
ns (13–18 and 9–13, respectively), and the high number of
vertebrae (39–44 vertebrae). Crenicichla (Lacustria) viata is
similar to species of Lugubria by having high counts of vertebrae
and scales in the E1 row but diers from all species of that genus
by females showing orange or reddish pigmentation on the lat-
eral portion of the abdomen, whereas in species of Lugubria this
paern is not observed—in most species, sexually dimorphic
females have a broad reddish or purplish broad pigmentation
on the ventral portion of the abdomen (belly). Crenicichla
(Lacustria) viata also shows a suborbital marking composed of
distinct dark spots that together form a stripe, which is charac-
teristic of other species of the subgenus Lacustria of Crenicichla,
whereas species of Lugubria do not show suborbital marking or
have it uniformly pigmented. Finally, most species of Lugubria
show a dark blotch posterior to the pectoral n that appears only
late during ontogeny and is absent in C. (La.) viata.
Species of Lugubria are further distinguished from Saxatilia
by the absence (vs. presence) of a humeral blotch that appears
early in ontogeny and remains in adults; from Hemeraia and
Teleocichla by having all post-lachrymal infraorbitals autogenous
(vs. infraorbitals 4 and 5 co-ossied) and posterior margin of
preopercle strongly serrated (vs. preopercle smooth or with weak
serrations irregularly distributed along the posterior margin);
and from Wallaciia by the absence (vs. presence) of serrations on
the posterior margin of the supracleithrum. Species of Lugubria
are additionally distinguished from the subgenus Crenicichla by
having all scales on head cycloid (vs. scales ctenoid on cheek and
predorsal area) and by the presence (vs. absence) of a dark blotch
on the caudal n, with the exception of Lugubria johanna and Lu.
monicae, which lack the caudal n blotch. Species of Lugubria are
futher distinguished from all species of the subgenus Batrachops
of crenicichla, except C. (B.) jegui, by the absence of a reticulate
paern of coloration on the ank (vs. presence of a reticulate pat-
tern resulting from the dark pigmentation on the base of each
scale on the ank).
Distribution: Lugubria is distributed in all major tributaries of the
Amazon basin, in the Río Orinoco basin, and coastal drainages
of the Guianas. ere are no records of Lugubria in the Atlantic
coastal drainages or the La Plata basin.
Remarks: Since Kullander (1991, 1997), species of the C. lugubris
group have been subdivided into the C. acutirostris group and
C. lugubris group s.s. (e.g. Montaña et al. 2008; Kullander and
Varella 2015). Crenicichla (Lacustria) viata and C. (Batrachops)
jegui were tentatively allocated in the C. acutirostris group, which
has recently been contested (Piálek et al. 2012; Burress et al.
2017, 2018; this paper; see Fig. 3). e distinction between
these subgroups was based on a series of characters in combin-
ation, including a depressed head and long, pointed snout in the
C. acutirostris group, with nostril halfway between the postlabial
skin fold and the anterior margin of orbit vs. blunt snout in
the C. lugubris group s.s., with nostril closer to postlabial skin
fold, as well as some details of coloration related to ontogeny.
All analyses performed herein recover a Lu. lugubris group s.s.
(Node 137) as monophyletic. Among the synapomorphies for
this group, some are related to the blunt snout (char. 13: 1=>0
and char. 96: 0=>1) and the changes on the coloration during
ontogeny, i.e. the presence of dark spots scaered on the an-
terior portion of the body (in juveniles) and the presence (in
adults) of a dark blotch posterior to the pectoral n and another
on the pectoral axilla (chars. 70: 0=>1; char. 71: 0=>1; and 81:
0=>1). e Lu. acutirostris group sensu Kullander (1991, 1997),
including all long-snout species of Lugubria, is recovered only in
the parsimony analyses under extended implied weighting but
not recovered in our main hypothesis (ML tree), in the BI tree
and in the parsimony analyses under equal weighting.
Character congruence and insights on phenotypic
diversication
Discussions on macroevolution of the clade of pike cichlids
(subtribe Crenicichlina) have been focused on the subgenus
Lacustria of Crenicichla, which is restricted to the rivers of
southern South America (Piálek et al. 2012, 2019a; Burress et
al. 2017, 2018). ese studies suggest the existence of adaptive
radiations in the C. (La.) missioneira complex in the Río Uruguay
and of the C. (La.) mandelburgeri complex in the Río Paraná.
Except for their occurrence in complex riverine (instead of lacus-
trine) habitats, these radiations were considered similar to the
species ocks in the African Great Lakes and Middle America
(Piálek et al. 2012), i.e. monophyletic assemblages of closely
related species (shallow genetic divergence) that coexist in the
same area with a high level of endemicity and ecomorphological
specialization.
Our results agree with these previous studies in demonstrating
the occurrence of parallel evolution of ecomorphological traits
in geographically isolated habitats between the C. missioneira
and C. mandelburgeri complexes, most of them related to feeding
and microhabitat occupation (e.g. Burress et al. 2018; Piálek et al.
2019a). Several characters present in these clades correlated to
rheophyly and durophagy were highly homoplastic among pike
cichlids. ese characters were clearly homoplastic and phylo-
genetically misleading when morphological characters were
analysed alone, resulting in lack of resolution and long-branch
araction.
More generally, ecomorphological convergence is rampant
beyond the well-known cases within the subgenus Lacustria,
with clear examples among multiple pike cichlid taxa coexisting
in complex environments such as rapids. Examples are the as-
semblages comprising C. (B.) jegui, C. (B.) cametana, and C. (B.)
cyclostoma occurring in sympatry in rapids of the Río Tocantins
basin and the diverse assemblages of Teleocichla and Lugubria
in rapids of the Río Xingu basin. Varella et al. (2018) discussed
characteristics related to microhabitat occupation and feeding
within the assemblage of Teleocichla from the Río Xingu, most
of them reanalysed herein in a phylogenetic framework and con-
rmed as convergent. Likewise, Varella and Ito (2018) explored
the large diversity of the C. lugubris group (=Lugubria) in the Río
Xingu basin and highlighted the convergent dark coloration be-
tween Lugubria dandara, Teleocichla preta, and Crenicichla (La.)
hu, which is comparable to that of Crenicichla (B.) cyclostoma
and C. (B.) cametana. Dark coloration in these species might
be understood as a convergent adaptation for inhabiting swi,
Downloaded from https://academic.oup.com/zoolinnean/advance-article/doi/10.1093/zoolinnean/zlad021/7204395 by Universidade de Sao Paulo (USP) user on 27 July 2023
New generic classication of pike cichlids • 37
clear-water rapids over beds of dark rocks widespread in the
Brazilian Shield.
On the other hand, the diversity of the genera Lugubria,
Saxatilia, and Wallaciia and of the subgenus Batrachops of
Crenicichla suggest widespread allopatric divergence resulting in
ecomorphologically similar species across river basins. Most spe-
cies or populations within their widespread species (see: Říčan
et al. 2021a) are distributed following hydrography or geological
history. In these groups, there is widespread variation in sexu-
ally selected traits (sexual dimorphism and dichromatism), but
traits associated with feeding or habitat occupation vary much
less, with ecomorphological variation mostly restricted to pairs
or small groups of sympatric species living in fast current envir-
onments.
From our study, it remains unclear if sympatric assem-
blages diversified in situ (like those of Lacustria radiations
in the Uruguay and Paraná rivers), because our taxonomic
sampling, character coding, and analyses focused on as-
sessing monophyly clarifying relationships among the major
groups of pike cichlids, sacrificing resolution of less-inclusive
groups. The taxonomic sampling of recent phylogenomic
analyses (Piálek et al. 2012, 2019a; Burress et al. 2017, 2018)
also prevents inference of relationships within these less
studied sympatric assemblages. Nevertheless, Říčan et al.’s
(2021a) recent study using a large taxonomic sampling but
only two mtDNA markers suggests monophyly of the assem-
blage of Lugubria in the Río Xingu basin and of Crenicichla
(Batrachops) in the Río Tocantins basin, but not of the large
sympatric assemblage of Teleocichla in the Río Xingu basin.
Říčan et al. (2021a, b) studies also suggest the existence of
smaller sympatric groupings of Crenicichla (Lacustria) that
evolved parallel ecomorphs in different rivers basins of the
Río Paraná basin.
In that context, resource partitioning apparently plays an im-
portant role in the origin or maintenance of the phenotypic di-
versity among rheophilic pike cichlids in dierent river basins or,
more generally, among coexisting but phylogenetically distant
species that occupy more or less specialized ecological niches.
Moreover, resource partitioning apparently plays a role even
within the already specialized rheophilic assemblage of Teleocichla
in the Río Xingu (Zuanon 1999; Varell a et al. 2016). From this
perspective, the role of ecological speciation may be a fruitful
avenue for subsequent studies on the evolution of pike cichlids
beyond the C. (La.) missioneira and C. (La.) mandelburgeri com-
plexes. Recent studies with broader taxonomic scope (e.g. Arbour
and López-Fernández 2014; Astudillo-Clavijo et al. 2015) place
the phenotypic and funcional diversication of pike cichlids
in the broader context of the tribe Geophagini and Neotropical
cichlids, highlighting that pike cichlids occupy a vast and largely
unique functional morphospace among their South American re-
latives. However, because few representatives of pike cichlids were
included in these analyses, much about the evolution within the
clade remains unexplored and will be the subject of future studies.
CONCLUSION
We performed multiple phylogenetic analyses on combined
datasets that included our original morphological data and
complementary published molecular data, being the first
phylogenetic analyses of pike cichlids to include morpho-
logical characters. In the morphological datasets, we codi-
fied features of several distinct complexes (coloration, fins,
squamation, sexual dimorphism, and osteology) that can be
assigned to a variety of putative functions: communication
and mate behaviour, feeding, foraging strategies, and swim-
ming. We included a comprehensive and well-balanced taxo-
nomic sampling that spans virtually all morphological and
ecological diversity of pike cichlids. We also used a variety
of data treatments and analytical approaches (criteria of op-
timality and weighting schemes) to assess variation among
our main hypotheses (ML tree) and multiple alternatives.
We discuss our results on the context of their fit with pre-
vious comprehensive studies and used our main phylogen-
etic hypothesis to revise the higher-level taxonomy of pike
cichlids.
Our results agree with almost all previous molecular phylo-
genetic studies of pike cichlids in recognizing the monophyly
of the major species groups traditionally used in the taxonomy
of the former genus Crenicichla, as well as the monophyly of
Teleocichla nested within Crenicichla s.l., making it paraphy-
letic. Accordingly, our results and previous studies agree
that the relationships between the groups are still not well
resolved, but also that the type species of Crenicichla (i.e. C.
macrophthalma) is consistently more closely related to species
formerly allocated in the C. reticulata group and C. lacustris
group than to any other group of pike cichlids. Considering
the relative stability of this scenario and the desirability of a
taxonomic classication reective of evoluationry relation-
ships, we propose a revision of generic and subgeneric nomen-
clature for the major monophyletic groups of pike cichlids. As
a result, we restricted the genus Crenicichla to the type species
(erected as a monotypic subgenus Crenicichla), and formalized
two of the formerly proposed species groups as the Crenicichla
subgenera Batrachops and Lacustria. Teleocichla is recognized
as a valid and easily diagnosable genus of rheophilic pike cich-
lids, and the remaining clades were elevated to new genera
(Wallaciia, Hemeraia, Saxatilia, and Lugubria). All newly de-
ned clades are relatively well supported and clearly diagnos-
able on the basis of morphological synapomorphies that are
consistently reconstructed on both ML tree and our most par-
simonious trees.
While the scope of this paper is strictly taxonomic, our re-
sults provide some insights on possible foci for phenotypic
and ecological diversication in pike cichlids. Our results sug-
gest that resource partitioning in enviroments with owing
water and rocky beds might have played a role in the origin
or maintainance of the great diversity of pike cichlids we see
today. In contrast, the diversity in other caldes of pike cich-
lids appears to be the result of passive, biogeographic pro-
cesses resulting in a variety of drainage-specic lineages with
muted ecomorphological diversication. Our study brings
a previously unavailable wealth of morphological characters
to the study of pike cichlids that, combined with emerging
phylogenomic datasets, should contribute to further our
understanding of the relationships and evolutionary history of
the South American pike cichlids.
Downloaded from https://academic.oup.com/zoolinnean/advance-article/doi/10.1093/zoolinnean/zlad021/7204395 by Universidade de Sao Paulo (USP) user on 27 July 2023
38 • Var e l l a et al.
KEY TO GENE AND SUBGENE OF PIKE CICHLIDS
We provide below a key for the identication of the six genera of pike cichlids (one of them with three subgenera), trying to be
as succinct as possible and using easy-to-observe characteristics to make it aractive to use. However, we strongly advice to also
check the more-in-depth comparisons performed in the diagnosis of each genus to reach reliable generic identications.
1. Presence of a humeral blotch even in earlier stages of development ......................................................................................Saxatilia
1’. Absence of humeral blotch or presence of a darkening just posterior to the pectoral n only in late stages of ontogeny (large
subadults and adults of Lugubria and Hemeraia hemera)........................................................................................................................2
2. Small scales in great number on the body, more directly expressed by the higher number of E1 scales (88–123 scales); high
number of total vertebra (38–44 vertebrae); ontogenetic colour change generally occuring: juveniles with several small dark
dots on the anterior portion of the body (disappearing in adults), and a round darkening posterior to the pectoral n appear-
ing in adults ........................................................................................................................................................................................Lugubria
2’. Low to moderate number of scales in the E1 row (41–79 scales), with the exception of C. (La.) viata with 79–93 scales and
Teleocichla centisquama (85–96 scales); moderate number of total vertebra (31–39 vertebrae; C. viata 40–41); absence of
such ontogenetic colour change (Hemeraia hemera as exception) ....................................................................................................... 3
3. Infraorbitals 4 and 5 co-ossied, forming a median pore .......................................................................................................................4
3’. All post-lachrymal infraorbital autogenous [see exceptions of this character in the diagnosis of the subgenus Crenicichla
(Lacustria)] ......................................................................................................................................................................................................5
4. Small size (max. SL 48–90mm, with exception of Teleocichla preta with max. 121.3mm SL); pelvic n long and pointed, with
third ray longest (almost reaching the genital papilla) and with a skin thickening on the lateral portion; mouth hypognathous;
cheek squamation restricted to its posterior portion; rostral pore of the nasal canal displaced posteriorly from the postlabial
margin of the snout; two or more caudal than precaudal vertebrae; absence of suborbital marking ...........................Teleocichla
4’. Medium size (max. SL 138–232mm); pelvic n short and rounded, with second ray longest not reaching the genital papilla;
mouth prognathous; regular cheek squamation; rostral pore of the nasal canal on the postlabial margin of the snout; more
precaudal than precaudal vertebrae or equal numbers of precaudal and caudal vertebra; presence of suborbital marking
.............................................................................................................................................................................................................Hemeraia
5. Small size (max. SL 52–85mm); large eyes (orbital diameter 7.8–12.6% of SL, with minimum between 7.8% and 8.6%);
presence of serrations on the posterior margin of supracleithrum (except Wallaciia heckeli); posterior margin of preopercle
strongly serrated, with spiny-like proeminent serrations; pectoral n with 13–14 rays (in modal values); reduced predorsal
squamation (not extending onto the NFL0) ..............................................................................................................................Wallaciia
5’. Medium size (max. SL 95–294mm); relatively small eyes in adults (stronger negative allometry—orbital diameter 4.6–
10.9% of SL, with minimum between 4.6% and 7.5%), with exception of C. macrophthalma (orbital diameter 9.4–10.6% of
SL); posterior margin of preopercle smooth; preopercle with various degrees of ornamentation (from smooth to weak or
short serrations) but not characterized as proeminent spiny projections as in Wallaciia; pectoral n with 15–17 modally;
predorsal squamation exting onto the NFL0 .......................................................................................................................... Crenicichla
Subgenera of Crenicichla
Sub1. Absence of sexual dimorphism of coloration; large eyes (orbital diameter 9.4–10.6% of SL); body almost entirely covered
by ctenoid instead of cycloid scales (scales ctenoid on cheek, ank squamation following paerns 0 and B1, and scales cten-
oid covering almost entirely the caudal n); absence of a caudal blotch ..................................................Crenicichla (Crenicichla)
Sub1’. Sexual dimorphism of coloration including orange pigmentation on the lateral abdomen of mature males (present in both
sexes in few species, but still more proeminent in males) and dark blotches on the dorsal n of mature females (occasionally
modied into an horizontal stripe in several species); relatively smaller eyes with strong negative allometry (orbital diameter
4.6–10.9% of SL, with minimum between 4.6% and 7.5%); ctenoid scales restricted to the anks and cycloid scales covering
the head and anteriormost portion of the dorsum and on pre-pelvic-n area (exception is C. viata that shows squamation
similar to C. macrophthalma); presence of dark caudal blotch .......................................................................................................Sub2
Sub2. Body laterally compressed; presence of dark suborbital punctulations (with various aspects); absence of reticulate colour
paern on ank resulted from the dark pigmentation on the base of the scales; nostrils halfway from the anterior margin of
orbit and the margin of postlabial snout ............................................................................................................. Crenicichla (Lacustria)
Sub2’. Body cylindrical or depressed (C. cyclostoma as exception); absence of dark suborbital punctulations (C. jegui has a uni-
formly pigmented dark suborbital bar); presence of a reticulate colour paern on ank resulted from the dark pigmentation
on the base of the scales (C. jegui as exception); nostrils situated on the margin of postlabial snout (exceptions are C. jegui,
C. cyclostoma and C. cametana) .......................................................................................................................... Crenicichla (Batrachops)
Downloaded from https://academic.oup.com/zoolinnean/advance-article/doi/10.1093/zoolinnean/zlad021/7204395 by Universidade de Sao Paulo (USP) user on 27 July 2023
New generic classication of pike cichlids • 39
SUPPLEMENTARY DATA
Supplementary data is available at XXXXXX Journal online.
Appendix S1. Commented revision of the literature about
pre-phylogenetic taxonomy of pike cichlids.
Appendix S2. Character statements: description, comments,
and illustrations of the 216 morphological characters used in the
DiscreteMatrix and ContinuousMatrix.
Appendix S3. Other alternative trees obtained from the ana-
lyses of the combined datasets: Bayesian Inference tree, based on
a partitioned analysis of 216 discrete morphological characters
and 5328bp distributed along 7 genes; strict consensus tree from
2 most parsimonious trees (12126.85 steps) obtained from the
Parsimony analysis of 16 continuous and 200 discrete morpho-
logical characters and 5328bp distributed along 7 genes using
equal weighting; strict consensus tree from 192 most parsimoni-
ous trees (436.1 steps) obtained from the Parsimony analysis of
216 discrete morphological characters and 5328bp distributed
along 7 genes using extended implied weighting (k=12); strict
consensus tree from 2 most parsimonious trees (478.8 steps)
obtained from the Parsimony analysis based on 16 continuous
and 200 discrete morphological characters and 5328bp distrib-
uted along 7 genes using extended implied weighting.
File S1. List of material used for character coding in the mor-
phological dataset.
File S2. Matrices of the 216 morphological characters used in
the combined analyses, including references of previous usage.
In DiscreteMatrix, all 216 characters are discretized a priori. In
ContinuousMatrix, the rst 16 characters are treated as continu-
ous and 200 as discrete.
File S3. Tables used for normalization of the continu-
ous data for the 16 rst morphological characters of the
ContinuousMatrix.
File S4. List of all the terminal taxa used in the integrative ana-
lysis with accession codes for GenBank or BOLD.
File S5. Resulting alignments derived from published sources
(seven genes, total 5328bp).
File S6. Information on the optimization of the morpho-
logical characters on the ML tree. Part I: Comparison between
the optimizations of the synapomorphies on the ML tree (main
text: Figure 2 le) and the most parsimonious trees obtained
from the analysis of the DiscreteMatrixTE under equal weight-
ing scheme (Pars EW/DiscreteMatrixTE – main text: Figure 2,
right). Part II: Synapomorphies of all clades according to the op-
timization on the ML tree. Part III: Reconstructions of all mor-
phological characters according to the optimization on the ML
tree. Part IV: For all characters, the minimum (min) and max-
imum possible steps (max), the observed steps according to the
optimization on the ML tree (obs.), the consistency (ci), reten-
tion (ri) and rescaled consistency index (rc) of each character.
ACKNOWLEDGEMENTS
anks to the curators of all the institutions that provided ma-
terial and/or welcomed the rst author during the visits to ich-
thyological collections: Ricardo Castro and Flávio Bockmann
(Laboratório de Ictiologia de Ribeirão Preto, Universidade de São
Paulo – [LIRP]-USP); Júlio Garavello and Alexandre Oliveira
(Laboratório de Ictiologia Sistemática da Universidade Federal de
São Calos [LISDEBE/UFSCar]), Carla Pavanelli and Weferson da
Graça (Núcleo de Pesquisas em Limnologia, Ictiologia e Aqüicultura,
Universidade Estadual de Maringá [NUPELIA]); Oscar Shibaa
(Museu de Zoologia, Universidade Estadual de Londrina [MZUEL]);
Vinicius Abilhoa (Museu de História Natural do Capão do Imbuia,
Curitiba [MHNCI]); Roberto Reis e Carlos Lucena (Museu de
Ciências e Tecnologia da Pontifícia Universidade Católica do Rio
Grande do Sul [MCP/PUCRS]); Amalia Miquelarena, Hugo López
and Jorge Cascioa (Instituto de Limnología de La PLata [ILPLA]
and Museo de La Plata [MLP]); Ricardo Ferriz (Museo Argentino de
Ciencias Naturales "Bernardino Rivadavia", Buenos Aires - MACN);
Lúcia Rapp Py-Daniel and Jansen Zuanon (Instituto Nacional de
Pesquisas da Amazônia [INPA]), Carolina Doria (Universidade
Federal de Rondônia [UFRO]); James Maclaine, Ralf Britz and Oliver
Crimmen (Natural History Museum [NHM/BMNH]); Sonia Fisch-
Muller and Raphael Covain (Museum d'Histoire Naturelle, Ville de
Genève [MHNG]); Ronald de Ruiter (Naturalis—ZMA, RMNH);
Ernst Mikschi and Anja Palandacic (Naturhistorisches Museum, Wien
[NMW]); Patrice Pruvost (Muséum National d'Histoire naturelle,
Paris [MNHN]); Kevin Swagel and Caleb McMahan (Field Museum
of Natural History, Chicago [FMNH]); Luiz Rocha, Dave Catania and
Jon Fong (California Academy of Sciences, San Francisco [CAS]);
Mark Sabaj (Academy of Natural Sciences of Drexel University,
Philadelphia [ANSP]). An expanded gratitude to the sta and students
of most of these institutions for the discussions during the development
of this study. anks to Dr Carlos Lucena, Prof. Maria Isabel Landim
(Museu de Zoologia da Universidade de São Paulo [MZUSP]), Prof.
Mônica Ragazzo (Instituto de Biociências da Universidade de São
Paulo [IB/USP]), Prof. Aléssio Datovo (MZUSP), and Prof. George
Maox (Universidade Federal de São Carlos, Sorocaba [UFSCar])
for the initial discussions on the proposal of the doctorate project and
for their appreciation of the resulting dissertation precursor to the
morphological portion of this study. anks to Prof. Vitor Tagliacolo
(Universidade de Uberlândia [UFU]), Dr Gustavo Ballen (MZUSP),
and to Prof Murilo Pastana (MZUSP) for discussions on the phylo-
genetic analyses presented in this paper. anks to Dr. Samuel Burstein
(UMMZ-UMICH) for the important assistance on obtaining data
from Genbank and on the phylogenetic analyses performed in this
paper. anks to Dr. Samuel Burstein (University of Michigan,
Museum of Zoology - UMICH) for the important assistance on
obtaining data from Genbank and on the phylogenetic analyses per-
formed in this paper. anks to Jens Gowald, Oliver Lucanus, José
Birindelli (MZUEL), Luiz Roberto Malabarba (Universidade Federal
do Rio Grande do Sul [UFRGS]), Marcelo Krause, and Nathan Lujan
Royal Ontario Museum [ROM] for the pictures of live specimens
that illustrate Appendix S2, as well as to many aquarists and cichlid
enthusiasts who made their pictures available in the aquarium lit-
erature, on the web, or made some pictures available directly to HV.
Special thanks to the curators, professors, sta, and students of the
Museu de Zoologia da USP that contributed to the development
of the doctorate project. HV was funded by Fundação de Amparo à
Pesquisa do Estado de São Paulo—FAPESP (grants 2011/14630-
0, 2013/17056-9, 2015/21901-1 and 2018/2018/24086-5) and
CNPq (grant 150304/2017-0). CO received nancial support from
FAPESP through grants 2018/20610-1, 2016/09204-6, 2014/26508-
3 and from Conselho Nacional de Desenvolvimento Cientíco e
Tecnológico - CNPq proc. 306054/2006-0. HLF was supported by
the University of Michigan, the Royal Ontario Museum, and Natural
Sciences, and Engineering Research Council of Canada (NSERC)
through Discovery Grants (371212-2009 and 2014-05374). HV
would like to dedicate this paper to Daniela Florenzano, as gratitude
for her unconditional support, emotionally and nancially, during the
whole period of development of this study.
Downloaded from https://academic.oup.com/zoolinnean/advance-article/doi/10.1093/zoolinnean/zlad021/7204395 by Universidade de Sao Paulo (USP) user on 27 July 2023
40 • Var e l l a et al.
DATA AVAILABILITY
e molecular data underlying this article are available in GenBank
Nucleotide Database at hps://www.ncbi.nlm.nih.gov/nucleotide/
and on Barcode of Life Data System V4 at hps://www.boldsystems.
org/. All sequences can be accessed using the accession codes avail-
able in the Supporting Information, File S4. e resulting alignments
of the seven genes used in the analyses are available in the Supporting
Information, File S5. e morphological data used in the combined
analyses are available in the Supporting Information, File S2 (morpho-
logical matrices) and Supporting Information, File S2 (tables used for
normalization of continuous data).
CONFLICTS OF INTEREST
e authors declare no conicts of interest.
REFERENCES
Agnarsson I, Miller JA. Is ACCTN beer than
DELTN?. Cladistics 2008;24:1032–1038. hps://doi.
org/10.1111/j.1096-0031.2008.00229.x
Arbour JH, López-Fernández H. Ecological variation in South American
geophagine cichlids arose during an early burst of adaptive morpho-
logical and functional evolution. Proc R Soc B 2013;280:1–9.
Arbour JH, López-Fernández H. Adaptive landscape and functional
diversity of Neotropical cichlids: implications for the ecology
and evolution of Cichlinae (Cichlidae; Cichliformes). J Evol Biol
2014;27:2431–42.
Astudillo-Clavijo V, Arbour JH, López-Fernández H. Selection to-
wards dierent adaptive optima drove the early diversication of
locomotor phenotypes in the radiation of Neotropical geophagine
cichlids. BMC Evol Biol 2015;15:1–13. hps://doi.org/10.1186/
s12862-015-0348-7
Baumgarten L, Machado-Schiano G, Henning F, et al. What big lips
are good for: on the adaptive function of repeatedly evolved hyper-
trophied lips of cichlid shes. Biological Journal of the Linnean Society
2015;115:448–455. hps://doi.org/10.1111/bij.12502
Bloch ME, Schneider JG. M. E. Blochii, Systema Ichthyologiae Iconibus
cx Ilustratum. Post obitum auctoris opus inchoatum absolvit, correxit,
interpolavit Jo. Golob Schneider, Saxo. Berolini. Sumtibus Auctoris
Impressum et Bibliopolio Sanderiano Commissum 1801;i–lx + 1–584, pls
1–110. hps://www.biodiversitylibrary.org/item/129130#page/7/
mode/1up
Bradford YM, Van Slyke CE, Ruzicka L, et al. Zebrash information
network, the knowledgebase for Danio rerio research. Genetics
2022;220:iyac016. hps://doi.org/10.1093/genetics/iyac016
Bremer K. Branch support and tree stability. Cladistics
1994;10:295–304.
Britski HA, Luengo JA. Sobre Crenicichla jupiaensis n.sp., espécie aberrante
do Río Paraná (Pisces, Cichlidae). Pap Avulsos Zool 1968;21:169–82.
Burress ED, Tan M. Ecological opportunity alters the timing and shape
of adaptive radiation. Evolution 2017;71:2650–2660. hps://doi.
org/10.1111/evo.13362
Burress ED, Piálek L, Cascioa JR, et al. Island- and lake-like par-
allel adaptive radiations replicated in rivers. Proc R Soc B: Biol Sci
2017;285:20171762. hps://doi.org/10.1098/rspb.2017.1762
Burress ED, Alda F, Duarte A, et al. Phylogenomics of pike cichlids
(Cichlidae: Crenicichla): the rapid ecological speciation of an incipient
species ock. J Evol Biol 2018;31:14–30. hps://doi.org/10.1111/
jeb.13196
Burress ED, Piálek L, Cascioa JR, et al. Rapid parallel morphological
and mechanical diversication of South American pike cichlids
(Crenicichla). Syst Biol 2022; syac018. hps://doi.org/10.1093/
sysbio/syac018r
Cascioa JR, Arratia G. Jaws and Teeth of American Cichlids (Pisces:
Labroidei). J Morphol 1993a;217:1–36.
Cascioa, JR, Arratia G. Tertiary cichlid shes from Argentina and re-
assessment of the phylogeny of New World cichlids (Perciformes:
Labroidei). Kaupia 1993b;2:195–240.
Cascioa JR, Almirón A, Aichino D, et al. Crenicichla taikyra (Teleostei:
Cichlidae), a new species of pike cichlid from the middle río Paraná,
Argentina. Zootaxa 2013;3721:379–86.
Chakrabarty P. A morphological phylogenetic analysis of Middle
American cichlids with special emphasis on the section ‘Nandopsis’
sensu Regan. Misc Publ Mus Zool Univ Mich 2007;198:1–31.
Chernomor O, von Haeseler A, Minh BQ. Terrace aware data structure for
phylogenomic inference from supermatrices. Syst Biol 2016;65:997–
1008. hps://doi.org/10.1093/sysbio/syw037
Cichocki F. Cladistic history of cichlid shes and reproductive strategies of the
American genera Acarichthys, Biotodoma and Geophagus. Unpublished
Ph.D. Dissertation, e University of Michigan, 1976.
Clark K, Karsch-Mizrachi I, Lipman DJ, et al. GenBank. Nucleic Acids Res
2016;44:D67–72. hps://doi.org/10.1093/nar/gkv1276
Colombo M, Diepeveen ET, Muschick M, et al. e ecological and gen-
etic basis of convergent thick-lipped phenotypes in cichlid shes.
MolEcol 2013;22: 670–684.
Datovo A, Bockmann FA. Dorsolateral head muscles of the catsh
families Nematogenyidae and Trichomycteridae (Siluriformes:
Loricarioidei): comparative anatomy and phylogenetic analysis.
Neotrop Ichthyol 2010;8:193–246.
Devincenzi GJ. Peces del Uruguay. Notas complementarias, III. Anales
Mus Hist Nat Montevideo 1939;4:1–37.
Edgar RC. MUSCLE: multiple sequence alignment with high accuracy
and high throughput. Nucleic Acids Res 2004;32:1792–7.
Eigenmann CH. On a collection of shes from Buenos Aires. Proc Wash
Acad Sci 1907;8 (for 1906):449–58.
Eigenmann CH. Catalogue of the fresh-water shes of tropical
and south temperate America. In: Sco WB, (ed), Reports of
the Princeton University Expeditions to Patagonia 1896-1899.
Zoology. E. Schweizerbarbart'sche. 1910, 375–511. hps://www.
biodiversitylibrary.org/item/18372#page/5/mode/1up
Eigenmann CH, Bray WL. A revision of the American Cichlidae. Ann N Y
Acad Sci 1894;7:607–24.
Elmer KR, Lehtonen TK, K au AF, et al. 2010. R apid sympatric ecological
dierentiation of crater lake cichlid shes within historic times. BMC
Biology 2010;8:1–15. hps://doi.org/10.1186/1741-7007-8-60.
Farias IP, Ortí G, Sampaio I, et al. Mitochondrial DNA phylogeny of the
family cichlidae: monophyly and fast molecular evolution of the neo-
tropical assemblage. J Mol Evol 1999;48:703–11.
Farias IP, Orti G, Meyer A. Total evidence: molecules, morphology, and
the phylogenetics of cichlid shes. J Exp Zool 2000;288:76–92.
Feilich KL, López-Fernández H. When does form reect function?
Acknowledging and supporting ecomorphological assumptions. Integr
Comp Biol 2019;59:358–70. hps://doi.org/10.1093/icb/icz070
Fowler HW. A collection of shes obtained by Mr. William C. Morrow in
the Ucayali River Basin, Peru. Proc Acad Nat Sci Phil 1940;91:219–89.
Fricke R, Eschmeyer WN. Eschmeyer’s Catalog of Fishes: Guide to
Fish Collections; 2022. hp://researcharchive.calacademy.org/re-
search/ichthyology/catalog/collections.asp (1 April 2022, date last
accessed).
Friedman M, Keck B, Dornburg A, et al. Molecular and fossil evidence
place the origin of cichlid shes long aer Gondwanan riing. Proc
R Soc B: Biol Sci 2013, 280: 20131733. hp://dx.doi.org/10.1098/
rspb.2013.1733
Gante HF, Salzburger W. Evolution: cichlid models on the runaway to
speciation. Curr Biol 2012;22:R956–8.
Golobo PA. Extended implied weighting. Cladistics 2014;30:260–72.
hps://doi.org/10.1111/cla.12047
Golobo P, Farris JS. Methods for quick consensus estimation. Cladistics
2001;17:S26–34.
Golobo PA, Farris J, Ni xon K. T.N.T. Tree Analysis Using New Technology,
2003. hp://www.zmuc.dk/public/phylogeny/tnt
Golobo PA, Maoni CI, Quinteros AS. Continuous charac-
ters analysed as such. Cladistics 2006;22:589–601. hps://doi.
org/10.1111/j.1096-0031.2006.00122.x
Downloaded from https://academic.oup.com/zoolinnean/advance-article/doi/10.1093/zoolinnean/zlad021/7204395 by Universidade de Sao Paulo (USP) user on 27 July 2023
New generic classication of pike cichlids • 41
Golobo PA, Carpenter JM, Arias JS, et al. Weighting against homoplasy
improves phylogenetic analysis of morphological data sets. Cladistics
2008a;24:758–73.
Golobo PA, Farris JS, Nixon KC. TNT, a free program for phylogenetic
analysis. Cladistics 2008b;24:774–86.
Golobo PA, Torres A, Arias JS. Weighted parsimony outperforms other
methods of phylogenetic inference under models appropriate for
morphology. Cladistics 2018;34:407–37. hps://doi.org/10.1111/
cla.12205
Günther ACLG. Catalogue of the shes in the British Museum. Catalogue of
the Acanthopterygii, Pharyngognathi and Anacanthini in the collection of
the British Museum, Vol. 4. London: British Museum, 1862, 1–534.
Günther A. On new species of shes from the Essequibo. Ann Mag Nat
Hist 1863;12:441–3.
Haseman JD. An annotated catalog of the cichlid shes collected by the
expedition of the Carnegie Museum to central South America, 1907-
10. Ann Carnegie Mus 1911;7:329–73.
Heckel JJ. Johann Naerer’s neue Flusssche Brasilien’s nach den
Beobachtungen und Miheilungen des Entdeckers beschrieben
(Erste Abtheilung, Die Labroiden). Ann Wiener Mus Naturgesch
1840;2:325–471.
Huelsenbeck JP, Ronquist F. MRBAYES: inference of phylogeny.
Bioinformatics 2001;17:754–5.
Ilves KL, Torti D, López-Fernández H. Exon-based phylogenomics
strengthens the phylogeny of Neotropical cichlids and identies re-
maining conicting clades (Cichliformes: Cichlidae: Cichlinae).
Mol Phylogenet Evol 2018;118:232–43. hps://doi.org/10.1016/j.
ympev.2017.10.008
Kitching IJ, Forey PL, Humphries CJ, et al.. Cladistics: e Teory and
Practice of Parsimony Analysis, 2nd edn. New York, NY: Oxford
University Press, 1998.
Kocher TD. Adaptive evolution and explosive speciation: the cichlid sh
model. Nat Rev Genet 2004;5:288–98.
Kullander SO. Cichlid shes from the La Plata basin. Part I. Collections
from Paraguay in the Muséum d’Histoire naturelle de Genève. Swiss
Journal of Zoology 1981;88:675–92.
Kullander SO. Cichlid shes from the La Plata basin. Part III. e
Crenicichla lepidota species group. Swiss Journal of Zoology
1982;89:627–61.
Kullander SO. A Revision of the South American Cichlid Genus Cichlasoma
(Teleostei: Cichlidae). Stockholm: Naturhistoriska Riksmuseet, 1983.
Kullander SO. e Cichlid Fishes of the Amazon River Drainage of Peru.
Stockholm: Swedish Museum of Natural History, 1986.
Kullander SO. Teleocichla, a new genus of South American rheophilic
cichlid shes with six new species. Copeia 1988;1988:196–230.
Kullander SO. Biotoecus Eigenmann and Kennedy (Teleostei: Cichlidae):
description of a new species from the Orinoco Basin and revised gen-
eric diagnosis. J Nat Hist 1989;23:225–60.
Kullander SO. A new species of Crenicichla (Teleostei: Cichlidae) from
the Río Tapajós, Brazil, with comments on interrelationships of small
crenicichline cichlids. Ichthyol Explor Freshw 1990a;1:85–94.
Kullander SO. Crenicichla hemera (Teleostei: Cichlidae), a new cichlid
species from the Rio Aripuanã drainage, Mato Grosso, Brazil. Ichthyol
Explor Freshw 1990b;1:213–8.
Kullander SO. Mazarunia mazarunii (Teleostei: Cichlidae), a new genus
and species from Guyana, South America. Ichthyol Explor Freshw
1990c;1:3–14.
Kullander SO. Crenicichla phaiospilus and C. percna, two new species of
pike cichlid (Teleostei: Cichlidae) from the Rio Xingu, Brazil. Ichthyol
Explor Freshw 1991;2:351–60.
Kullander SO. Heroina isonycterina, a new genus and species of cichlid sh
from western Amazonia, with comments on cichlasomine system-
atics. Ichthyol Explor Freshw 1996;7:149–72.
Kullander SO. Crenicichla rosemariae, a new species of pike cichlid
(Teleostei, Cichlidae) from the upper Rio Xingu drainage, Brazil.
Ichthyol Explor Freshw 1997;7:279–87.
Kullander SO. A phylogeny and classication of the South American
Cichlidae. In: Malabarba L, Reis RE, Vari RP, Lucena ZM, Lucena
CAS, (eds), Phylogeny and Classication of Neotropical Fishes. Porto
Alegre: EDIPUCRS, 1998, 461–98.
Kullander SO. Family Cichlidae. In: Reis RE, Kullander SO, Ferraris JCJ,
(eds), Check list of the eshwater shes of South and Central America.
Porto Alegre: EDIPUCRS, 2003, 605–54.
Kullander SO, Lucena CAS. A review of the species of Crenicichla
(Teleostei, Cichlidae) from the Atlantic coastal rivers of southeastern
Brazil from Bahia to Rio Grande do Sul States, with descriptions of
three new species. Neotrop Ichthyol 2006;4:127–46.
Kullander SO, Lucena CAS. Crenicichla gillmorlisi, a new species of cichlid
sh (Teleostei: Cichlidae) from the Paraná river drainage in Paraguay.
Zootaxa 2013;3641:149–64.
Kullander SO, Nijssen H. e Cichlids of Surinam. Leiden: E.J. Brill, 1989.
Kullander SO, Varella HR. Wallace’s Pike Cichlid gets a name aer 160
years: a new species of Cichlid sh (Teleostei: Cichlidae) from the
Upper Rio Negro in Brazil. Copeia 2015;103:512–9.
Kullander SO, Norén M, Friðriksson GB, et al. Phylogenetic relationships
of species of Crenicichla (Teleostei: Cichlidae) from southern South
America based on the mitochondrial cytochrome b gene. J Zool Syst
Evol Res 2010;48:248–58.
Ladiges W. Bemerkungen zu einigen Neuimporten. Die Aquarien- und
Terrarienzeitschri 1958;11:203–4.
Landim MIPF. Relações logenéticas na família Cichlidae Bonaparte, 1840
(Teleostei: Perciformes). Unpublished Ph.D. Dissertation, Universidade
de São Paulo, 2006.
Lewis PO. A likelihood approach to estimating phylogeny from discrete
morphological character data. Syst Biol 2001;50:913–25.
Lippitsch E. Scale morphology and squamation paerns in cich-
lids (Teleostei, Perciformes): a comparative study. J Fish Biol
1990;37:265–91. hps://doi.org/10.1111/j.1095-8649.1990.
tb05858.x
Lippitsch E. A phyletic study on lacustrine haplochromine shes
(Perciformes, Cichlidae) of East Afr ica, based on scale and squamation
characters. J Fish Biol 1993;42:903–46.
Lippitsch E. Scale and squamation character polarity and phyletic assess-
ment in the family Cichlidae. J Fish Biol 1995;47:91–106. hps://doi.
org/10.1111/j.1095-8649.1995.tb01876.x
López-Fernández H, Honneycu RL, Stiassny MLJ, et al. Morphology,
molecules, and character congruence in the phylogeny of South
American geophagine cichlids (Perciformes, Labroidei). Zool Scr
2005a;34:627–51.
López-Fernández H, Honeycu RL, Winemiller KO. Molecular phyl-
ogeny and evidence for an adaptive radiation of geophagine cichlids
from South America (Perciformes: Labroidei). Mol Phylogenet Evol
2005b;34:227–44.
López-Fernández H, Winemiller KO, Honeycu RL. Multilocus phyl-
ogeny and rapid radiations in Neotropical cichlid shes (Perciformes:
Cichlidae: Cichlinae). Mol Phylogenet Evol 2010;55:1070–86.
López-Fernández H, Taphorn DC, Liverpool EA . Phy logenetic diagnosis
and expanded description of the genus Mazarunia Kullander, 1990
(Teleostei: Cichlidae) from the upper Mazaruni River, Guyana, with
description of two new species. Neotrop Ichthyol 2012;10:465–86.
hps://doi.org/10.1590/S1679-62252012000300001.
López-Fernández H, Arbour J, Winemiller KO, et al. Testing for an-
cient adaptive radiations in Neotropical cichlid shes. Evolution
2013;67:1321–37.
Lucena CAS. Two new species of the genus Crenicichla Heckel, 1840 from
the upper rio Uruguay drainage (Perciformes: Cichlidae). Neotrop
Ichthyol 2007;5:449–56.
Lucena CAS, Kullander SO. e Crenicichla (Teleostei: Cichlidae) spe-
cies of the Uruguai River drainage in Brazil. Ichthyol Explor Freshw
1992;3:97–160.
Luengo JA, Britiski HA. Una Crenicichla nueva del Rio Paraná, Brasil
(Osteichthyes, Cichlidae). Acta Biol 1974;8:553–64.
Lujan NK, Conway KW. Life in the fast lane: a review of
rheophily in freshwater fishes. In: Riesch R, Tobler M,
Plath M (eds), Extremophile Fishes: Ecology, Evolution, and
Physiology of Teleosts in Extreme Environments. Switzerland:
Springer International Publishing, 2015, 107–36. https://doi.
org/10.1007/978-3-319-13362-1_6
Manousaki T, Hull PM, Kusche H, et al. Parsing parallel evolution: eco-
logical divergence and dierential gene expression in the adaptive
Downloaded from https://academic.oup.com/zoolinnean/advance-article/doi/10.1093/zoolinnean/zlad021/7204395 by Universidade de Sao Paulo (USP) user on 27 July 2023
42 • Var e l l a et al.
radiations of thick-lipped Midas cichlid shes from Nicaragua. Mol
Ecol 2013;22(3):650–69. hps://doi.org/10.1111/mec.12034
McGee MD, Borstein SR, Meier JI, et al. e ecological and genomic
basis of explosive adaptive radiation. Nature 2020;586:75–9.hps://
doi.org/10.1038/s41586-020-2652-7
McMahan CD, Chakrabarty P, Sparks JS, et al. Temporal paerns of di-
versication across global cichlid biodiversity (Acanthomorpha:
Cichlidae). PLoS One 2013;8:e71162. hps://doi.org/10.1371/
journal.pone.0071162
Montaña CG, López-Fernández H, Taphorn DC. A new species of
Crenicichla (Perciformes: Cichlidae) from the Ventuari River,
Upper Orinoco River Basin, Amazonas State, Venezuela. Zootaxa
2008;1856:33–40.
Musilová Z, Říĉan O, Novák J. Phylogeny of the Neotropical cichlid sh
tribe Cichlasomatini (Teleostei: Cichlidae) based on morphological
and molecular data, with the description of a new genus. J Zool Syst
Evol Res 2009;47:234–47.
Near TJ, Dornburg A, Eytana RI, et al. Phylogeny and tempo of diversi-
cation in the superradiation of spiny-rayed shes. Proc Natl Acad Sci
USA 2013;110:12738–43.
Nguyen LT, Schmidt HA, von Haeseler A, et al. IQ-TREE: a fast and
eective stochastic algorithm for estimating maximum likeli-
hood phylogenies. Mol Biol Evol 2015;32:268–74. (hp://dx.doi.
org/10.1093/molbev/msu300)
Nixon KC, Carpenter JC. On outgroups. Cladistics 1993;9:413–26.
Norman JR. Descriptions of nine new freshwater shes from French
Guiana and Brazil. Ann Mag Nat Hist 1926;18:91–7.
Parins-Fukuchi C. Use of continuous traits can improve morphological
phylogenetics. Syst Biol 2018;67:328–39.
Pellegrin J. Description de Cichlidés nouveax de la collection du Muséum.
Bull Mus Natl Hist Nat 1903;9:120–5.
Pellegrin J. Contribution à l’étude anatomique, biologique eaxinomique
des poissons de la famille des Cichlidés. Mem Soc Zool Fr
1904;16:41–400.
Pellegrin J. Sur deux poissons du genre Crenicichla de la collection du
muséum de Paris. Bull Soc Zool Fr 1905;30:167–9.
Perugia R. Di alcuni pesci raccolti nell’ alto Paraguay dal Cav. Guido
Boggiani. Ann Mus Civ Stor Nat Giacomo Doria 1897;18:147–50.
Piálek L, Říĉan O, Cascioa J, et al. Crenicichla hu, a new species of
cichlid sh (Teleostei: Cichlidae) from the Paraná basin in Misiones,
Argentina. Zootaxa 2010;2537:33–46.
Piálek L, Říĉan O, Cascioa J, et al. Multilocus phylogeny of Crenicichla
(Teleostei: Cichlidae), with biogeography of the C. lacustris group:
species ocks as a model for sympatric speciation in rivers. Mol
Phylogenet Evol 2012;62:46–61.
Piálek L, Dragová K, Cascioa J, et al. Description of two new species
of Crenicichla (teleostei: cichlidae) from the lower Iguazú river with
a taxonomic reappraisal of C. iguassuensis, C. tesay and C. yaha. Hist
Nat 2015;3:5–27.
Piálek L, Burress E, Dragová K, et al. Phylogenomics of pike cichlids
(Cichlidae: Crenicichla) of the C. mandelburgeri species complex:
rapid ecological speciation in the Iguazú River and high endemism in
the Middle Paraná basin. Hydrobiologia 2019a;832:355–75. hps://
doi.org/10.1007/s10750-018-3733-6
Piálek L, Cascioa JR, Almirón AE, et al. A new pelagic predatory pike
cichlid (Teleostei: Cichlidae: Crenicichla) from the C. mandelburgeri
species complex with parallel and reticulate evolution. Hydrobiologia
2019b;832:377–95.
de Pinna MCC. Concepts and tests of homology in the cla-
distic paradigm. Cladistics 1991;7:367–94. hps://doi.
org/10.1111/j.1096-0031.1991.tb00045.x
Ploeg A. e cichlid genus Crenicichla from the Tocantins River, State
of Pará, Brazil, with descriptions of four new species. Beaufortia
1986;36:57–80.
Ploeg A. Review of the cichlid genus Crenicichla Heckel, 1840 from
Surinam, with descriptions of three new species (Pisces, Perciformes,
Cichlidae). Beaufortia 1987;5:73–98.
Ploeg A. Zwei neue Arten der Gaung Crenicichla Heckel, 1840 aus dem
Amazonasbecken, Brasilien (Pisces, Perciformes, Cichlidae). DATZ
Aqu ar. Te rr. 1989;42:163–7.
Ploeg A. Revision of the South American cichlid genus Crenicichla Heckel,
1840 with descriptions of een new species and considerations on spe-
cies groups, phylogeny and biogeography (Pisces, Perciformes, Cichlidae).
Unpublished Ph.D. Dissertation, Universiteit van Amsterdam, 1991.
Rabosky DL, Santini F, Eastman J, et al. Rates of speciation and morpho-
logical evolution are correlated across the largest vertebrate radiation.
Nat Commun 2013;4:1958.
Rambaut A, Suchard MA, Xie D, et al. Tracer v1.6, 2014. hp://beast.bio.
ed.ac.uk/Tracer
Ratnasingham S, Hebert PDN. bold: the barcode of life data system
(hp://www.barcodinglife.org). Mol Ecol Notes 2007;7:355–64.
hps://doi.org/10.1111/j.1471-8286.2007.01678.x
Regan CT. A revision of the shes of the South-American cichlid genera
Crenacara, Batrachops, and Crenicichla. Proc Zool Soc 1905;1905:152–68.
Regan CT. A synopsis of the cichlid shes of the genus Crenicichla. Ann
Mag Nat Hist 1913;8:498–540.
Říĉan O, Kullander SO. e Australoheros (Teleostei: Cichlidae) species
of the Uruguay and Paraná River drainages. Zootaxa 2008;1724:1–51.
Říĉan O, Zardoya R, Doadrio I. Phylogenetic relationships of Middle
American cichlids (Cichlidae, Heroini) based on combined evidence
from nuclear genes, mtDNA, and morphology. Mol Phylogenet Evol
2008;49:941–57.
Říčan O, Dragová K, Almirón A, et al. MtDNA species-level phylogeny
and delimitation support signicantly underestimated diversity and en-
demism in the largest Neotropical cichlid genus (Cichlidae: Crenicichla).
PeerJ 2021a;9:e12283. hps://doi.org/10.7717/peerj.12283
Říčan O, Říčanová Š, Almirón A, et al. Parallel evolution of a new sym-
patric species pair of Crenicichla (Teleostei: Cichlidae) from
Misiones, Argentina with a review of biogeography and mitonuclear
discordance in the C. mandelburgeri species complex. J Zoolog Syst
Evol Res 2021b;59:2134–49. hps://doi.org/10.1111/jzs.12566
Roa-Fuentes CA, Zeni J, Varella HR, et al. First record of Crenicichla
semifasciata (Heckel, 1840) (Teleostei: Cichlidae) in the upper rio
Paraná basin, Brazil. Check List 2015;11:1610.
Ronquist F, Huelsenbeck JP. MRBAYES 3: Bayesian phylogenetic infer-
ence under mixed models. Bioinformatics 2003;19:1572–4.
Sereno PC. Logical basis for morphological characters in
phylogenetics. Cladistics 2007;23:565–87. hps://doi.
org/10.1111/j.1096-0031.2007.00161.x
Smith WL, Chakrabarty P, Sparks JS. Phylogeny, taxonomy, and evolution
of Neotropical cichlids (Teleostei: Cichlidae: Cichlinae). Cladistics
2008;24:625–41.
Sparks JS, Smith WL. Phylogeny and biogeography of cichlid shes
Teleostei: Perciformes: Cichlidae). Cladistics 2004;20:501–17.
Steindachner F. Ichthyologische Beiträge (XVIII). Anz Kaiserl Akad Wiss,
Wien Math Naturwiss Kl 1915;52:346–9.
Steinhäuser JM. Crenicichla niederleinii (Holmberg, 1891) - eine seit über
125 Jahren "verschwundene" Art: ein Rätsel und seine mögliche Lösung.
DCG-Informationen (Deutsche Cichliden-Gesellscha) 2019;50:229–33.
Stiassny MLJ. e phyletic status of the family Cichlidae: a comparative
anatomical investigation. Neth J Zool 1981;31:3275–314.
Stiassny MJL. e relationships of the neotropical genus Cichla: a phy-
letic analysis including some functional considerations. J Zool
1982;197:427–53.
Stiassny MJL. Cichlid familial intrarelationships and the placement of the
neotropical genus Cichla. J Nat Hist 1987;21:1311–31.
Stiassny MJL. Phylogenetic intrarelationships of the family Cichlidae:
an overview. In: Keenleyside MHA (ed.), Cichlid Fishes. Behaviour,
Ecology and Evolution. New York, NY: Chapman and Hall, 1991,
1–35.
Stiassny MJL. Atavisms, phylogenetic character reversals, and the origin
of evolutionary novelties. Neth J Zool 1992;42:260–76.
Stiassny MLJ, Jensen JS. Labroid intrarelationships revisited: morpho-
logical complexity, key innovations, and the study of comparative di-
versity. Bull Mus Comp Zool 1987;151:269–319.
Strong EE, Lipscomb D. Character coding and inapplicable data. Cladistics
1999;15:363–71. hps://doi.org/10.1111/j.1096-0031.1999.tb00272.x
Taylor WR, Van Dyke GC. Revised procedures for staining and clearing
small shes and other vertebrates for bone and cartilage study. Cybium
1985;9:107–19.
Downloaded from https://academic.oup.com/zoolinnean/advance-article/doi/10.1093/zoolinnean/zlad021/7204395 by Universidade de Sao Paulo (USP) user on 27 July 2023
New generic classication of pike cichlids • 43
Varella HR. Revisão taxonômica das espécies de Crenicichla Heckel das
bacias dos rios Paraná e Paraguai (Teleostei: Cichlidae). Unpublished
M.Sc. esis, Universidade de São Paulo, 2011.
Varella HR, Ito PM. Crenicichla dandara, new species: the black jacundá
from the Rio Xingu (Teleostei: Cichlidae). Proc Acad Nat Sci Phil
2018;166:1–12. hps://doi.org/10.1635/053.166.0104
Varella HR, Moreira CR. Teleocichla wajapi, a new species of cichlid from
the rio Jari, Brazil, with comments on T. centrarchus Kullander, 1988
(Teleostei: Cichlidae). Zootaxa 2013;3641:177–87.
Varella HR, Kullander SO, Lima FCT. Crenicichla chicha, a new species
of pike cichlid (Teleostei: Cichlidae) from the rio Papagaio, upper rio
Tapajós basin, Mato Grosso, Brazil. Neotrop Ichthyol 2012;10:233–44.
Varella HR, Zuanon J, Kullander SO, et al. Teleocichla preta, a new spe-
cies of cichlid sh from the Rio Xingu basin in Brazil (Teleostei:
Cichlidae). J Fish Biol 2016;89:1551–69.
Varella HR, Loeb MV, Lima FCT, et al. Crenicichla ploegi, a new species of
pike-cichlid of the C. saxatilis group from the Rio Juruena and upper
Rio Paraguai basins in Brazil, with an updated diagnosis and biogeo-
graphical comments on the group (Teleostei: Cichlidae). Zootaxa
2018;4377:361–86. hps://doi.org/10.11646/zootaxa.4377.3.3
Wagner CE, Harmon LJ, Seehausen O. Ecological opportunity and sexual
selection together predict adaptive radiation. Nature 2012;487:366–9.
Webb JF. Ontogeny and phylogeny of the trunk lateral line system in
cichlid shes. J Zool 1990;221:405–18.
Young , Snoeks J, Seehausen O. Morphological diversity and the roles
of contingency, chance and determinism in african cichlid radiations.
PLoS One 2009;4:e4740.
Zuanon JAS. História natural da ictiofauna de corredeiras do Rio Xingu, na
região de Altamira, Pará. Unpublished Ph.D. Dissertation, Universidade
Estadual de Campinas, 1999; 199.
Downloaded from https://academic.oup.com/zoolinnean/advance-article/doi/10.1093/zoolinnean/zlad021/7204395 by Universidade de Sao Paulo (USP) user on 27 July 2023
