FACULTY OF SCIENCES Department of Biology Evolutionary Morphology of Vertebrates Early development of Corydoras aeneus (Siluriformes, Callichthyidae): a case study for understanding the evolutionary basis of loricarioid ontogenetic patterning Part 1 – Text Frank Huysentruyt Thesis submitted to obtain the degree of Doctor in Sciences (Biology) Proefschrift voorgedragen tot het bekomen van de graad Doctor in de Wetenschappen (Biologie) Academiejaar 2007-2008 Rector: Prof. Dr. Paul van Cauwenberge Decaan: Prof. Dr. Herwig Dejonghe Promotor: Prof. Dr. Dominique Adriaens Early development of Corydoras aeneus (Siluriformes, Callichthyidae): a case study for understanding the evolutionary basis of loricarioid ontogenetic patterning Part 1 - Text Frank Huysentruyt Thesis submitted to obtain the degree of Doctor in Sciences (Biology) Proefschrift voorgedragen tot het bekomen van de graad Doctor in de Wetenschappen (Biologie) Academiejaar 2007-2008 Rector: Prof. Dr. Paul van Cauwenberge Decaan: Prof. Dr. Herwig Dejonghe Promotor: Prof. Dr. Dominique Adriaens SAMENSTELLING LEES*- EN EXAMENCOMMISSIE Prof. Dr. Wim Vyverman Prof. Dr. Dominique Adriaens Prof. Dr. Peter Bossier Prof. Dr. Michel Chardon Dr. Tom Geerinckx Prof. Dr. Ann Huysseune Dr. Isaäc Isbrücker Prof. Dr. Giorgos Koumoundouros (Univeristeit (Universiteit (Universiteit (Universiteit (Universiteit (Universiteit (Universiteit (Universiteit Gent, voorzitter) Gent)* Gent) Luik)* Gent) Gent) Amsterdam, Nederland) Patras, Griekenland)* DANKWOORD Er zijn mensen die zeggen dat ik soms te veel praat. Ik denk dat dat terecht is. Veel valt daar niet aan te doen, het is wie ik ben. Maar af en toe heb ik ook echt iets te zeggen, zoals nu, en dus schrijf ik het beter op, zodat het blijft. In dit deel wil ik alle mensen bedanken die, elk op hun manier, aan dit werk hebben bijgedragen, hetzij in een actieve rol, hetzij in een meer passieve rol, de ene daarom niet minder belangrijk dan de andere. In eerste instantie wil ik natuurlijk Prof. Dr. D. Adriaens bedanken voor zowat alles. Vanzelfsprekend in de eerste plaats omdat hij in mij heeft geloofd en me de kans heeft gegeven om dit tot een goed einde te brengen. Daarnaast is er niemand die zelfs maar in de buurt komt van de mate waarin hij actief aan dit werk heeft bijgedragen. Van het lezen en verbeteren van teksten (die in het begin echt zeer slecht moeten geweest zijn) tot het aanbrengen van ideeën en het bediscussiëren van allerlei resultaten. Daarnaast wens ik hem zeker en vast ook te bedanken voor de kansen die hij me heeft gegeven om buitenlandse congressen bij te wonen waardoor ik niet alleen allerlei leuke locaties heb kunnen bezoeken, maar waar ik ook onschatbare ervaring in het presenteren en bediscussiëren van wetenschappelijke data mocht opdoen. Ik weet dat het zijn bewuste en niet altijd even evidente keuze is om zijn studenten deze kansen te bieden en ben hem daar, samen met mijn collega-studenten, zeer dankbaar voor. Ten tweede wil ik natuurlijk ook mijn collega’s bedanken. Ik kan oprecht zeggen dat ik altijd graag naar het labo ben gekomen en dat was vooral aan hen te danken. Ik denk niet dat er mensen tussenzitten met wie ik niet kon opschieten en sommige zijn hopelijk vrienden voor het leven geworden. In alfabetische volgorde: Barbara, Celine, Fatemeh, Heleen, Joachim, Marleen, Mimi, Natalie, Paul, Soheil, Stijn en Tom, bedankt voor een onvergetelijke tijd! In het bijzonder wil ik de hulp van Barbara, Joachim en Marleen bij al het praktische werk nog eens extra in de bloemetjes zetten, hun expertise vormt de hoeksteen van alle wetenschappelijk werk dat in ons labo wordt en zal worden verricht. Ook bij het analyseren van sommige data in dit werk heb ik de hulp gekregen zonder welke het simpelweg niet zou zijn gelukt: Bieke, bedankt voor het vele werk waar ik je te pas en te onpas mee heb lastig gevallen. In een voorlaatste paragraaf wil ik mijn familie bedanken. Mijn broer, schoonbroers en schoonzus omdat ze er gewoon altijd zijn als je ze nodig hebt. Mijn schoonouders daarnaast voor de vele steun en het niet gespeelde enthousiasme waarmee ze dit doctoraat ondersteund hebben. En natuurlijk vooral mijn ouders. Jullie hebben mij gedurende mijn ganse studies (en daarmee bedoel ik mijn volledige schoolcarrière) voortdurend gestimuleerd om het onderste uit de kan te halen. Iets wat ik aanvankelijk meer niet dan wel heb gedaan, maar het verstand komt nu eenmaal met de jaren. Het is me nu toch gelukt en het besef dat er twee mensen waren die op de momenten dat het wat slabbakte toch in mijn bleven geloven heeft daar een zeer grote rol in gespeeld. Bedankt. In laatste instantie wil ik mijn gezin bedanken. Eerst en vooral Jo(ke), om mij te leren studeren (en dat je dat in blokken van langer dan een half uur moet doen) en verder gewoon om er nu toch al tien jaar te zijn, elke dag. Ik hou zeer veel van je! Verder wil ik ook mijn twee zoontjes, Bas en Daan (en bij voorbaat mijn ongeboren kinderen, zodat er later geen ruzie is omdat de ene wel en de andere niet in papas doctoraat mocht), die absoluut in geen enkel opzicht ook maar iets met dit doctoraat te maken hebben, bedanken. Ik wil ze bedanken omdat ze, en dit kan contradictorisch klinken, ondanks de slapeloze nachten, ziektes, het vele werk en het constante weg-en-weer-gerij, het afwerken van dit doctoraat zo gemakkelijk hebben gemaakt. Het hebben van kinderen zorgt nu eenmaal voor de nodige zin voor relativering en rust die het schrijven van dit werk tot een plezier heeft gemaakt. Aan iedereen dus nogmaals bedankt ! Frank. A hen is only an egg’s way of making another egg. Samuel Butler TABLE OF CONTENTS TABLE OF CONTENTS 1. General Introduction 1.1. 2. 1 General context 1 Aims of the study 3 Outline of this thesis 6 Material and Methods 2.1. 2.2. 3. General context and aims Material 9 Choice of species 9 Material examined 11 Methods 15 In toto clearing and staining 15 Dissections 16 Serial sectioning 17 3D-reconstructions 18 Adult Morphology 3.1. 3.2. Adult Osteology 19 Abstract 19 Brief introduction 19 Brief material and methods 20 Results and discussion 20 Adult Myology 35 Abstract 35 Brief introduction 35 Brief material and methods 36 Results 36 Discussion 42 i TABLE OF CONTENTS 4. Ontogeny 4.1. 4.2. 4.3. 4.4. 4.5. 4.6. The Egg Abstract 49 Introduction 49 Brief material and methods 50 Results and discussion 51 Early Development and Growth 53 Abstract 53 Brief introduction 54 Brief material and methods 54 Results 56 Discussion 59 The Cranium 63 Abstract 63 Brief introduction 64 Brief material and methods 65 Results 65 Discussion of chondrocranial development 73 Discussion of osteocranial development 80 The Cranial Myology 87 Abstract 87 Brief introduction 88 Brief material and methods 89 Results 89 Discussion 91 The Postcranial Skeleton 97 Abstract 97 Brief introduction 97 Brief material and methods 98 Results 98 Discussion 101 Ontogeny Of Corydoras pygmaeus, a Miniaturized Congeneric Abstract ii 49 105 105 TABLE OF CONTENTS 5. Introduction 105 Brief material and methods 108 Results 108 Discussion 111 General Discussion 5.1. 5.2. The Ontogeny of Corydoras aeneus 117 The theory of saltatory ontogeny 117 The Ontogeny of Corydoras aeneus 118 Corydoras pygmaeus and Ancistrus cf. triradiatus 121 Thresholds in the development of Loricarioidea 122 The Evolution of Algae Scraping in Loricarioidea 125 Scaridae, another example of adaptive radiation an 6. algivory in teleosts 125 Three phases of evolution 126 Summary & Samenvatting 6.1. Summary 137 6.2. Samenvatting 143 7. References 149 8. Publication List 165 iii TABLE OF CONTENTS iv Chapter 1 Introduction 1 GENERAL CONTEXT AND AIMS 1.1 GENERAL CONTEXT AND AIMS General context In 1859 Darwin formulated his answer to the one question which had been troubling biologists over the centuries: 'Why are there so many different species?’. The answer he proposed explained how species evolved by random variation with an associated selection of those phenotypes with the highest fitness. He later added the theory of sexual selection to explain the presence of some unfavourable traits in extant species (Darwin, 1871). In addition to this, speciation in sexually reproducing organisms is explained by reproductive isolation between subpopulations, descended from a single interbreeding parent population. More recently, various authors have altered the focus point of this question toward: ‘Why does one group have so many species and another so few?’ or, if the number of extant species is held as a measure for evolutionary success, ‘why is one group more successful than another?’. Turner (2007) formulates this question in terms of rate and asks: ‘What makes some groups diversify faster and more extensively than others?’. This question becomes even more relevant if a great discrepancy in evolutionary success is noted between closely related groups. The process of evolution in which very low constraints apply, strongly increasing rate and quantity of the speciation process in a single group is often referred to as adaptive radiation. Various definitions of this term, however, exist. According to Simpson (1953), it is the simultaneous diversification of a lineage into numerous sublineages and specializations, while Seehausen (2004) states it to be the evolution of ecological and phenotypic diversity within a rapidly multiplying lineage. In conclusion, a general consensus could be to describe adaptive radiation as an evolutionary process in which both the rate of evolution as well as the diversity obtained increase significantly. In this context, the rate of evolution can be viewed as both the rate at which new characters get fixed in populations as well as the rate at which new species1 arise. 1 Mayr (1996, 2001) defined species as groups of interbreeding natural populations that are reproductively isolated from other such groups. 1 1 GENERAL CONTEXT AND AIMS In addition, all species of a radiation constitute a monophyletic group, and they often share some innovative trait or sets of traits (sometimes called key innovations (Liem, 1973)) (West-Eberhard, 2003). Currently, the interest in describing global biodiversity has again stimulated modern biologists to further contemplate on the evolutionary mechanisms generating this diversity. In this context, various studies have dealt with groups showing clear examples of such radiation processes, like Darwin’s finches, Caribbean anole lizards, reef-dwelling parrotfish, African cichlids,… and studies on these groups continue to date (e.g., Grant & Grant, 2002; Losos et al., 2003; Streelman et al., 2002; Albertson et al., 2003, 2005; Turner, 2007). A similar case of adaptive radiation is found in the evolution of catfishes of the family Loricariidae. This family is the most speciose of all Siluriformes (ca. 23% of all species) and the fifth most speciose of all teleosts (ca. 3% of all species), containing 716 species over 96 genera (Nelson, 2006; Ferraris, 2007) (fig. 1.1). Members of this family are also characterized by the peculiar trophic niche they occupy. All loricariids exhibit a large ventral suckermouth, which allows them to adhere to smooth surfaces and in addition, their trophic morphology has evolved into a scraping feeding tool (fig. 1.1). The use of this specialized feeding mechanism for (algae) scraping is, however, not obligatory since a wide variety of items associated with underwater substrates have been found in the diets of loricariids (Aranha et al., 1998; Delariva & Agostinho, 2001; Nelson, 2002). Loricariidae belong to the superfamily of the Loricarioidea (fig. 1.2; 1.3), of which Schaefer & Lauder (1986, 1996) have argued that the various families constructing it exhibit a trend of increasing morphological complexity in which various elements have been decoupled2, creating opportunities for the evolution of new structures and functions (see 5.2). In members of the more basal loricarioid family of the Callichthyidae a suckermouth and algae scraping apparatus is still lacking; but the mouth already has a more ventral position than most non-loricariid siluriform families, while in the family of the Astroblepidae such a suckermouth is present, although no algivory is known (fig. 1.3). 2 A decoupling of biological components refers to the unlinking of developmental pathways, tightly linked functions, aspects of stereotyped behaviour patterns, mechanical associations of bones, ligaments and muscles, or the reduction of a high genetic correlation (Lauder et al., 1989). The recognition of ‘new structures’ (sometimes termed novelties) varies among literature, but in many cases so-called new structures are actually the result of profound decoupling events (Geerinckx, 2006). 2 1 GENERAL CONTEXT AND AIMS The Loricarioidea in general and Loricariidae specifically have to date not been the subject of many studies on the mechanisms of radiation in this superfamily, perhaps due to their specioseness. As mentioned, Schaefer & Lauder (1986, 1996) have tackled the matter of the decoupling of various cranial elements as one of the mechanisms facilitating the evolution of such highly specialized morphologies, but most studies dealing with this group have been limited to a descriptive morphology of the adult, often in the context of phylogenetic analyses (e.g. Schaefer, 1990; Arratia & Huaquín, 1995; Reis, 1998; Aquino & Schaefer, 2002). A more thorough approach, focused on the evolutionary mechanisms behind the radiation in Loricarioidea could yield valuable information on the evolution of algae scraping in this group, since it occurred at both the superfamily and family level, with clear and recognizable steps of increase in morphological complexity. An algae scraping trophic morphology has also evolved in the family of the Mochokidae, a catfish group found on the African continent (fig. 1.4). In contrast to the Loricarioidea, however, algivory in this family has evolved at an infrafamiliar level. It has not led to a radiation of comparable diversity, since the family comprises only 193 species, which are confined to 10 genera, with algivory present in only 57 species belonging to 3 genera (Friel & Vigliotta, 2006; Ng & Bailey, 2006; Wright & Page, 2006; Ferraris, 2007; Vigliotta, 2007). Aims of the study This evolution of algae scraping trophic specializations in both loricarioids and mochokids has formed the basis of an ongoing FWO project: ‘Evolutionary adaptiveness for a highly specialized feeding niche: algae scraping in tropical catfishes’ led by the Evolutionary Morphology of Vertebrates lab at the Ghent University (UG). For this project, partnership was found in the Laboratory for Functional Morphology of the Antwerp University (UA) and the Ichthyology Department of the Africa Museum at Tervuren (RMCA). The project is interpreted as a study of morphology and function throughout the ontogeny of both plesiomorphic and specialized members of both loricarioids and mochokids. This 3 1 GENERAL CONTEXT AND AIMS way, the extent of convergence in both groups can be estimated, but, in addition, a comparison of developmental sequences of related taxa within a phylogenetic framework also serves the purpose of identifying possible heterochronic events (Mabee & Trendler, 1996). As it is risky to infer process from pattern (Hanken & Wake, 1993), adult morphology alone may be a poor guide to these processes. Therefore, an adequate assessment of differences in ontogenetic patterns between the different families will substantially improve my understanding of morphological divergence (Straus, 1985). In this context, the major aspects dealt with in this project are: - The ontogeny of a loricariid (Ancistrus cf. triradiatus) (T. Geerinckx – UG). - The functional morphology of several loricariid fishes (A. cf. triradiatus, Otocinclus vestitus, Farlowella acus) (T. Geerinckx, K. Nijs, D. Maes – UG). - A biomechanical study using X-rays and EMG of a loricariid (Pterygoplichthys lituratus) (T. Geerinckx – UG, A. Herrel – UA). - A kinematic analysis of suction feeding in a callichthyid (Corydoras splendens) (T. Lieben, A. Herrel – UA). - A comparative study of non-algae scraping (Synodontis sp.) and algae scraping (Atopochilus savorgnani) African Mochokidae (C. Ide, J. De Puysseleir, F. Huysentruyt, T. Geerinckx – UG). And, finally, the subject of this thesis: - The ontogeny of a basal, non-algae scraping, loricarioid (Corydoras aeneus). The analysis of such a basal, non-algae scraping representative allows to obtain a view on the plesiomorphic situation within loricarioid evolution. Groups that are situated at a more basal position in a phylogeny generally exhibit characters that have a more plesiomorphic character (i.e. more resembling the situation found in the common ancestor). These plesiomorphic characters are evidently also found throughout ontogeny and are not confined to the adult morphology. Therefore, the entire ontogeny of these groups can be considered as a plesiomorphic state, which makes the study of these ontogenetic trajectories highly informative. A 4 1 GENERAL CONTEXT AND AIMS comparison of these trajectories to those of more specialized groups would indicate the point(s) and manner(s) of divergence in both groups, yielding information on the mechanisms behind the evolution of a specialized morphology in the latter group. It is long known that ontogeny and evolution are intimately and reciprocally interrelated, since evolutionary changes in morphology require changes in development that produce relevant structures of interest, whose variation provides the material for evolution by natural selection (Klingenberg, 1998). The subject of this research was subsequently divided into five main objectives: • To provide a detailed description of the adult cranial and postcranial morphology, identifying the different structures and determining their homology in comparison to related taxa. This allows a more precise comparison of the ontogeny of those structures in my study species with a loricariid • species (see 3.1; 3.2). To perform an analysis of the early ontogeny of the species through an analysis of both external morphology as well as of growth. Such an analysis points out the timing and rate of the development and the subsequent shifts in developmental properties throughout ontogeny. These results also provide a framework onto which the results of the ontogeny of cranial • structures can be mapped (see 4.2; 4.7; 5.1). To present description a of detailed the (macroscopic ontogeny of the and microscopic) cranial structures: chondrocranium, osteocranium and musculature (see 4.3; • 4.4; 4.5) and of the postcranial skeleton (see 4.6). To present an overview of the ontogeny of Corydoras pygmaeus, a miniaturized congener of my subject species, in order to determine rigidity of the ontogenetic pattern in Callichthyidae and the effects of heterochronic events acting • on it (see 4.7). To compare all these results with similar results described for Ancistrus cf. triradiatus by Geerinckx (2006) (see 5.1). 5 1 GENERAL CONTEXT AND AIMS • Finally, to perform the implementation of all above results and comparisons in the context of the evolution of algae scraping and subsequent radiation in Loricariidae (see 5.2) As a general hypothesis in this research I expected the main differences found in the adult morphology of both a specialized versus a basal loricarioid to originate early in ontogeny. The evolution of algae scraping has generally been believed to have evolved via one of two possible evolutionary patterns. A first option consists of an evolutionary pattern in which the suckermouth attachment develops first, followed by the development of the scraping feeding mode, while in a second the order of events is the other way around (Adriaens, 2003; Geerinckx, 2006). From a phylogenetic perspective, the first option would appear to be the most plausible, since in the more basal family of the astroblepids such a suckermouth has evolved without the evolution of an algae scraping feeding apparatus. Therefore, I hypothesized that a suckermouth has evolved prior to an algae scraping feeding mode and it was expected that the earliest differences in the ontogenetic patterns of callichthyids and loricariids would be related to the formation of such a suckermouth. By studying the early ontogeny in both families, I further expected to find evidence that corroborates the decoupling hypothesis of Schaefer & Lauder (1986, 1996). Finally, the possibility existed that, next to the formation of a suckermouth and/or algae scraping mechanism, additional differences between both ontogenies were found, which had also facilitated the evolution of both elements in the more specialized groups. Outline of this thesis The body text of this thesis is divided into five main chapters. In the first chapter, a general introduction is provided, followed by the general aims of the study itself, in which the central questions of this research are formulated. 6 1 GENERAL CONTEXT AND AIMS The second chapter presents an overview of the material used in this study and of the general methods used in the study of these specimens. This chapter is then referred to in all chapters presenting my results under a ‘brief material and methods’ section. Specific methods used only in certain parts of my study are also commented on in these sections instead of in the general ‘Material and methods’ chapter. The third chapter is the first chapter in which results are presented. It deals with the adult morphology of the species studied and is divided into two parts: one discussing the skeletal morphology and another dealing with the species’ cranial and body musculature. In this dissertation I chose to start with a description of the adult morphology, since ontogeny is best studied with a thorough knowledge of its endpoint. In chapter four, the ontogenetic component of the thesis is dealt with. It is divided into seven parts, of which the first six deal with Corydoras aeneus itself. The first part describes the species’ egg morphology while the second part deals with early development and growth. In parts three to six results on the ontogeny of the species’ morphology are presented, dealing with respectively: the chondrocranium, osteocranium, cranial musculature and appendicular skeleton. Finally, in the last part, an overview of the ontogeny of a miniaturized congeneric (Corydoras pygmaeus) is presented. In the final chapter, these results are discussed. In the first part of this chapter, an overview of ontogeny of Corydoras aeneus is presented and important thresholds in the development of this species are recognized. This is also done for C. pygmaeus and Ancistrus cf. triradiatus and the occurrence of these thresholds in all three species are placed in an evolutionary perspective. In a second part of this chapter, all data gathered in this study, combined with additional data from literature are used to create an overview of the evolution of algae scraping in the superfamily of the Loricarioidea in general. Finally, in the last part of this chapter general conclusions of this thesis are formulated. 7 1 GENERAL CONTEXT AND AIMS 8 Chapter 2 Material and Methods 2 MATERIAL AND METHODS - MATERIAL 2.1 MATERIAL Choice of species The general aim of this study was to perform a detailed microscopical anatomical study of the development of the cranial elements in a basal loricarioid representative. To obtain this level of detail (using histological serial sections) through an entire ontogenetic series for both soft and hard tissues, time did not allow the study of multiple species and therefore, the research needed to be limited to a single species. Subsequently, the choice of the species became highly important, since it needed to reflect the plesiomorphic loricarioid condition as well as possible. This would allow the use of results as a basal model for further comparative research on members of this group. Therefore, as a first option, members of the group Nematogenyidae + Trichomycteridae, the sistergroup to all other loricarioids (fig. 1.2), were taken into consideration. However, none of the species in both families showed the possibility for generating a reliable ontogenetic series, mainly due to the scarcity of material and the absence of good breeding protocols. In addition, those species which were more readily available were the ones showing a high degree of specialization. This way, e.g. the nematogenyid family is monotypic, with its only extant species Nematogenys inermis appearing only in Chile where it is very rare (Ferraris, 2007; de Pinna, 2003). Trichomycterids, on the other hand, are very speciose holding 207 species in 41 genera (Ferraris, 2007), but very few data exist on the reproductive behaviour of the various species (needed to create breeding protocols). In addition, a high number of species which belong to this family are highly specialized (e.g. parasitic members in Vandellinae and Stegophilinae (de Pinna & Wosiacki, 2003)), which also affects breeding as well as their suitability as a plesiomorphic model. The Callichthyidae, on the other hand, the sister family to Scoloplacidae + Astroblepidae + Loricariidae (fig. 1.2), does include several species suited for breeding in captivity. The family comprises 194 species distributed over 8 genera (Ferraris, 2007): Callichthys, Dianema, Hoplosternum, Lepthoplosternum, Megalechis, Aspidoras, Scleromystax, and Corydoras. The latter genus is the 9 2 MATERIAL AND METHODS - MATERIAL most speciose genus of the family (152 species (Ferraris, 2007)), widespread in South America (Gosline, 1940; Nijssen, 1970; Kramer & Braun, 1983) and well known among aquarists for its many ornamental species (Burgess & Quinn, 1992). Given this and given the fact that several studies on members of the genus in general have already been published (Strauss, 1985; Howes & Teugels, 1989; Reis, 1998, Olivera et al., 1992, 1993), it was made my genus of choice. In addition, most of the studies done on the genus deal with the species Corydoras paleatus and C. aeneus (fig. 2.1). Therefore, the choice of one of both was logical. In comparison to the studies on C. paleatus, those published on C. aeneus showed a wider variety of subjects dealing with both morphology, physiology (e.g. Kramer & McClure, 1980, 1981; Kramer & Braun, 1983; Oliveira et al., 1992, 1993; Shiba et al., 1982; Sire & Huysseune, 1996; Huysseune & Sire, 1997) as well as reproductive biology (Kohda et al., 1995, 2002; Pruzsinsky & Ladich, 1998). In addition, C. aeneus is also very popular in the trade of freshwater ornamental fish and it is annually bred and shipped in large quantities all over the world (Tamaru et al., 1997). As a result, both specimens and breeding protocols (e.g. Fuller, 2001) for this species are easily available. Given this, C. aeneus (Gill, 1858) was chosen as the species of interest in this study. However, despite it being commercially bred, almost nothing is known about its ontogeny. Some attention has been paid to the early ontogeny of some aspects of the head in other callichthyids (Hoedeman, 1960c), but still a lot of relevant information is lacking. The same accounts for the adult morphology: a complete overview of the cranial and postcranial morphology is absent, despite of its relevance for ongoing phylogenetic research on Loricarioidea, to which these callichthyids belong (Reis, 1998; Britto & Castro, 2002). In studies by Reis (1998) and Britto (2003), morphological data provided a phylogenetic framework for both the family (Callichthyidae), and subfamily (Corydoradinae) and Shimabukuro-Dias et al. (2004) used molecular data to investigate the phylogenetic affinities within the family. The phylogeny of the genus itself, however, remains unresolved. (Schaefer, 1990; Reis, 1998; Aquino & Schaefer, 2002), as well as the generic relationships within the callichthyids (Reis, 1997, 1998) and very little information exists on the phylogeny of the highly diverse genus Corydoras. In this context, even the species of choice in this research still raises several questions in a taxonomical perspective. Where most Corydoras species have a 10 2 MATERIAL AND METHODS - MATERIAL distribution that is very often confined to a single basin, the distribution of Corydoras aeneus, covers almost the entire South American tropics. Therefore, it has been argued that the current definition of the species probably comprises several separate nominate species. At this point, several geographically distinct populations of Corydoras aeneus can be discerned by body coloration, a characteristic which disappears in preserved specimens (Isbrücker, pers. comm.). However, since no major additional morphological differences were identified to date and since the focus of this research is situated at a morphological level, this does not affect the results presented here. To allow a future identification of the species used, five additional specimens from the same strain were deposited at the Zoology Museum of the Ghent University (UGMD 175375-379). Material examined Breeding Specimens were commercially obtained from an aquarium shop (Poisson d’Or Belgium). Since origin of the adult specimens is unknown and since long-term breeding under artificial conditions could affect early development, a projection of the results of this study onto the whole species should be considered with caution. Still, since my main interest is to describe the general patterns of early development of the species (determined by size and age) the use of this brood stock seems justifiable, apart from the practical reasons (ability to obtain and breed them). A total of 35 specimens (25 were males), were put together in a 1.0 x 0.5 x 0.6 m tank. The tank had a 3 cm sand layer, a box filter, a heating apparatus and was heavily furnished with plant material (Microsorum, Echinodorus and Anubias-species). A temperature of 24-26°C, pH-level of 8-8.5 and hardness of 9-12° was maintained. The specimens were fed on a diet of commercial food flakes (TetraMin). In the tank, dry season conditions were imitated by lowering the water level to 10 cm and lowering light intensity for a period of 20 to 30 days. This was followed by an imitation of rain season conditions (daily adding of fresh, colder water and sufficient aeration) (Fuller, 2001). About a week after the start of this artificial rain season, several egg 11 2 MATERIAL AND METHODS - MATERIAL clutches (maximum 10) were found on the glass walls of the tank for several consecutive days. Hatching After hatching, which took place at about 3 days after fertilization, the hatchlings were moved to a smaller tank with similar water conditions (24-26°C and pH 8-8.5) and a photoperiod of 12 hours of darkness/light. From three days up to three weeks after hatching the larvae were fed on a diet of decapsulated Artemia-nauplii. After that, the diet was changed to crushed flakes and two weeks later to the adult diet of whole flakes (TetraMin). Collection of specimens Specimens from various nests were removed post-hatching at different ages, sedated and killed with an overdose of MS-222 (3-aminobenzoic acid ethyl ester, Sigma). Small specimens were photographed with the use of a ColorView digital camera, mounted onto an Olympus SZX9 binocular microscope. Afterwards, these specimens were measured to the nearest 0.1 mm using AnalySIS 5.0 software. Larger specimens were measured to the nearest 0.5 mm. In the smallest specimens in which the caudal fin had not yet developed and in which notochord flexion did not yet occur, notochord length (NL) was used as an alternate measurement to SL (figure 4.3B). The combined use of NL and SL in the analysis of growth is a method previously used in other ichthyological studies (see Kavanagh & Alford, 2003). After measurement, all specimens were given a collection number (AA, AB,…) (table 2.1) and preserved in a buffered 4% paraformaldehyde fixative (pH 7.0). Additional species In order to determine the importance and extent to which heterochronic events have played in callichthyid evolution, in chapter 4.7, a brief comparative study on a miniaturized congener was done. Species of choice for this was Corydoras pygmaeus (for a motivation on this choice see 4.7). Specimens were kept, bred and collected as described for C. aeneus. Specimens were also given a collection number (Cp 01, Cp 02,…) (table 2.2). 12 2 MATERIAL AND METHODS - MATERIAL TABLE 2.1 Specimens of Corydoras aeneus used in this study, excluding specimens that were only measured for use in chapter 4.2. (AB, Alcian Blue; AR, Alizarin Red; T, Toluidine Blue). No. Coll. No. NL/SL (mm) Age (Ph) Method Staining 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49. 50. 51. BC BG AU Y AM AV BN AA AX AN BP AW BR AD AY AJ AP AZ AG BT BW BA BK AR BB BZ BJ CA BQ BM CC BO CD BS CF CI CH BV BY CK CB CE CM CO CQ CS CL CG CJ CN CT 3.3 4.0 4.0 4.4 4.8 4.8 4.9 5.1 5.2 5.3 5.3 5.4 5.6 5.7 5.7 5.7 5.9 6.2 6.2 6.3 6.4 6.6 6.6 7.2 7.2 7.3 7.7 7.8 7.9 8.2 8.6 9.1 9.3 9.7 9.8 10.7 11.1 11.2 11.7 12.4 12.5 13.1 13.7 13.9 15.0 15.5 15.6 15.6 16.0 16.9 18.0 0h 17 h 1d 3d 9d 2d 3d 4d 4d 9d 4d 4d 5d 5d 5d 7d 11 d 6d 7d 6d 8d 7d 10 d 13 d 8d 10 d 9d 11 d 13 d 11 d 12 d 12 d 13 d 14 d 14 d 18 d 16 d 16 d 18 d 20 d 20 d 22 d 22 d 25 d 29 d 33 d 30 d 24 d 27 d 33 d 37 d Sections (2µm) Sections (2µm) Clearing Sections (2µm) Clearing Clearing Sections (2µm) Clearing Clearing Clearing Sections (2µm) Clearing Clearing Clearing Clearing Clearing Clearing Clearing Sections (2µm) Clearing Clearing Clearing Clearing Clearing Clearing Clearing Clearing Clearing Clearing Clearing Clearing Sections (2µm) Clearing Clearing Clearing + CT Clearing Clearing Clearing Clearing Clearing Clearing Clearing Clearing Clearing Sections (5µm) Clearing Clearing Clearing Clearing Clearing T T AB + AR T AR AB + AR T AR AB + AR AB + AR T AB + AR AR AB + AR AR AR AB + AR AR T AR AB + AR AB + AR AR AB + AR AR AB + AR AB + AR AB+ AR AB +AR AR AB + AR T AB + AR AR AR AR AB +AR AB + AR AR AB + AR AB + AR AR AR AR T AB + AR AB + AR AB + AR AB + AR AR 13 2 MATERIAL AND METHODS - MATERIAL 52. 53. 54. 55. 56. 57. 58. 59. 60. 61. 62. 63. 64. 65. 66. 67. 68. 69. 70. 71. 72. 73. 74. 75. 76. CR CP CZ CW DE DH DJ DF DW EI EL DV EJ EH EE EK DT DS DU EA EB DZ ED EC DY 18.5 19.0 19.5 20.5 22.0 25.0 26.0 30.0 35.0 35.0 36.0 36.0 37.0 37.0 39.0 40.0 41.0 43.0 43.0 43.0 44.0 45.0 47.0 48.0 50.0 41 d 37 d 50 d 50 d 70 d 80 d 90 d 70 d - Clearing Clearing Sections (5µm) Clearing Clearing Clearing Sections (5µm) Clearing Clearing Dissection Clearing Clearing Clearing Dissection Sections (5µm) Clearing Clearing Clearing Clearing Dissection Dissection Dissection Dissection Dissection Sections (5µm) AB +AR AB + AR T AB + AR AB + AR AB + AR T AB + AR AR AB + AR AR AB + AR T AB + AR AR AR AR T In addition to table 2.1, figure 2.2 presents a growth curve showing the different specimens of Corydoras aeneus used, depicting their specific treatment. TABLE 2.2 Specimens of Corydoras pygmaeus used in this study, excluding specimens that were only used for measurements in chapter 4.7. (AB, Alcian Blue; AR, Alizarin Red; T, Toluidine Blue). (*indicates the use of an adjusted clearing and staining procedure, see 2.2) No. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 14 Coll. No. Cp Cp Cp Cp Cp Cp Cp Cp Cp Cp Cp Cp Cp Cp Cp 01 02 08 09 04 10 16 06 17 22 47 21 31 25 56 NL/SL (mm) Age (Ph) Method 3.7 4.3 4.8 5.4 5.6 5.9 6.6 7.5 7.9 8.5 9.1 9.2 10.8 10.0 11.0 0h 1d 3d 4d 3d 7d 12d 14d 13d 22d 16d 22d 35d 29d - Clearing Clearing Clearing Clearing* Clearing Clearing* Clearing Clearing Clearing* Clearing* Clearing Clearing* Clearing* Clearing* Clearing Staining AB AB AB AB AB AB AB AB AB AB AB AB AB AB AB + + + + + + + + + + + + + + + AR AR AR AR AR AR AR AR AR AR AR AR AR AR AR 2 MATERIAL AND METHODS - METHODS 2.2 METHODS In toto clearing and staining As mentioned in table 2.1 and 2.2 and in figure 2.2, several specimens were subjected to a clearing and staining procedure. For this purpose, a slightly modified version of the protocol as described by Taylor & Van Dyke (1985) was used (table 2.3). Their method differs from Hanken & Wassersug’s (1981) method in, a.o., avoiding extensive damage from the acid cartilage stain on the bony structures by the alcohol series and the use of borax, and by using KOH instead of trypsin for the clearing. Small specimens often benefit from trypsin instead of KOH though, and therefore, in these cases, KOH was replaced by trypsin in the first clearing step of the protocol (table 2.3). Since the acetic acid used in the cartilage staining step could decalcify bone, this step was usually not exceeded over 12h and for control, this step was omitted in several specimens (table 2.1; 2.3). TABLE 2.3 Clearing and staining procedure adapted from Taylor & Van Dycke (1985). Step Dehydration Solution/Action Duration 50% alcohol 12h 75% alcohol 12h 96-100% alcohol 12h 96-100% alcohol 12h Neutralization alcian blue (8GX, Sigma): 0.009-0.03% in (40ml glacial acetic acid/60ml 96-100% alcohol) saturated borax solution (Na2B4O7.10H2O) Bleaching 3-10% H2O2 in 0.5% KOH 0.5h-... Clearing 1-4% KOH / 0.15% trypsin in 30% NaBO3 12h-... Bone staining alizarin red (Sigma): 0.5% KOH in 0.1% alizarin red, until colour switches from deep red to purple (stir) 24h Further clearing 0.5-4% KOH 12h... Preservation 25% glycerin + 75% 0.5% KOH 12h 50% glycerin + 50% 0.5% KOH 12h 75% glycerin + 25% 0.5% KOH 12h 100% glycerin storage Cartilage staining 8-24h 48h 15 2 MATERIAL AND METHODS - METHODS The method described in table 2.3, however, proved unreliable in the clearing and staining of Corydoras pygmaeus specimens, where several specimens were lost in the procedure due to overstaining with alcian blue. To counter this, the adapted method by Gavaia et al. (2000) was used as described in table 2.4 for the specimens marked with * in table 2.2. TABLE 2.4 Clearing and staining procedure adapted from Gavaia et al. (2000). Step Washing Solution/Action Duration TBST (50 mM Tris, pH 7.4; 150 mM NaCl; 0.1% Triton X-100) 0.25h aq. dest. 0.25h Neutralization alcian blue (8GX, Sigma): 0.009-0.03% in (40ml glacial acetic acid/60ml 96-100% alcohol, filtered) (pH ca. 1) 0.1% 1% KOH in alcohol Hydration 80% alcohol 2h 70% alcohol 2h 50% alcohol 2h 30% alcohol 2h aq. dest. 2h Clearing 0.5-5% KOH 4h-... Bone staining alizarin red (Sigma): 0.5% KOH in 0.1% alizarin red, until colour switches from deep red to purple (stir) 0.5h-2h Final clearing 0.5-4% KOH with a drop of H2O2 0.5h-4h Preservation 25% glycerin + 75% 0.5% KOH 2h 50% glycerin + 50% 0.5% KOH 2h 75% glycerin + 25% 0.5% KOH 2h 100% glycerin storage Cartilage staining Washing 0.2-0.5h 48h Dissections For the study of soft tissue, such as ligaments and muscles, dissections were performed on otherwise untreated fixed specimens (marked as ‘dissection’ in the method section in table 2.1). Visualization of muscle fiber arrangement was enhanced by the use of an iodine solution (Bock & Shear, 1972). For the study of cartilaginous and bony elements dissections were also performed on cleared and stained specimens. This allowed the study of separate elements and of those regions that were invisible in an overall dorsal, ventral and lateral view (like e.g. 16 2 MATERIAL AND METHODS - METHODS a detailed view of the gill arches, a ventral view of the neurocranium floor, a median view of the lower jaw,…). Study of specimens was done using an Olympus SZX7 stereomicroscope, equipped with a camera lucida for generating line drawings. Digital images of specimens (and eggs, see 4.1) were taken using an Olympus SZX9 stereomicroscope, equipped with a ColorView 8 digital camera driven by AnalySIS 5.0 software (Olympus). Serial sectioning In order to study the detailed anatomy of the earliest stages in ontogeny, serial sections were used. For this purpose, several specimens (table 2.1) were embedded in a Technovit 7100 medium and 2 or 5 µm thick sections were made using a Reichert-Jung Polycut microtome. After this, these sections were mounted onto glass slides, stained with toluidine blue and covered. Examination and digital imaging of these sections was done on a Reichert-Jung Polyvar light microscope, also equipped with a ColorView 8 camera driven by AnalySIS 5.0 software (Olympus). TABLE 2.5 Technovit 7100 embedding procedure. Step Solution/Action Duration Vacuum fixation 4% buffered formalin days to weeks Washing Tap water 8h Decalcification Decalc (Histolab) 36h Washing tap water 5h Dehydration 30% alcohol 12h 50% alcohol 12h 70% alcohol 12h 96% alcohol (two times alcohol renewal) 12h Technovit 7100 solution A min. 24h Technovit 7100 solution A renewal min. 48h add Technovit 7100 Harder II 12h place in deepfreeze 12h place at room temperature (check progress) approx. 2h place in oven 1h Embedding Polymerization 17 2 MATERIAL AND METHODS - METHODS 3D-reconstructions For the 3-dimensional visualization of the detailed anatomical structures studied through the use of serial sections, graphical reconstructions were generated. To do this, a series of digital images of the serial sections were made at fixed distance intervals between the slides photographed. This way, e.g., an image was taken every 4 slides of a 2 µm thick series, leaving a fixed 8 µm distance interval between all images. These intervals were then recalculated into a voxel coordinate system (3D-pixels), after which this value was assigned to images when uploading them into the Amira 3.1.1 (T.G.S.) software. There, images were aligned and the different anatomical structures of interest were manually traced. After this, surface models of each of these structures were generated, and digitally smoothed. 3D-meshes of these structures obtained this way were exported using the .dxf format, in which the object’s coordinates are included, allowing an automatic repositioning of all structures in relation to each other in other software packages used. In a next stage, these .dxf meshes were imported into the Rhinoceros (McNeel) software package. This software allowed rendering of different surfaces onto the 3D-meshes as well as the ability to obtain bitmap (.tiff) images taken from different angles of view. Afterwards, the reconstructions obtained were compared to cleared and stained specimens of similar size. Several comparisons made this way showed no substantial differences in the size and shape of comparable elements (hard tissue) between both methods. 18 Chapter 3 Adult Morphology 3 ADULT MORPHOLOGY – ADULT OSTEOLOGY 3.1. ADULT OSTEOLOGY Modified from: Huysentruyt F & Adriaens D Descriptive osteology of Corydoras aeneus (Gill, 1858) (Siluriformes: Callichthyidae) Cybium 2005, 29(3): 261-273. Abstract Both the cranial and postcranial osteology of Corydoras aeneus is described based on the study of five adult specimens. The results were compared to the condition described for other loricarioid fishes. Most results corresponded with the expected pattern based on studies dealing with parts of the callichthyid osteology, although some differences were observed. One of these differences was the presence of a suprapreopercular bone, a bone that was previously unmentioned in any callichthyid species. Further, several homologies were resolved and mainly confirmed existing hypotheses. This way, for example, based on the presence of several branches of the lateral line system, the compound dorsolateral bone of the otic region was identified as the posttemporopterotico-supracleithrum. Further, the presence of several otoliths in the compound bone forming the neurocranial floor in the occipital region, confirmed the fact that the bone comprized out of a fusion of both basi- and exoccipital bones. This study, however, failed to resolve some other questions regarding homologies (i.e. lacrymo-antorbital, suprapreopercle) pointing out the need for further ontogenetic research. In this light, this study provides the basis for this further comparative and ontogenetic research on callichthyids specifically and on loricarioids in general. Brief introduction The aim of this part of the study is to provide a full description of the osteology of the species, as a basis for further ontogenetic research on this species, as well as to contribute to future phylogenetic studies. 19 3 ADULT MORPHOLOGY – ADULT OSTEOLOGY Brief material and methods The specimens examined are presented in table 2.1. For clearing and staining procedures, see 2.2. For the nomenclature of the skeletal elements I followed Schaefer (1990) and Reis (1998). The homology of the autopalatine with the dorsal part of a premandibular arch follows Daget (1964) and Jarvik (1980). Results and discussion In Corydoras aeneus the neurocranium is pyriform, with a small ethmoid and orbital region, broadening at the temporal region into a large occipital region (fig 3.1B). Ethmoid region (figs. 3.1, 3.2) The mesethmoid in C. aeneus is narrow anteriorly and broadens posteriorly (see also Fink & Fink, 1996; Arratia, 2003a). It lacks cornua and broadens substantially toward its posterior margin. Although a general trend toward a reduction of the cornua is present in all catfishes (Lundberg, 1982; Schaefer, 1987), a total lack of these cornua is only present within the Callichthyidae, with the exception of the species of the former genus Brochis, in which extremely reduced cornua are still present (Schaefer, 1990; Reis, 1998). Hoedeman (1960d) suggested an initial formation of these cornua in Hoplosternum and Callichthys, but without further ossification, implying a reduction (presumably as a result of allometric growth). Further ontogenetic research will have to reveal whether this also holds for C. aeneus. On its posterior margin the mesethmoid contacts the frontals with a V-shaped suture dorsally and a W-shaped wedge with the prevomer ventrally. The posterior, V-shaped suture with the frontals in C. aeneus is also found in other callichthyids, in Nematogenys inermis and in the Scoloplacidae, but not in Loricariidae, Astroblepidae and Trichomycteridae, which reflects the plesiomorphic condition in the Loricarioidea (Schaefer, 1990). Laterally, the mesethmoid contacts the lateral ethmoids and the autopalatine ventroposteriorly. In addition, it is connected by a ligament to both the reduced premaxilla and the maxilla. The presence of such ligaments is a condition which 20 3 ADULT MORPHOLOGY – ADULT OSTEOLOGY all Callichthyidae share with the Scoloplacidae, Loricariidae and Astroblepidae (Schaefer, 1990). The well developed lateral ethmoids, together with the mesethmoid, the frontal, the autopalatine and the lacrymo-antorbital, surround the nasal cavity, which is different from the situation in the Callichthyinae, where a large depression in the lateral ethmoid forms the total nasal capsule (Reis, 1998). In this cavity the free, tube-like, nasal bone encloses the anterior part of the supraorbital canal. This canal directly enters the frontal bone at the posterior margin of the nasal (fig. 3.3) in contrast to the situation in the Callichthyinae where the supraorbital canal consequently first enters the lateral ethmoid (Reis, 1998). The nasal bone has the typical catfish, tube-like shape, although it only bears two pores, in contrast to the three pores found in most diplomystids and primitive loricarioids (Arratia & Huaquín, 1995). The toothless prevomer is dropshaped and forms an elongated V-shaped suture with the parasphenoid posteriorly. An independent prevomer is, within the Loricarioidea, present in all families except the Scoloplacidae (Schaefer, 1990; Arratia, 2003a). Orbital region (figs. 3.1, 3.2B) The first bone of the infraorbital series of C. aeneus, the lacrymo-antorbital, is a large, plate-like bone, which forms most of the ventral border of the orbita. This fusion of the first infraorbital bone, the lacrimal, with the antorbital bone is common among catfishes (Schaefer, 1990), but in C. aeneus some questions regarding the true nature and origin of the infraorbital bones remain. A first question is whether this bone really is the lacrymo-antorbital, merely an expansion of the lacrimal bone or whether both lacrimal and/or antorbital are totally absent (Schaefer, 1990; Reis, 1998; Arratia, 2003a). Subsequently, the homologies of the remaining infraorbital bones and of possible fusions within this series remain unclear. Regardless, the first two infraorbitals have become platelike in all Callichthyidae, a condition which they share with more primitive nonsiluriform teleosts, and some other Siluriformes (e.g. Clarias gariepinus (Adriaens et al., 1997)). In general, in Siluriphysi, the infraorbital series is reduced to tube-like bones bearing the infraorbital canal. Plate-like infraorbitals are therefore believed to be secondarily derived (Fink & Fink, 1996). The lacrymo-antorbital bears the first part of the infraorbital canal with two of its sensory pores. This canal further continues through the smaller second infraorbital bone (fig. 3.3). Furthermore, the anterior extension of this 21 3 ADULT MORPHOLOGY – ADULT OSTEOLOGY infraorbital canal into the first bone of the series occurs in Corydoras, Aspidoras and Scleromystax but is not present in other callichthyid species (Schaefer, 1990; Britto, 1998; Reis, 1998). On its dorso-posterior margin the second infraorbital broadens and connects both the sphenotic, of which the dermal part is in fact the last infraorbital bone (Gosline, 1975), and the posttemporopterotico-supracleithrum. The orbital skull roof is formed by the two large frontal bones, separated posteriorly by the anterior cranial fontanel. This fontanel is divided into two openings by the ossified epiphyseal bridge and is elongated posteriorly. As in the genera Hoplosternum, Megalechis, Lepthoplosternum, Dianema and members of the former genus Brochis, the anterior fontanel enters the parieto-supraoccipital bone in C. aeneus (Reis, 1998). The fontanel itself is minute, in contrast to that in other callichthyids, where a larger fontanel is present (Schaefer, 1990; Reis, 1998). In astroblepids, scoloplacids and loricariids, however, no open cranial fontanels are found. The frontals further contact the sphenotics laterally and the orbito- and pterosphenoid ventrally. The frontals, as in other teleosts, bear the supraorbital canal, but in C. aeneus an additional central pore is present (fig. 3.3). According to Reis (1998) this pore represents the parietal branch of that canal in other Siluriformes. Arratia & Huaquín (1995), however, report the absence of a parietal branch as a loricarioid synapomorphy, but, they, on the other hand, do report the presence of an epiphyseal branch in several loricarioids. Therefore, and based on the position and orientation of this pore I believe it to be homologous with this epiphyseal branch. The wall of the orbital region is formed by the anterior orbito- and the posterior pterosphenoid, which both ventrally contact the orbital floor at the level of the parasphenoid. The orbito-, and pterosphenoid in C. aeneus all posses the typical shape found in other Siluriformes (Schaefer, 1990; Reis, 1998; Arratia, 2003a). The orbitosphenoid is hour-glass-shaped in ventral view and holds a large foramen. The parasphenoid is fairly narrow anteriorly and bears two elongated, anterior processes (in between lies the prevomer). Posteriorly, it broadens widely at the level of the prootics, further ending in a narrow, sharp region. Anteriorly, the bone shows a strong, midline ridge. The posterior ‘wings’ of the parasphenoid suture with both prootic bones and the posterior tip connects to the occipital bone complex. Further, the bone is much broader than what is described for all Corydoradinae in Reis (1998). In between the orbitosphenoid, 22 pterosphenoid, parasphenoid and prootics, as in most 3 ADULT MORPHOLOGY – ADULT OSTEOLOGY siluriformes, the combined foramen for the fasciculus opticus and the trigeminofacial nerve is situated. A connection between the parasphenoid and pterosphenoid, thereby splitting the foramen in an orbital and trigemino-facial fenestra, is absent. Otic region (figs. 3.1, 3.2B) The floor of this region is formed by the posterior tip of the parasphenoid, flanked by the prootic bones. These square-shaped prootics further connect to the sphenotics and posttemporo-pterotico-supracleithrum laterodorsally and to the occipital complex posteriorly. The prootics enclose the utriculus with the lapillus and furthermore bear a large foramen. These foramina are not homologous to the auditory foramen (Schaefer, 1987), although their true purpose remains unknown. The lateral margins of the otic region are made up by the sphenotics that are also square and make additional contact with the parietosupraoccipital dorsally and the posttemporo-pterotico-supracleithrum posteriorly. In the sphenotics, both the supraorbital and infraorbital canal come together into the otic canal, which exits the sphenotic at its posterior margin and continues into the posttemporo-pterotico-supracleithrum (fig. 3.3). The latter bone has often been described as the fusion of both supracleithrum and pterotic with the absence of the posttemporal (Regan, 1911; Lundberg, 1975), while other authors described it as the fusion between a compound supracleithrum (supracleithrum, posttemporal and an ossified Baudelot’s ligament) and pterotic (Schaefer & Lauder, 1986; Schaefer, 1987, 1990; de Pinna, 1993; Reis, 1998). A compound nature of this bone is also supported by the presence of several segments and branches of the lateral line system (Schaefer & Aquino, 2000). At about one-third of the bone length, the preopercular canal branches off and from that point on the otic canal continues in the postotic canal, from which, at about two-thirds of the bone length, a first postotic branch, the pterotic branch, branches off. The canal then continues through the remaining one-third of the bone into the posttemporal branch and leaves the bone at its posterior margin (fig. 3.3). The presence of both the preopercular and pterotic branch indeed confirm the fact that the pterotic bone is included in this bony complex since both branches are generally inherent to the pterotic (Schaefer & Aquino, 2000). This situation, in which only one postotic branch (the pterotic branch) is present, occurs in all corydoradinae. In other callichthyid species two postotic branches 23 3 ADULT MORPHOLOGY – ADULT OSTEOLOGY are present (Schaefer & Aquino, 2000). Furthermore, the presence of a posttemporal branch confirms the presence of the posttemporal bone in the complex. Further, in C. aeneus, the posttemporo-pterotico-supracleithrum also bears a large articular cavity for the articulation of the pectoral girdle with the skull. Finally, it remains unclear whether the epiotic bones have disappeared during the development of C. aeneus or whether they have become incorporated within other bone complexes (Arratia, 2003a). Occipital region (figs. 3.1, 3.2B) The skull roof in this region is formed by one, large, compound bone, the parieto-supraoccipital. The fusion of the dermal parietal bones and perichondral supraoccipital bones during ontogeny is typical for Siluriphysi (Bamford, 1948; Lundberg, 1975; Fink & Fink, 1996). Here, the bone neither bears a posterior fontanel nor latero-sensory canals and has a large posterior process which contacts the nuchal plate and the first pair of laterodorsal bony scutes. The neurocranium floor in this region is formed by the compound occipital bone, a fusion between the basi- and exoccipitals. Within the Loricarioidea, a similar fusion only occurs in Scoloplacidae and Callichthyidae (Reis, 1998). Evidence for the presence of the basioccipital bone within the complex is found in the position of the bone contacting the parasphenoid anteriorly and articulating with the centre of the first vertebrae posteriorly, where it forms the posterior rim of the neurocranium (Rojo, 1991). Another indication is the fact that the bone encapsulates the asteriscus. Similar, the encapsulation of the sagitta confirms the presence of the exoccipitals within the bone complex. Further, this bone contacts the complex vertebral centre and its outgrowths on its posterior side and the posttemporo-pterotico-supracleithrum through the ossified Baudelot’s or transscapular ligament on its dorsolateral side (for a nomenclature on this structure, see Lundberg (1975), Fink & Fink (1981), Schaefer (1987) and Reis (1998)). Maxillary bones (figs. 3.1, 3.2) The highly reduced premaxillary bone is toothless in adult specimens of C. aeneus and bears a small dorsal process. The absence of teeth on the premaxilla is common to all callichthyids, although teeth are present in the early stages of C. aeneus (Machado-Allison & Garcia, 1986; Huysseune & Sire, 1997). The 24 3 ADULT MORPHOLOGY – ADULT OSTEOLOGY maxillary bone is also reduced to a small bone lacking dentition and supporting the maxillary barbel. In C. aeneus the bone is comma-shaped and bears a small process on its postero-lateral face. The bone articulates with the autopalatine through two articular facets which creates a hinge-joint. Both premaxillary and maxillary bones are ligamentously connected to the mesethmoid and next to that another ligament connects the maxillary bone to the palatine. A similar highly mobile and reduced premaxilla is present in all Callichthyidae. The fact that this increased mobility is caused by a ligamentous junction with the mesethmoid is a character shared with Astroblepidae and Loricariidae, but not with Scoloplacidae (Schaefer & Lauder, 1986). The shape and function of the maxillary bone (small, toothless and supporting the maxillary barbel) is the same as in all Siluriformes, with the exception of the Diplomystidae and †Hypsidoridae (Grande, 1987; Fink & Fink, 1996; Grande & de Pinna, 1998; Arratia, 2003a). Also, the presence of a pair of palatine condyles on the bone is common to all Siluriformes, except Astroblepidae and Helogenes-species (de Pinna, 1993). The bone’s posterolateral process serves as an insertion site for the retractor tentaculi muscle. Premandibular arch (figs. 3.1, 3.2) The autopalatine is rod-shaped and straight, with a flat lateral surface. It bears a small posterior process, which contacts the lateral ethmoid, a shape which is different from that in less advanced catfish families like Diplomystidae and †Hypsidoridae (Grande, 1987; Schaefer, 1990; Arratia, 1992). Anteriorly, the bone bears a large cartilaginous condyle for articulation with the maxillary bone, to which it is also ligamentously connected. The posterior process, on the other hand, bears no cartilaginous tip and is small compared to primitive catfishes (Arratia, 1992; Reis, 1998), but not as small as in Callichthys (Reis, 1998). The absence of this cartilaginous tip is variably present among catfishes, but common to all non-nematogenyid loricarioids (de Pinna, 1993). The process serves as the insertion site of the extensor tentaculi muscle (Fink & Fink, 1981; Reis, 1998). Mandibular arch (figs. 3.2A, 3.4) The metapterygoid bone is nearly triangular and has a narrow, elongated, anterior process that ends near the autopalatine. This process is single in Corydoras, as it is in Aspidoras, whereas, in Callichthys, it is bifurcated (Reis, 1998). The metapterygoid itself was first described as a fusion of ecto-, ento- 25 3 ADULT MORPHOLOGY – ADULT OSTEOLOGY and metapterygoid by Howes & Teugels (1989), although other authors reported the ecto- and entopterygoid to be absent (Regan, 1911; Arratia, 1990; Reis, 1998). In addition, the hypothesis by Howes & Teugels (1989) was, due to the lack of ontogenetic evidence, contradicted by Arratia (1992), who thus defined the bone as being the metapterygoid only. Next to this commonly used hypothesis, Diogo et al. (2001) also hypothesized the catfish hyomandibula as being a fusion of the hyomandibula and metapterygoid of other teleosts and the quadrate as being a fused quadrate + symplectic. Next to this, he also proposed a homology between the metapterygoid and the gymnotiform entopterygoid. In this case, I chose to follow the hypothesis as presented by Arratia (1990) since no evidence against it is presented here and it is commonly used in catfish literature. The metapterygoid further contacts the hyomandibula on its posterior margin through a serrated suture, as in all Corydoradinae (Reis, 1998). The metapterygoid is also joint synchondrally to the quadrate bone and ligamentously attached to the lateral ethmoid. The quadrate bone is a simple, small, triangular bone, the typical condition found in Diplomystidae, as well as in most Siluriformes (Arratia, 1992; Reis, 1998). The bone connects synchondrally to both the metapterygoid and the hyomandibula. On its postero-ventral margin it articulates with the articular bone complex. This complex is considered to consist of the fused angular, the articular and the retro-articular bone (Arratia, 2003a). This compound bone is small, not canal-bearing, and connected to the dentary bone complex. It articulates with the quadrate dorsally and is ligamentously connected with both the interopercle and posterior ceratohyal bone. The articular bone complex further bears a laminar coronoid process, which serves as an insertion site for parts of the adductor mandibulae muscle complex (Reis, 1998). The last bone of the mandibular arch is another compound bone called the dentary complex. The bone is a fusion of the mento-meckelium and the dental bone. It forms the main part of the lower jaw and is toothless in adult specimens of C. aeneus, a condition that is different from that in the early ontogenetic stages (Huysseune & Sire, 1997). It bears a small process anteromedially for insertion of the anterior intermandibular muscle. Further it medially encloses the Meckel’s cartilage. The fact that the Meckel’s cartilage is small and that no coronomeckelian bone is present are conditions the Callichthyidae share with Astroblepidae, Loricariidae, Trichomycteridae and several other nonloricarioid catfishes (de Pinna, 1993). As in the articular complex, the dentary 26 3 ADULT MORPHOLOGY – ADULT OSTEOLOGY complex does not bear a part of the preoperculo-mandibular branch of the lateral line system, a condition shared by all Loricarioidea, except N. inermis (Schaefer, 1990). Hyoid arch (figs. 3.4A, 3.5) The hyomandibula articulates with the neurocranium through the sphenotic and posttemporo-pterotico-supracleithrum. It also bears a large process on its dorsoposterior margin for the articulation with the opercle. The perichondral part of the hyomandibula is long and bears a bony plate on its ventro-anterior side, which contacts the metapterygoid and quadrate. On its medial side the bone articulates with the rest of the hyoid arch through the small interhyal bone. This interhyal articulates with the posterior ceratohyal, which, in turn, synchondrally contacts the anterior ceratohyal. The anterior ceratohyal has a twisted surface with a medial, bony outgrowth and articulates with three branchiostegal rays on its medial posterior margin and with the larger, fourth ray on its lateral posterior margin. The anterior part of the hyoid arch consists of both a ventral and a dorsal hypohyal, both square-shaped and articulating with the ventral, plate-like parurohyal. The presence of both dorsal and ventral hypohyals in C. aeneus and in most other Corydoradinae (Reis, 1998) is in contrast to other Loricarioidea. According to Arratia & Schultze (1990) most catfishes have two pairs of hypohyals, except for Trichomycteridae, Loricariidae and Callichthyidae, which is contradicted by my findings. The former study, however, was solely based on observations on Callichthys, where indeed only the ventral hypohyals are present (Arratia & Schultze, 1990; Reis, 1998). Schaefer (1987) confirms this and mentions a loss of the dorsal hypohyal only in Trichomycteridae and Loricariidae, but contrary to Arratia & Schultze (1990), he also mentions a similar loss in Astroblepidae. Branchial arches (figs. 3.5) In C. aeneus, the branchial basket bears the typical siluriform configuration in which five branchial arches are present. Only basibranchials II and III are present as distinct, ossified elements. The posterior copula remains cartilaginous. Ossified hypobranchials I and II are present, whereas separate hypobranchials III and IV show no ossification. The fifth hypobranchial is absent. The ceratobranchials of all five arches are well ossified, bearing cartilaginous tips 27 3 ADULT MORPHOLOGY – ADULT OSTEOLOGY (with exception of the posterior tip of the fifth one). All ceratobranchials support hemibranchs. The fifth ceratobranchial bears the lower pharyngeal tooth plate and is the only ossified bone in this arch. The first four epibranchials are very variable in shape, with the second and fourth bearing an uncinate process. All four are fully ossified and bear hemibranchia. Furthermore, the two first epibranchials are synchondrally connected to each other distally. They contact the ossified third infrapharyngobranchial bone through a fused cartilaginous first and second infrapharyngobranchial. The latter is synchondrally connected to the third epibranchial and to the fourth infrapharyngobranchial. This fourth infrapharyngobranchial is connected to the fourth epibranchial bone and supports the upper pharyngeal tooth plate. Opercular series (figs. 3.2A, 3.3) The opercular series consist of the opercular, interopercular, preopercular and suprapreopercular bones. This condition differs within different groups of catfishes and even loricarioids. In Loricariidae, for example, the interopercular bones have been lost entirely. The opercle itself is large, more or less triangular, and is connected to the interopercle on its ventro-anterior margin. It also bears a process for the articulation with the hyomandibula on its dorso-anterior margin. In Astroblepidae and Loricariidae this articulation shifts toward the dorsal side of the opercular bone (Schaefer, 1987, 1988). The interopercle is a small, triangular bone, which is ligamentously connected to the lower jaw at the level of the angulo-retroarticular bone. Dorso-anteriorly from the interopercle and anterior to the ventral part of the opercle, lies the preopercle. This bone, present in all loricarioid families, bears part of the preopercular canal with two of its pores, one centrally and one anteriorly, which are homologous to pores 4 and 5 in Diplomystes (Schaefer, 1988). The part of the preopercular canal running through the preopercle and suprapreopercle in C. aeneus, as in all Callichthyidae, does no longer connect to the part of the preopercular canal that is present in the posttemporo-pterotico-supracleithrum. In N. inermis and in several nonloricarioid catfishes, this canal continues into the mandible and is consequently referred to as the preoperculo-mandibular branch. In trichomycterids, on the other hand, the preopercular canal is extremely reduced and does not even enter the preopercle, but remains limited to an opening in the pterotic bone (Baskin, 1972; Schaefer, 1988). Finally, the presence of a suprapreopercular bone in C. 28 3 ADULT MORPHOLOGY – ADULT OSTEOLOGY aeneus, is a condition that has never been mentioned within the Callichthyidae but which was present in all specimens examined. Therefore, further ontogenetic research will focus on the development of this bone, attempting to reveal whether this bone is truly homologous to the suprapreopercular bone found in other fish groups. Weberian apparatus (fig. 3.1B) In C. aeneus the Weberian apparatus is part of a complex structure, comprising a fusion between several vertebrae. Normally, the complex vertebral centre of the Weberian apparatus is a fusion of the second to the fifth vertebral centre in all Siluriformes, except Diplomystes, where the fifth centre is excluded from the complex (Arratia, 1987; Fink & Fink, 1996). Additionaly, in Loricarioidea the first vertebral centre is also fused to the complex (Schaefer, 1990; Reis, 1998). Coburn & Grubach (1998), however, discovered, after ontogenetic research, that in Corydoras paleatus only three vertebrae are fused within the complex and that the first two vertebrae are missing. The gas bladder is divided into two chambers which are encapsulated in the expansions of the transversal processes of this complex centre. Laterally to the compound centre, two foramina are situated through which passes the duct that connects these two chambers. Gas bladder contact with the external medium occurs through an aperture in the posttemporo-pterotico-supracleithrum, covered by a hollow expansion bearing the latero-sensory canal. This condition is possibly homologous to the condition found in Astroblepidae and Loricariidae, where the aperture is completely covered by the posttemporo-pterotico-supracleithral bone (Reis, 1998). The connection between the gass bladder and inner ear is made up of one compound bone referred to as the compound tripus (Schaefer, 1990). The compound tripus found in C. aeneus was suggested to be a fusion between the tripus, intercalarium, scaphium and interossicular ligament, typically found in all Siluriformes, but with the loss of the claustrum in Callichthyidae (Schaefer, 1990; Reis, 1998). However, since Coburn & Grubach (1998) mention the loss of the first two vertebrae, their derivatives (claustrum, scaphium and intercalarium) are also missing and their results show the tripus to be a myoseptal tripus (formed in the paravertebral sac and the dorsal myoseptum of vertebra III). 29 3 ADULT MORPHOLOGY – ADULT OSTEOLOGY Vertebral column (figs. 3.1B, 3.6) In the specimens examined, the total number of vertebrae, including the first five incorporated in the Weberian apparatus, was 28. This number equals that found in several other Corydoras and callichthyid species, and is one more than the number found in Scoloplacidae and some Loricariidae (e.g. Otocinclus, Hypoptopoma) (Schaefer, 1990). Britto (2000) mentions the presence of 28-31 vertebrae in several Aspidoras-species, which corresponds to the 27-32 vertebrae described by Regan (1911) for the family of the Callichthyidae. The first articulating vertebra, the sixth vertebra, has two large parapophyses that articulate with the complex centre of the Weberian apparatus. These parapophyses further support a pair of large, hollow ribs, which contact the first ventrolateral bony scutes behind the pectoral girdle. Vertebrae 7-12 each carry a pair of small, thin ribs. The presence of a pair of such large, hollow ribs on the parapophysis of the sixth vertebral centre, followed by several pairs of small ribs is typical for all Callichthyidae (Regan, 1911; Hoedeman, 1960d; Reis, 1998; Britto, 2000). In contrast to Hoedeman (1960d) mentioning only four to five of these small ribs in Corydoras-species, six were found here. The number of caudal vertebrae is 14, of which, in the first three to four, the haemal spines are expanded and plate-like. These haemal spines are ventroposteriorly oriented, thus forming a protective, posterior wall for the abdominal cavity. Furthermore, the last preural vertebra is incorporated within the ural complex together with the last vertebra, the first ural vertebra (Lundberg & Baskin, 1969). Dorsal fin (fig. 3.6A) In C. aeneus the dorsal fin bears a first small fin ray, modified to serve a spinelocking mechanism, followed by a second, large one (Alexander, 1965). After this, seven branched dorsal fin rays are present, of which the last is split up to its base. The pterygiophores of these spines plus the first five fin rays are connected to the 8th to 11th vertebral neural spine. The first pterygiophore bears a large transverse process, which connects to the lateral body scutes. In callichthyids, this process is further ligamentously connected to the sixth rib, whereas in scoloplacids, astroblepids and loricariids, this ligament ossifies into a lateral bone (Schaefer, 1990). Preceding the first dorsal fin spine, a nuchal plate is present, which is connected to the seventh vertebra and contacts the parieto- supraoccipital at its anterior side. The condition where two fin spines with seven 30 3 ADULT MORPHOLOGY – ADULT OSTEOLOGY fin rays are present, as is the case here, fits the plesiomorphic nine fin rays found in Diplomystes (Alexander, 1965). Although most authors do not count the first modified fin spine, the number of remaining true fin rays fits my findings. It must, however, be stated that the low number of specimens studied here does not allow a detailed description of meristic characters and that such a description is also beyond the scope of this study. Nonetheless, seven fin rays, with the last ray branched up to the base, does corresponds with the original description of C. aeneus by Gill (1858) and of other Corydoras-species (Isbrücker & Nijssen, 1973, 1992a; Nijssen, 1970) and could indicate a low variability. The number of dorsal-fin rays is a character of great taxonomic value within the Corydoradinae and a number of 7-9 is determinative for Corydoras-species, whereas a number of 10 or more is determinative for species formerly belonging to the genus Brochis (and different from the plesiomorphic siluriform condition). The distinction between Aspidoras, on the one hand, and Corydoras, on the other hand, can also be done based on dorsal fin morphology. Here, a lack of contact between the nuchal plate and the posterior process of the parieto-supraoccipital is held as being typical for Aspidoras-species (Reis, 1996). Within the Loricarioidea, all families have the plesiomorphic siluriform number of branched dorsal fin rays, except for the Scoloplacidae, where a reduction of the number has occurred and only four are present (Reis, 1998). Anal fin The anal fin consists of a single unbranched and seven branched fin rays. The bases of the first four rays articulate with the haemal spines of vertebrae 20 to 22. The number of fin rays (n=7) found in C. aeneus corresponds with the number given in the original description by Gill (1858), although the presence of a possible eighth branched ray is mentioned. The presence of a single, unbranched, anal fin ray is a derived condition within the Callichthyidae, which only occurs in some Corydoras-species and all Lepthoplosternum-species. Among the Loricarioidea, this condition is also found in Scoloplacidae, Astroblepidae and Loricariidae, in contrast to Trichomycteridae and N. inermis, where two unbranched rays are present (Reis, 1998). 31 3 ADULT MORPHOLOGY – ADULT OSTEOLOGY Adipose fin The adipose fin consists of a single spine, derived from a small dorsal bony plate. The homology of this spine initially was unclear. The spine could be considered as a transformed bony scute or a true fin ray that is covered by these dermal plates. Hoedeman (1960d), however, mentions the presence of ‘two strong muscle bundles’ used for erection of the spine in Callichthyinae, which would mean that the spine in Corydoras, although not movable in the Corydoradinae, is homologous to a fin ray. Caudal skeleton (fig. 3.6B) As mentioned before, in many Siluriformes and Cypriniformes, the first preural vertebra is fused to the complex centre of the caudal skeleton (Lundberg & Baskin, 1969). The caudal skeleton in C. aeneus is of the pleurostyl type and consists of two bony plates. The dorsal plate is formed by a fusion of the urostyl and the dorsal hypurals III, IV and V, a fusion which, as well as the development of a plate-like lamina on the epural (the neural spine of the first preural centre), could be revealed by preliminary ontogenetic data (see 4.6). The ventral plate comprises the parhypural and hypurals I and II. Hypurals II and III are variably fused on their left and/or right side or on neither side. The fact that the dorsal hypurals are also fused to the compound centre is common to all loricarioids, except for N. inermis (Schaefer, 1990). The number of principal rays is 7/7, which is common among Corydoras species (Isbrücker & Nijssen, 1973, 1992a). Surprisingly, the number found here differs from that given in the original C. aeneus description by Gill (1858) (n=6/6). Further, both the neural and haemal spine of the preural vertebral centre II are heavily ossified and branched. This state is, to a lesser extent, also present in the preural vertebral centre III. Pectoral girdle (fig. 3.7) The pectoral girdle consists of the cleithrum, which articulates with the supracleithrum, part of the posttemporo-pterotico-supracleithrum and embedded in the skull. The cleithral bones are medially connected by a simple suture. The ventral part of the pectoral girdle consists of the scapulocoracoid. As in all Siluriformes the scapulacoracoid bone is a compound bone, comprising the scapula, the coracoid and the mesocoracoid. In Callichthyidae, the posterior process of this scapulacoracoid and of the cleithrum are sutured behind the 32 3 ADULT MORPHOLOGY – ADULT OSTEOLOGY articulation of the fin with the girdle, this way forming a bony shield around the entire base of that fin (Reis, 1998). The scapulocoracoid bones also connect ventrally, but in contrast to the cleithral bones, here a heavily interdigitating suture is present. Cleithrum and scapulacoracoid connect medially by means of a coracoid bridge (see also Diogo et al., 2001). The pectoral spine is pungent as in all Corydoradinae (this in contrast to the Callichthyinae), bears serrations on both its anterior and posterior face and a large articulation head, which also suits a spine locking mechanism (Hoedeman, 1960d; Alexander, 1965). Ten branched rays together with two proximal radials are present. Pelvic girdle (fig. 3.8) The pelvic girdle consists of two basipterygia, which bear both an internal and an external anterior process. The homology of both these processes was questioned by Shelden (1937), and both were referred to as ‘projections’. Since no obvious motivation was given to support this idea, I do not follow his views on this matter and consider them to be the internal and external process (see 4.6). The internal process is well developed and bears a small dorsal lamina. The presence of a dorsal lamina on the internal process is a typical callichthyid feature (Reis, 1998). Second, the external process also bears a lamina, which, in C. aeneus is connected to a scute of the lower, lateral series of bony scutes, by means of connective tissue (Reis, 1998). A third laminar process is present on the ischiac process, where it connects to the ventral tip of a scute of the lower, lateral series (Reis, 1998). This ischiac process is further divided into a dorsal and a ventral process, of which the dorsal part is bent laterally. These callichthyid features in the pelvic girdle are possibly related to reproductive strategies (Reis, 1998). Furthermore, the pelvic fin bears six branched rays. 33 3 ADULT MORPHOLOGY – ADULT OSTEOLOGY 34 3 ADULT MORPHOLOGY – ADULT MYOLOGY 3.2. ADULT MYOLOGY Modified from: Huysentruyt F, Geerinckx T & Adriaens D A descriptive myology of Corydoras aeneus (Gill, 1858) (Siluriformes: Callichthyidae), with a brief discussion on adductor mandibulae homologies Animal Biology 2007, 57(4): 433-452. Abstract Cranial and postcranial myology of adult Corydoras aeneus is described and results discussed in comparison to other ostariophysan, siluriform and/or loricarioid fishes. Further, a brief discussion on m. adductor mandibulae homologies is given providing arguments for the use of the terminology proposed by Diogo & Chardon (2000) in future studies dealing with the myology of Siluriformes. Doing this, I here identified an A1OST, A2, and A3’ section in C. aeneus and recognized the homology of the m. retractor tentaculi with the A3”. Next to this, the opercular system is discussed, focusing on similarities in this system in both a callichthyid (C. aeneus) and loricariid (Ancistrus cf. triradiatus) representative. In both these families, the m. dilatator operculi is enlarged and the direction of the operculo-hyomandibular articulation has shifted. In addition, in both lineages, the m. hyohyoidei abductor has also shifted its orientation, acting as an adductor. These similarities seem to corroborate the ‘decoupling hypothesis’ by Schaefer & Lauder (1986) in which a decoupling of elements in primitive members of a lineage leads to a higher morphological diversity within that lineage combined with the possible acquisition of new functions. Brief introduction In chapter 3.1, a descriptive study on the osteology of the species is presented, but a complete overview of the cranial and postcranial myology is still absent. This information, however, is quite relevant for ongoing phylogenetic research on Loricarioidea, to which these callichthyids belong (Reis, 1998; Britto & Castro, 2002; Britto, 2003). In his review, Britto (2003) already stated that the poor knowledge about the phylogenetic relationships among the genera of the 35 3 ADULT MORPHOLOGY – ADULT MYOLOGY Corydoradinae is largely due to the scarcity of data on internal anatomy. The review aims to counter this and provides 71 osteological characters, resolving a large part of this intergeneric phylogeny. However, no myological characters are included in his study. Reis (1998) also mentioned the importance and scarcity of anatomical data on the family level, but he also provided no myological data in his analysis. The lack of myological characters in these various phylogenetic analyses of this group is even more puzzling, considering the fact that Schaefer & Lauder in 1986 already found that, as a consequence of structural innovations in the jaw adductor complex in Loricarioidea, a lot of confusion exists about the homology of these different parts, and thus about their evolutionary history. In addition, Schaefer (1990) has also proposed the presence of a divided musculus adductor mandibulae as a synapomorphy for the Callichthyidae. Finally, even recent molecular research on the family of the Callichthyidae pointed out that “further analyses of additional morphological data may be very helpful in the understanding of the phylogeny of the Callichthyidae” (Shimabukuro-Dias et al., 2004). Consequently, the aim of this study is to provide a full description of the species’ adult myology, as a basis for further ontogenetic research on this species, as well as to contribute to future phylogenetic studies. Brief material and methods The specimens examined are presented in table 2.1. For clearing and staining procedures, procedures on serial sectioning and 3D-reconstructions see 2.2. For details on skeletal elements, I refer to chapter 3.1. Muscle terminology follows Winterbottom (1974), except where indicated otherwise. Results Jaw musculature (figs. 3.9, 3.11, 3.12) The m. adductor mandibulae complex in Corydoras aeneus consists of two different muscles, the A2 and A1OST-A3’. The A2 originates laterally on the 36 3 ADULT MORPHOLOGY – ADULT MYOLOGY hyomandibula, on a ridge near the suture with the metapterygoid and has its insertion on the ligamentum primordium. The A1OST-A3’ also originates on the hyomandibula, just below the A2, but the site of origin of this muscle expands further ventrally, covering the entire caudal margin of the quadrate bone. The muscle then runs forward, covering the entire suspensorial arch, splitting up at the level of the caudal margin of the lower jaw, with a lateral bundle (A1OST) inserting on the dorsolateral side of the lower jaw and a medial bundle (A3’) on the dorsomedial side. Both bundles attach to the lower jaw between the suspensorial joint caudally and the coronoid process rostrally. The m. retractor tentaculi in C. aeneus originates on both the hyomandibula and metapterygoid. The muscle then runs medially from both the A2 and A1OSTA3’ and, like the A2, also attaches to the lig. primordium. The m. intermandibularis anterior connects the medial sides of both dentary bones at their rostral margin. The m. intermandibularis posterior is subdivided into a pars dorsalis and a pars ventralis which merge rostrally and insert on the lower jaw. Both parts follow the same path and are difficult to be discerned through dissection. Serial sections, however, clearly show a subdivision into a dorsal and ventral part. These serial sections also revealed that the muscle was innervated by the inferior mandibular nerve branch of the trigeminal nerve (V) and no myocomma was present. This indicates that the muscle merely consists out of the m. intermandibularis posterior and that no m. protractor hyoidei part is present. The m. extensor tentaculi originates both from the lateral side of the orbitosphenoid and the caudal side of the lateral ethmoid bone. Rostrally, the muscle inserts as a single bundle on the caudal margin of the autopalatine, posterior to its articulation with the lateral ethmoid. Suspensorial musculature (figs. 3.9, 3.10) The m. levator arcus palatini originates from the skull at the level of the sphenotic bone and inserts on the laterodorsal surface of the hyomandibula. The m. adductor arcus palatini originates from the base of the skull along the lateral side of the parasphenoid bone and inserts on the medial side of both hyomandibula and metapterygoid. Posterior to the m. adductor arcus palatini, a m. adductor hyomandibulae is present, originating on the ventral side of the prootic and inserting on the medial surface of the hyomandibula. 37 3 ADULT MORPHOLOGY – ADULT MYOLOGY Opercular musculature (fig. 3.10) The m. dilatator operculi is a conical muscle originating from a cavity in the sphenotic and from the hyomandibular bone and running ventroposteriorly to insert, through a tendon, on the anterodorsal side of the opercle, above the articulation with the hyomandibula. The anterior fibers of this muscle are partially continuous with the m. levator arcus palatini. The m. adductor operculi originates entirely on the posttemporo-pteroticosupracleithrum. The muscle runs ventrally to insert on a ridge on the dorsomedial side of the opercle. The m. levator operculi is a broad muscle originating on the posttemporopterotico-supracleithrum and the prootic. This muscle runs ventrally to insert medially on the operculare on a large ridge that starts just below the rostral side of the insertion ridge of the dilatator operculi and that runs caudoventrally to the middle of the opercle. Hyoid musculature (fig. 3.12) As mentioned, no m. protractor hyoidei is present. The m. hyohyoideus inferior connects both hyoid arches and the left and right part of this muscle are medially fused through an aponeurosis. The m. hyohyoideus abductor in C. aeneus originates on the most medial branchiostegal rays, runs rostrally and fuses at the midline just behind the caudal margin of the hyohyoideus inferior. The aponeurosis through which these two halves are fused connects to the midline aponeurosis of the m. sternohyoideus (q.v.). However, a second bundle runs from the branchiostegal rays caudally to insert on the scapulocoracoid. Further, the mm. hyohyoidei adductores interconnect all branchiostegal rays, the opercle and the interopercle. Finally, the m. sternohyoideus is a large muscle, originating from the cleithrum and running rostrally, inserting on the parurohyal bone and further connecting to the dorsal and ventral hypohyal through a double ligamentous connection. This muscle is split up over its entire length into a right and left half, which medially connect through an aponeurosis. Gill arch musculature (figs. 3.13) Dorsally, in C. aeneus, gill arches I-IV bear a m. levator externus. These muscles originate grouped on the ventrolateral side of the prootics and insert on the respective epibranchial bones. On the third arch, however, the insertion site 38 3 ADULT MORPHOLOGY – ADULT MYOLOGY of the m. levator externus has also extended onto the cartilage connecting the epibranchial and infrapharyngobranchial III to the infrapharyngobranchial IV, with some fibers even inserting on the latter bone. Only arches I, II and IV bear a m. levator internus. This muscles is a single muscle in the first two arches, originating on the prootic, on the same site as the mm. levatores externus, and inserting on the cartilaginous infrapharyngobranchial I+II and epibranchial II. Since the infrapharyngobranchials of the first two arches are fused, I chose to consider this muscle as the combined m. levator internus of the first two arches. On the fourth arch the m. levator internus also inserts on the infrapharyngobranchial IV, with the upper pharyngeal toothplate attached, but the site of origin is on the ventromedial part of the posttemporo-pteroticosupracleithrum. A m. obliquus dorsalis is only present on arches III and IV. These muscles interconnect the epibranchials and infrapharyngobranchials of both arches. Only the third arch bears a m. transversus dorsalis which interconnects the left and right infrapharyngobranchial. Additionally, the m. obliquus posterior connects the ceratobranchial V to the epibranchial IV. It is a short muscle, with its sites of origin and insertion close to the lateral tips of both bones. No m. retractor dorsalis or m. levator posterior is present. In gill arches I, II, III and IV a m. adductor arcuum branchialium is present connecting the ceratobranchials to the epibranchials. Ventrally, all arches bear a m. rectus ventralis, with the exception of arch III. This muscle connects the ceratobranchial of the respective arch to the hypobranchial of the proceeding one, except for the first arch, where this muscle inserts on the dorsal hypohyal. Arch III bears the m. rectus communis, which has the same origin as the m. rectus ventralis in the other arches, but with its insertion on the dorsal hypohyal. A m. obliquus ventralis is only present in arches II and III, connecting the arches’ hypobranchials to their ceratobranchials. In arch II, these muscles also insert on the basibranchial, connecting the right and left m. obliquus ventralis at this point. A m. transversus ventralis was found in arches IV and V, interconnecting the left and right cartilaginous hypobranchial in the fourth arch and the left and right ceratobranchial in the fifth. Further, both the m. pharyngoclavicularis internus and externus connect the ventral part of the gill basket to the pectoral girdle. Posteriorly, both ceratobranchial V and epibranchial IV are confluent with the m. sphincter oesophagi. 39 3 ADULT MORPHOLOGY – ADULT MYOLOGY Eye musculature (fig. 3.14A) The extrinsic eye musculature in C. aeneus is similar to the pattern commonly found in teleosts. Two oblique muscles, the m. obliquus superior and m. obliquus inferior originate on the orbitosphenoid bone and insert on the most dorsal and most ventral part of the eyeball respectively. At the same sites respectively, both the m. rectus superior and m. rectus inferior insert, originating on the parasphenoid. The last two extrinsic eye muscles, the m. rectus externus and the m. rectus internus both originate in a large posterior myodome between the prootics and the parasphenoid at the level of the prootic bridge, inserting respectively on the most caudal and most rostral side of the eyeball. Pectoral musculature (fig. 3.14B, 3.15) The pectoral girdle is connected to the splanchnocranium through the m. sternohyoideus and to the posttemporo-pterotico-supracleithrum through the m. protractor pectoralis (fig. 3.9, 3.10). Ventrally, the main muscle mass for movement of the pectoral fin is made up of the m. abductor pectoralis superficialis pars ventralis. This muscle originates on the ventrolateral face of the cleithrum and scapulocoracoid and runs backwards where it attaches to the bases of all fin rays except the pectoral spine. Dorsal to this muscle lies the m. abductor pectoralis superficialis pars dorsalis, which also inserts on the fin rays, but originates from the scapulocoracoid. Lateral to these muscles lies the m. arrector pectoralis ventralis, which originates from the dorsolateral side of the cleithrum and inserts ventrally on the pectoral fin spine. Dorsal to this the m. arrector pectoralis dorsalis originates on the dorsomedial side of the cleithrum, runs through a foramen in the cleithrum and inserts on the anterodorsal margin of the pectoral fin spine. Also originating on the dorsomedial face of the pectoral girdle, but on the scapulacoracoid are the m. adductor pectoralis superficialis and the m. adductor pectoralis profundus, which also pass through a foramen to insert dorsally on the bases of all fin rays except the pectoral spine. Here, the lateral muscle fibers finally insert on the more medial fin rays and the medial fibers insert on the lateral rays, this way resulting in a ‘cross-over’ of muscle fibers. The pectoral spine is abducted by contraction of the m. abductor pectoralis profundus, which also originates on the medial side of the scapulacoracoid, but ventrally, and inserts on the head of this pectoral fin spine. 40 3 ADULT MORPHOLOGY – ADULT MYOLOGY Additionally, the pectoral girdle is connected to the pelvic girdle through the m. infracarinalis anterior. Pelvic musculature (fig. 3.16) Ventrally, the pelvic girdle is covered by three separate muscles. The most ventral muscle is the m. abductor pelvicus superficialis, which originates medially on the pelvic girdle and inserts on all but the most lateral fin ray. Underneath this muscle, but with a more anterior origin lies the m. abductor pelvicus profundus, which also inserts on all fin rays except the first. Also ventrally lie the m. arrector pelvicus ventralis and the m. arrector pelvicus dorsalis, which both originate on the ventrorostral margin of the anterior internal process and insert on the first pelvic fin ray. On the anteromedial margin of the anterior internal process of the pelvic girdle the m. adductor pelvicus superficialis originates and posterior to this up to the posterior margin of the girdle, the m. adductor pelvicus profundus has its origin. Both these muscles insert on all pelvic fin rays. The pelvic girdle further attaches to the anal fin through the m. infracarinalis medius and to the pectoral girdle through the m. infracarinalis anterior. Unpaired fin and body musculature (fig. 3.17, 3.18) The caudal fin musculature mainly consists of a m. flexor dorsalis and a m. flexor ventralis, which both originate on the third preural vertebral centre and insert on the bases of the principal fin rays. In addition, dorsal to this muscle lies the m. flexor dorsalis superior which originates on the neural spine of the third preural vertebra and inserts on the dorsal proximal fin rays and the upper principal fin ray. Also, ventrally a m. flexor ventralis inferior is present, which originates on the haemal spine of the third preural vertebra and inserts on the ventral proximal fin rays. Additionally, on the hypurals I, II and the parhypural, the m. hypochordalis longitudinalis originates, which ligamentously inserts on the upper principal fin ray. Further, the principal fin rays are interconnected through the mm. interradiales, which insert on the lepidotrichia shafts, distal to the insertion of the flexor muscles. Dorsally, the posterior margin of the dorsal fin is connected to the neural spine of the third preural vertebra by the m. supracarinalis posterior, which inserts on this spine through a tendon. Ventrally, the haemal spine of the third preural vertebra is connected to the anal fin by the m. infracarinalis posterior through a musculous insertion. 41 3 ADULT MORPHOLOGY – ADULT MYOLOGY Three types of muscles control the movement of the dorsal fin: the mm. erectores dorsales, the mm. inclinatores dorsales and the mm. depressores dorsales. The first two types of muscles, of which the mm. inclinatores dorsales lie more lateral and cover the mm. erectores dorsales, originate from the anteroventral part of the pterygiophore supporting the proceeding fin ray, while the m. depressor dorsalis of each fin ray has its origin on the ventral side of the pterygiophore of the corresponding fin ray. On the first dorsal fin spine only a m. erector dorsalis inserts which originates on the anterodorsal side of the seventh vertebra. A m. erector dorsalis also inserts on the second dorsal fin spine, which also originates on the seventh vertebra, albeit on the posterodorsal side. Additionally, this second spine bears a m. depressor dorsalis which originates on the anterodorsal side of the eighth vertebra and runs posterior to the transverse process to insert on the spine’s posterior margin. Posteriorly, the dorsal fin is connected to the third preural vertebra through the m. supracarinalis posterior. Anteriorly, the dorsal fin connects to supraoccipital bone through the m. supracarinalis anterior. Lateral to all other muscles of the anal fin lie the mm. inclinatores anales, which originate on the pterygiophores and insert laterally on the fin rays bases. Underneath these muscles lie the mm. erectores anales and the mm. depressores anales. Both these muscles originate on the pterygiophores and insert on the fin rays, the mm. erectores anales anteriorly and the mm. depressores anales posteriorly. Posteriorly, the m. infracarinalis posterior connects the anal fin to the third preural centre and anteriorly, the m. infracarinalis medius connects the anal fin to the pelvic girdle. Discussion Next to structural changes in the morphology of the m. hyohyoideus inferior, the presence of a cartilage plug between the lower jaw and the hyoid in Astroblepidae and Loricariidae, and the shift and neoformation of several ligamentous connections in the jaw region, the Loricarioidea differ structurally from the plesiomorphic siluriform situation found in Diplomystidae as described by Diogo & Chardon (2000) at the level of the jaw musculature. Because of this, 42 3 ADULT MORPHOLOGY – ADULT MYOLOGY the superfamily Loricarioidea has been considered “a clade showing a pattern of progressive increase in the mechanical complexity of structures associated with feeding”, with consequent homology ambiguities (Schaefer & Lauder, 1986, 1996). A first example of this is the nature of the dorsal division of the m. adductor mandibulae inserting onto the maxillary bone, the so-called A1 (Winterbottom, 1974; Gosline, 1986, 1993). Since such a muscle in ostariophysans is believed to have evolved independently from the A1 in eurypterygians (Fink & Fink, 1996), a true A1 part, as present in Acanthomorpha, is absent in Ostariophysi (Adriaens & Verraes, 1996; Diogo & Chardon, 2000; Wu & Shen, 2004). Alexander (1965), however, did recognize an A1 part in Callichthyidae and Loricariidae, an error later adapted by Howes (1983), Schaefer & Lauder (1986) and Schaefer (1997). Diogo & Chardon (2000) chose to name the different muscle bundles according to their position, giving the more lateral bundles inferior numbers. This way, they named this dorsolateral section in Ostariophysi A2 (medial to A0 and A1). Following the same logic, they suggested the term A1OST for the ventrolateral cheek muscle inserting on the dorsal face of the lower jaw (the term OST was added to avoid confusion with the acanthomorph A1). Wu & Shen (2004), on the other hand, avoided the use of the term A1 in their terminology, using only derivatives of the A2 and A3 terms. Still, I believe the terminology proposed by Diogo & Chardon (2000) to be more appropriate, given the thoroughness of their comparative work on siluriforms, a group dealt with in a rather limited way by Wu & Shen (2004). However, it must be noted that in the terminology of Diogo & Chardon (2000) the term A1 in A1OST is somewhat misleading due to the lack of any relation of this muscle with the maxillary, as already stated by Wu & Shen (2004); as well as that their A2 is not homologous to the A2 as described by Vetter (1878), and maybe the term A2OST would have been more appropriate. Given this, I identified the dorsolateral division of the m. adductor mandibulae in Corydoras aeneus as the A2 according to the definition by Diogo & Chardon (2000). In C. aeneus, in contrast to other non-loricarioid catfishes, this muscle has shifted its insertion onto the lig. primordium, through which it indirectly inserts on both upper and the lower jaw, altering the entire jaw mobility (Howes, 1983; Schaefer & Lauder, 1986). Next to this, in C. aeneus, the m. retractor tentaculi also inserts on the lig. primordium, a muscle I believe to be homologous to, or derived from the A3”, as already suggested by various authors 43 3 ADULT MORPHOLOGY – ADULT MYOLOGY (Lubosch, 1938; Alexander, 1965; Howes, 1983; Adriaens & Verraes 1996, 1997a; Diogo & Chardon, 2000; Diogo, 2005). In the ventral region of this cheek muscle complex, the m. adductor mandibulae is composed of two separate bundles: one attaching to the lower jaw laterally and a second one attaching to the lower jaw on the medial side. Given the ventrolateral position of the former bundle I identified it as the A1OST. The latter bundle was identified as the A3’, based on its medial position. Gosline (1989, 1993) and Adriaens & Verraes (1996) also confirm the presence of a compound lower jaw muscle in catfishes, and conclude that this muscle is the composed A2A3’, as suggested earlier by Takahashi (1925). Given the synonymy between the A2 as described by Takahashi (1925) to the A1OST as described by Diogo & Chardon (2000), this conclusion also fits my findings and I put this compound muscle in homology with the A1OST-A3’. This insertion of the A1OSTA3’ directly onto the medial surface of the dentary is considered a derived state in Siluriformes (Diogo, 2005). Diogo (2005), also claims that in Corydoras, the insertion of the A3’ is lateral to that of the A2 and A3”, and that the A3” inserts partially on the mandible and partially on the lig. primordium. When regarding the A3” as the m. retractor tentaculi, the latter statement is indeed correct, but my observations contradict the lateral position of the A3’ as it is situated on the medial side of the A2. Wu & Shen (2004) do not mention a further subdivision of the A3 and place the A3 in synonymy with the A3’ as described by Takahashi (1925). They, however, do not mention the m. retractor tentaculi in their survey, and since this is believed to be homologous to the A3”, retaining the name A3’ for the medial bundle in C. aeneus seems justifiable. Finally, in C. aeneus both an A0 and Aω are absent, a condition common in Siluriformes (Alexander, 1965; Diogo & Chardon, 2000). Next to this, the subdivision of the m. extensor tentaculi in C. aeneus is a character already mentioned by Diogo (2005) as a derived siluriform state but which was not reported by Schaefer & Lauder (1986, 1996). This subdivision of the m. extensor tentaculi fits the general trend of an increase in morphological complexity of the functional design through decoupling present throughout the loricarioid evolution (Schaefer & Lauder, 1986, 1996). The definition of decoupling in this case can be stated as the repetition of individual elements as redundant design components, followed by the specialization of one or more of these 44 elements as a mechanistic basis for the evolution of novel 3 ADULT MORPHOLOGY – ADULT MYOLOGY structure/function (Schaefer & Lauder, 1996) (see also 1.1). In these loricarioids, such a decoupling has mainly occurred at the level of the jaws. The upper jaws have been decoupled from the cranium, while the lower jaw has been decoupled from the opercular series, increasing mobility and independency of both upper and lower jaw (Schaefer & Lauder, 1996). Next to this, the lower jaws and hyoid musculature have become decoupled from their plesiomorphic bilaterally constrained midline attachments and a new redundant linkage was acquired (Schaefer & Lauder, 1996). All these factors, combined with the increase in myological complexity at the level of the m. adductor mandibulae has ultimately facilitated the evolution of a suckermouth in astroblepids and loricariids and the evolution of an algae scraping feeding apparatus in the latter family. Another example in which an evolutionary pattern of increased morphological complexity through decoupling is present throughout the loricarioid lineage involves the opercular system. This way, in the family of the Loricariidae, the opercle has been decoupled from the lower jaw and has lost his function in expiration (Geerinckx & Adriaens, 2006). Within this family, in the Ancistrinitribe, this has ultimately led to the formation of an erectile opercle, armoured with large denticles, acting as a defensive mechanism (Geerinckx & Adriaens, 2006). One of the most important innovations in this process has been the development of a hypertrophied m. dilatator operculi, which has ultimately led to the formation of several myodomes in the skull bones in Ancistrus cf. triradiatus, thus forming a ‘secondary skull roof’ (Geerinckx & Adriaens, 2006). In C. aeneus, the m. dilatator operculi is not hypertropied nor is it split into several different bundles. This corresponds to the plesiomorphic condition for Siluriformes as described by Diogo (2005), who only mentions a subdivided m. dilatator operculi in members of the Aspredinidae and Trichomycteridae. He, however, does not mention this for loricariids, though only two species of Hypoptopoma were included in the study. However, in C. aeneus, the m. dilatator operculi does extend rostrally, originating in a cavity in the sphenotic bone, a situation which is not mentioned in the study by Diogo (2005). In his study, the origin of the m. dilatator operculi on the dorsal surface on the neurocranium is mentioned as a derived character state in Trichomycteridae and members of the genus Plotosus but the presence of a sphenotic cavity was not detected in any of the studied siluriforms. As in A. cf triradiatus, the cavity in C. aeneus splits up the sphenotic bone in a deeper and outer layer, although both 45 3 ADULT MORPHOLOGY – ADULT MYOLOGY situations differ structurally. In C. aeneus the deeper layer of this cavity is formed by perichondral part of this bone (the autosphenotic), where the outer layer is formed by the dermosphenotic part (as shown from preliminary ontogenetic data). In the skull bones where a myodome is formed in A. cf triradiatus, however, the myodome is positioned in between layers of mixed origin, since canals of the lateral line system are present in the deeper layers (Geerinckx & Adriaens, 2006). Therefore, both cavities in A. cf. triradiatus and C. aeneus can not be considered homologous and the term myodome was not applied to the situation in C. aeneus. In addition, in its most narrow sense the term ‘myodome’ is restricted to those cavities housing the external eye muscles in teleosts (Rojo, 1991). In a broader sense, as applied by Geerinckx & Adriaens (2006) the term applies to a cavity formed in a bone housing a muscle, which is clearly the case in the A. cf. triradiatus skull, but not in the C. aeneus sphenotic, where the dermosphenotic merely forms an outgrowth covering the m. dilatator operculi. Nonetheless, the rostral expansion of the m. dilatator operculi in C. aeneus, combined with the more oblique direction of the opercle-hyomandibular articulation, implies an increase in efficiency in the dilatation of the opercle. Next to this, the anterior bundle of the m. hyohyoidei abductor has shifted its orientation to the transverse plane, acting as an adductor of the branchiostegal membrane. A similar state is found in A. cf triradiatus, in which the m. hyohyoidei abductor has also shifted its direction medially, acting as an adductor of the branchiostegal membrane, whereas fibers of the m. hyohyoideus inferior assist in the abduction (Geerinckx & Adriaens, in press). This way, the closing of the branchiostegal membrane is at least partially decoupled from the opercle in both species, a factor which could have facilitated the alteration of the opercle into a defensive mechanism in A. cf. triradiatus. This hypothesis would again fit the general hypothesis of increased morphological complexity through decoupling, which shows the importance of decoupling combined with shifts in function as a common evolutionary pathway for accomplishing innovative structural design. It is my belief that further investigation of not only the adult morphologies, but also extensive ontogenetic and functional comparative research of several loricarioid lineages will elucidate the full impact of such an evolutionary pathway and quantify their role in the evolution of these extremely successful and diverse taxa. 46 3 ADULT MORPHOLOGY – ADULT MYOLOGY 47 3 ADULT MORPHOLOGY – ADULT MYOLOGY 48 Chapter 4 Ontogeny 4 ONTOGENY - EGG 4.1 THE EGG Modified from: Huysentruyt F & Adriaens D Adhesive structures in the eggs of Corydoras aeneus (Gill, 1858; Callichthyidae) Journal of Fish Biology 2005, 66: 871-876. Abstract The surface structure of the eggs of the catfish Corydoras aeneus was examined and showed to be a unique pattern among teleosts. The surface was covered with small protuberances, which resemble attaching-filaments of teleost eggs. Eggs were, however, found to be very adhesive and since the species is known to inhabit turbid waters, this rare egg attachment mode could well be related to these environmental settings. Introduction Modifications in the egg morphology of various teleosts often reflect the ecological challenges a species is faced with during its embryonic life stages. In this context, for instance, the thickness of the zona radiata interna, which has a protective function, is directly related to the exposure to mechanical strains (Riehl, 1996). Next to this, the zona radiata externa has been known to mediate egg adhesion in several teleosts and various modes of egg attachment are known (Laale, 1980; Patzner & Glechner, 1996; Riehl & Patzner, 1998; Rizzo et al., 2002). Again, the presence and manner of egg attachment could reflect environmental constraints placed upon the eggs, or the species in general, during these developmental stages. Patzner & Glechner (1996) found fishes from different environments, which exhibited the same variety in attachment structures. Rizzo et al. (2002), on the contrary, related the absence or presence of egg adhesion in different species to their migratory behavior and Morin & Able (1983) related the nature of adhesive structures found in the eggs of Fundulus heteroclitus to a variance in egg deposit sites. In addition Rizzo et al. (2002) 49 4 ONTOGENY - EGG point out that adhesive eggs are usually larger and laid in smaller numbers, as well as they relate egg adhesiveness to both the sedentary nature of species and possible parental care. In any case, the presence of adhesive eggs certainly indicates a behavioral strategy in which eggs are confined to a single locus until hatching, a strategy which could relate to environmental settings, making it useful to take environmental factors into account when discussing egg morphology. Therefore, in the armored catfish Corydoras aeneus, known to inhabit fast current fresh water systems in Southern America (Gosline, 1940; Nijssen, 1970; Kramer & Braun, 1983; Burgess & Quinn, 1992; Kohda et al., 1995; Fuller, 2001), adaptations in egg morphology in order to cope with a similar turbulent environment can be expected. More so, since Kohda et al. (1995) found adaptations in the species’ insemination strategy, presumably to ensure a high insemination rate even in a turbulent habitat. This unique mode of insemination in C. aeneus consists of the female swallowing sperm and quickly transporting it through the intestine, emitting it at the anal opening into a previously emitted ventral egg pouch (Kohda et al., 1995). However, since no motivation was presented by these authors describing how the sperm is protected from the intestinal environment and since this method has not been observed in any other species, this hypothesized insemination mode remains doubtfull and needs further study. For this study, eggs of C. aeneus were examined in order to obtain information on their morphology, this way possibly gathering some information on the environmental demands met by the species during its primary developmental stages. Brief materials and methods Breeding in Corydoras aeneus was induced as described under 2.1. The fertilized eggs were photographed and measured as described under 2.2. After this, eggs were fixed in 4% glutaraldehyde in a 0,2 M cacodylate buffer (pH 7,4). After washing, the eggs were dehydrated in a graded ethanol series and critical point dried under liquid CO2. The eggs were further mounted and coated with 50 4 ONTOGENY - EGG gold using a Balzers SCD040 sputtercoater. For examination a Jeol SM840 scanning electron microscope was used at 15 kV. Results and discussion The eggs (n = 21) examined have a diameter of 1.47 ± 0.20 mm (± SD) and were always laid against either a leaf or the aquarium walls. This size matches the egg size of 1.5 mm described by Fuller (2001) for the species. A large yolk sac is present and, like in most teleosts (Riehl & Patzner, 1998), eggs are spherical in shape (fig. 4.1). The egg-surface is covered with little projections, with the exception of a small circular region surrounding the micropyle (fig. 4.1B; 4.2E). These projections are c. 10 µm long, c. 5 µm wide and continuous with the outer layer of the zona radiata, i.e. the zona radiata externa (fig. 4.2A), which is in total c. 13 µm thick. Both the size and position of these projections would suggest them to be villi-like protuberances as according to the definition by Riehl & Patzner (1998). However, when comparing them to the villi described in Leuciscus leuciscus, Alburnoides bipunctatus and Rutilus rutilus by Patzner & Glechner (1996), the shape of the villi seems to differ from all these species. The protuberances found here were regularly hexagonally shaped (fig. 4.2C), whilst the villi found in the former species were irregularly shaped. In addition, after spawning, these protuberances can lengthen considerably and act as adhesive structures in contact to both the substrate or other eggs (fig. 4.2B). After lengthening the projections more resemble attaching-filaments than villi-like protuberances. Attaching-filaments, however, are only fixed at the vegetal or animal pole or arranged in a disc or annular bulge (Riehl, 1996; Riehl & Patzner, 1998) excluding these projections from that category. In addition, this part of the zona radiata is perforated in Corydoras aeneus with several small pores that lie in a clear hexagonal pattern in between the projections (fig. 4.2C). Similar hexagonal structures are suggestive of the honey-comb like pattern characteristic of fishes of the family Percidae (Riehl & Bless, 1995; Riehl & Patzner, 1998). This way, the arrangement and attachment mode found in the eggs of C. aeneus was not described in other teleosts. Further, under the zona radiata externa lies the zona radiata interna, which is c. 7 µm thick and exist 51 4 ONTOGENY - EGG entirely out of pillar-shaped structures separated by pore-canals (fig. 4.2A). The thickness fits the teleost average, where a zona radiata of 5-15 µm is commonly found (Riehl, 1999), which wrongfully suggests that the eggs of C. aeneus do not have to cope with more than average mechanical stress. The deepest zone of the zona radiata is the zona radiata subinternus, which forms the boundary with the oocytoplasma and is ± 0.2 µm thick. This part of the zona radiata is also highly perforated but pores are more randomly organized throughout this inner layer. The micropylar apparatus in C. aeneus exists of a flat, ellipse-shaped, micropylar pit which is approximately 50 µm long and 20 µm wide (fig. 4.2D). In this pit lies the micropylar canal, which is c. 13 µm wide (fig. 2D; F). Although the latter measurement is somewhat unreliable due to the unclear borders of the canal, this condition fits the type 2 micropylar configuration as described by Riehl & Götting (1974) and Riehl (1999). The micropyle itself lies at the end of the pit and is, after insemination, closed off by a micropylar plug to avoid polyspermy (Riehl, 1996). Conclusively, it can be stated that the eggs in C. aeneus exhibit a unique surface pattern with small villi-like protuberances which resemble attachingfilaments of teleost eggs. The presence of these structures could be related to the turbid habitat in which this species lives, but further studies on eggs of phylogenetically related and of non-related sympatric fishes will further have to clarify this hypothesis and demonstrate the true relation. The results of this study could possibly also be used in a taxonomic perspective, since morphological characters of teleost eggs have already been used this way (Kim & Park, 1996; Park & Kim, 2001) and phylogenetic relations within both Corydoras and the Callichthyidae are still unclear (Britto & Castro, 2002). However, Patzner & Glechner (1996) also question such an application since no association between egg morphology and taxonomy exists, at least not within the family Cyprinidae. Morin & Able (1983) seem to confirm this by reporting interspecific variation in Fundulus heteroclitus, but, however, do state that egg morphology has proved to help Cyprinodontiformes. identify Therefore, major groups information on and egg species within morphology of the other callichthyids will have to reveal whether a similar pattern is also present in those species and to what extent, if any, the use of a similar character in the phylogenetic reconstruction of this family would be advisable. 52 4 ONTOGENY – EARLY DEVELOPMENT AND GROWTH 4.2 EARLY DEVELOPMENT AND GROWTH Modified from: Huysentruyt F, Moerkerke B, Devaere S & Adriaens D Early development and allometric growth in the armoured catfish Corydoras aeneus (Gill, 1858) Hydrobiologia, submitted. Abstract Corydoras aeneus larvae were bred in captivity and collected at different ages in a closely set ontogenetic series. The development of external morphology and of allometries of several body parts was studied, attempting to reveal important steps in the species’ early life history. Based on external morphology, the different stages in early development of C. aeneus were identified, as described by Balon (1975). After hatching, at a SL of 3.5 mm, an eleutherembryonic phase was present, followed by the protopterygiolarval phase (4.4-5.7 mm SL), the pterygiolarval phase (5.7-14.0 mm SL) and the juvenile period. In addition, an overall growth curve was established and inflexion points were determined. Hence, it was of interest to determine the growth coefficient k in SL = b agek with the possibility to allow for changing k over different time points. To this end, both variables were transformed on the log-scale (log(SL) = log(b)+k log(age)) and a piecewise linear regression method was applied where I followed the ideas of regression spline smoothing procedures. This way, the growth curve was divided into six different stages of growth rate. Initially, the slope was 0.05 until 0.7 dph, then increasing to 0.18 until 4 dph, and 0.36 until 10 dph. After this, growth rate reached a maximum of 0.76 until 24 dph, slowed down to 0.47 until 37 dph and then finally again slowed down to 0.36. A similar analysis was also done on the data of growth in the different body parts and these results were compared to both morphological and literature data. This led to the conclusion that the inflexion points found in the early development of C. aeneus matched the different key-events known in teleost early development. The transition from endo- to exogenous feeding, when priorities also focus on respiratory functions, was the first point at which allometries changed together with functional demands. A second, similar alignment occurred at the transition to the pterygiolarval phase, when priorities shift toward locomotory needs. Finally, my results also indicated a transition to a carangiform swimming mode at approximately 8 mm SL. 53 4 ONTOGENY – EARLY DEVELOPMENT AND GROWTH Brief introduction Morphogenesis and differentiation are very intense during early life stages (van Snik et al., 1997) and teleost larvae commonly functionally optimize growth to increase fitness (Fukuhara, 1992). Both of these processes lead to a discontinuous larval growth (in terms of rate) (Gisbert, 1999), of which a thorough knowledge would help to understand functional trends and ecology of the species at different developmental stages (Fukuhara, 1992). Therefore, it was my objective to study the ontogeny of C. aeneus and, as a first step in this study, I gathered information on growth of the species in order to obtain insights into the timing of changes in ontogenetic allometries. Further, knowledge on the ontogeny can also serve aquacultural applications (Fukuhara, 1992). Materials and methods Breeding in Corydoras aeneus was induced as described under 2.1. Afterwards, specimens from various nests were removed post hatching at 1h, 2h, 3h, 6h, 12h, 24h, 2d, 3d up to 14d, every two days up to 24d, at 25d, 27d, 29d, 30d, 33d then every four days up to 45d, followed by one specimen every five days until 70d, one specimen every ten days until 90d and the last two specimens at 120d (in total 41 specimens were used). All specimens were collected and photographed as described under 2.1. Based on standard length (SL) in function of age (in days), an overall growth curve was established and inflexion points were determined. Hence, it was of interest to determine the growth coefficient k in SL = b agek (Fuiman, 1983), with the possibility to allow for changing k over different time points. To this end, both variables were transformed on the log-scale (log(SL) = log(b)+k log(age)) and a piecewise linear regression method was applied where I followed the ideas of regression spline smoothing procedures (see e.g. Friedman & Silverman, 1989). 54 4 ONTOGENY – EARLY DEVELOPMENT AND GROWTH Suppose we have n data points (xi,yi) (i = 1,…,n). In this case, xi and yi respectively represent the age and standard length of observation i on the logscale. Assume further that the data points satisfy: y i = f ( xi ) + ε i ε i ~ N (0, σ 2 ); i = 1, K, n (*) with f ( xi ) = β 0 + β1 xi + ∑ β j +1 ( xi − t j ) I ( xi > t j ) . m j =1 m is the number of ‘knots’ or inflexion points and {t j ; j = 1,K, m} is the set of these knots ( {10 j ; j = 1,K, m} is then the set on the original scale). Note that t min( xi ) < t1 < K < tm < max( xi ) and that we restrict {t j ; j = 1,K, m} to be a subset of {xi ; i = 1,K, n} . I ( xi > t j ) = 1 as soon as xi > t j and 0 otherwise. Model (*) implies that the slope or growth rate before the first knot equals β1 . After the first l knots ( l ≤ m ), the slope becomes β1 + ∑β l j =1 j +1 . The goal is to estimate f ( xi ) based on the observed data points. This means that a set of knots needs to be chosen and that the parameters β = ( β 0 , β 1 , β 2 , K , β m +1 ) need to be estimated. In the absence of knots, f ( xi ) = β 0 + β1 xi and the model to be fitted is a simple linear regression model. I opted to select a number of knots using a forward search procedure and to estimate β using ordinary least squares. More in particular, I followed the forward addition strategy as described in Lee (2002) who gives an overview of algorithms for ordinary least squares regression spline fitting. The generalized cross-validation (GCV, see Friedman & Silverman, 1989) was used as the criterion to select the ‘best’ model. The best model is then the model which minimizes 1 / n∑ ( y i − fˆ ( xi )) 2 n GCV = i =1 (1 − (3m + 1) / n) 2 with fˆ ( xi ) the estimate for f ( xi ) . The sequential search procedure works as follows. The initial model is the model without knots and its GCV-value is calculated. In each step of the search procedure, a knot is added such that the largest decrease or smallest increase in 55 4 ONTOGENY – EARLY DEVELOPMENT AND GROWTH GCV is seen. This process is repeated until a user-defined maximum number of knots is obtained. In all my analyses, this maximum was set equal to 10. After this procedure, the model with the smallest GCV among all candidate models is chosen as the final model. In order to investigate allometric growth, a similar analysis was done on growth of the different body parts, this time using SL as the independent variable. For this purpose, head length, abdominal length, post-anal length, caudal fin length, pectoral fin length and eye diameter (measured in a horizontal plane) were measured (fig. 4.3) and used as dependent variables. Not all measurements could be made on all specimens (e.g. a damaged caudal fin, anal opening not yet present, etc.), where this was the case, specimens were omitted from the analysis. This resulted in slightly different sample numbers for these measurements. Afterwards, the specimens were drawn and external morphology was described. The occurrence of important events was described providing the minimal SL at which these events were observed. Based on morphological criteria different stages of early development were demarcated and terminology was used according to Balon (1975, 1999). However, testing whether ontogeny in C. aeneus occurred saltatorily, was beyond the scope of this study, since this would require a more extensive study of disruptive morphogenetic events throughout the species’ ontogeny (see 5.1 for a more thorough approach of this matter). Results Hatching occurs after an incubation period of three to four days. Early development after eleutherembryonic hatching phase was divided (between into hatching four and different exogenous parts: the feeding), protopterygiolarval phase (until start of finfold differentiation), pterygiolarval phase (until completion of finfold differentiation) and juvenile period (until sexual maturation). 56 4 ONTOGENY – EARLY DEVELOPMENT AND GROWTH Morphological development At hatching, larvae have a SL of 3.5 ± 0.2 mm (± SD) (n = 6). From this size up to 4.4 mm SL a yolk sac is present. During this eleutherembryonic phase (fig. 4.4A), specimens are unpigmented and the mouth and anal opening are still closed. Distinctly large, ovally shaped, pectoral fins are present, which are more like large skin folds than true fins and thus do not yet bear any fin rays. Two pairs of oral barbels, the maxillary and external mandibular barbels, are present and the median finfold is undifferentiated. The eye diameter in these free swimming embryos is small and the eyes are still almost devoid of pigment. From three to five dph, at a SL of 4.4 to 5.7 mm, a yolk sac is no longer present and an extra pair of oral barbels, the internal mandibular barbels, begin to develop (fig. 4.4B). The anus and mouth open, as does the opercular cavity. At this protopterygiolarval phase, faint body pigmentation appears on the lateral side and head of these specimens. The pigmentation on the head is concentrated in a double line, which runs from the buccal area over the eye, to the back of the head. Further pigmentation is concentrated in spots on the lateral side of the body and in the finfold and pectoral fins. In these fins, pigmentation is concentrated where future fin rays are to develop. The eye in these specimens is larger and more pigmented. Near the end of this phase the median finfold begins to differentiate, which starts with an indentation behind the future dorsal fin and a slight narrowing of the finfold near the caudal peduncle. In specimens aged 7-23 dph, with a SL between 5.7 and 14.0 mm (the pterygiolarval phase), the finfold further differentiates (fig. 4.4C; D; E). At 11 dph (7-8 mm SL) the dorsal fin is almost fully detached from the finfold, pelvic fins start to develop and pigmentation is present at the base of the future anal and adipose fin, where in both cases the finfold also starts to indent. At 16 dph (11-12 mm SL) all fins are fully detached, except for the adipose fin. This fin is still surrounded anteriorly by a small part of the median finfold, which totally disappears at a SL of 14 mm. The fin rays in the different fins develop in a partially different order. Fin rays begin to develop in the caudal fin, pectoral and dorsal fins first and at a SL of 8 mm both dorsal and caudal fin have reached their definite number of fin rays. At this time the pelvic and anal fin rays start to develop, both of which become fully developed at a SL of 9-10 mm. A spine within the adipose fin is present from a SL of 11-12 mm on and in the pectoral fins the definite number of fin rays is reached at 22-25 mm SL. At 11 mm SL, the 57 4 ONTOGENY – EARLY DEVELOPMENT AND GROWTH first bony scutes, typical for callichthyid fishes, start to develop as well. The first scute that develops is the one anterior to the adipose fin, which partially covers the adipose fin spine. From there on three more, small, dorsal scutes develop in an postero-anterior order. At a SL of 15 mm, several small scutes develop dorsally and ventrally on the caudal peduncle. Finally, at approximately 16 mm SL, the first lateral scutes start to develop postero-anteriorly along the midline of the body with a further centripetal development of all scutes. Allometric growth In the growth curve of Corydoras aeneus, which ranges from 0 until 120 dph, five different inflexion points were calculated with subsequent changes in growth rate (slope) (fig. 4.5). The first inflexion in the growth curve is located at 0.7 dph (which corresponds to a fitted value of 3.9 mm SL), at which the estimated growth rate increases from 0.05 to 0.18 (p<0.01). Further, the growth curve in specimens over 0.7 dph shows a second inflexion at 4 dph (5.4 mm SL), where the slope further increases to 0.36 (p<0.01). A subsequent third inflexion is present at 10 dph (7.4 mm SL), where growth rate increases even further up to 0.76 (p<0.001). After this, at 24 dph (14.5 mm SL), growth slows down to a rate of 0.47 (p<0.01). Finally, the last inflexion in this growth curve is located at 37 dph (17.9 mm SL), where growth rate again decreases to 0.36, but this growth rate is not significantly different from the rate found before this point (p = 0.20). Note that I used the GCV-criterion to select the best model and not statistical significance. In the context of body part allometries, the model with one inflexion point (R² = 0.958) has the smallest GCV. This model shows that growth of the head is positively allometric (slope = 1.44) up to a SL of 13.1 mm, after which the rate drastically decreases to 0.69 (p<0.001) (Fig. 4.6A). Growth of the abdomen is also best modelled by a single inflexion point model (R² = 0.947) and in this case, growth starts nearly isometric (slope = 1.19), and then declines to 0.80 (p<0.05) from 9.8 mm SL on (fig. 4.6B). In the regression which describes growth in the postanal region, no improvements over the simple linear regression model were found in GCV by introducing inflexion points. This way, growth rate in the postanal region in relation to SL was found to remain constant at 0.81 and significantly different from 0 (p<0.001) throughout the entire ontogenetic series studied here (R² = 0.909) (fig. 4.6C). The analysis of growth 58 4 ONTOGENY – EARLY DEVELOPMENT AND GROWTH in the caudal region showed a two inflexion point model as best suited (R² = 0.978). Growth in this region starts out highly positively allometric (slope = 4.72) up to a SL of 6.2 mm, after which it slows down between 6.2 and 21 mm SL to a rate of 1.38 (p<0.001). After this it finally decreases to a rate of -0.20 (p<0.001) (fig. 4.6D). Further, a single inflexion point was also found in the growth curve of the eye (R² = 0.973). This inflexion was situated at a SL of 9.8 mm where the rate drastically decreases (slope = 1.89 to 1.02, p<0.001) (fig. 4.7A). Finally, in the growth of the pectoral fins, no inflexions were detected, which again leaves the simple linear regression as the best fit (R² = 0.924), with a constant growth rate of 1.22 (significantly different from 0 with p<0.001) as a result (fig. 4.7B). Discussion Hatching size in Corydoras aeneus is found to be quite constant (3.5 ± 0.2 mm (n = 6)) at 25°C, which is presumed to be correlated to the low variance in egg size (1.47 mm ± 0.20 (n = 21)) (Osse & van den Boogaart, 1995) (see also 4.1). The scope of this paper, however, was to study growth allometries in C. aeneus, and the timing of shifts in growth rates, both for overall growth as for growth of specific body regions (head, abdomen, post-anal region, caudal fin, pectoral fin and eye diameter). I hypothesized that the chronology of these allometric changes would be related to the chronology of important early life history events, and would therefore reflect an evolutionary ontogenetic response to functional demands. It has already been suggested that allometric growth patterns closely match the expected priorities for executing the necessary biological roles (Osse et al., 1997), or as Kováč & Copp (1999) put it: “a certain level of development is necessary,... , to coincide in functional readiness”. Therefore, in this study, the chronology of relevant shifts in allometric growth rate was statistically assessed by calculating inflexion points in growth rates of different body regions during ontogeny. In addition, literature on early life history traits in teleosts in general provided the framework to verify possible correlations with the allometries found in this study. 59 4 ONTOGENY – EARLY DEVELOPMENT AND GROWTH A first correlation I hypothesized was based on the fact that predation is one of the main agents that result in larval mortality (Bailey & Houde, 1989). As a result, a higher efficiency in the use of structures that counter this threat can be expected in these early life stages. In contrast to some callichthyid congeners that produce foam nests for their hatchlings (e.g., Megalechis), such hiding facilities are not provided for larval C. aeneus. Consequently, predator avoidance would require an increase in swimming efficiency, and thus burst swimming can be expected to be an important escape response in C. aeneus hatchlings. Weihs (1980), in his study on Engraulis mordax, found a continuous swimming mode to be more efficient in larvae under 5 mm, while in larvae with a length between 5 and 10 mm burst-and-coast swimming became more effective. Blaxter (1986), on the other hand, mentions that a change in swimming mode occurs as the caudal fin develops and inertial forces start to play a bigger role in locomotion. My observations indeed show a higher growth rate of the caudal region (slope = 4.72) up to 6.2 mm SL, which provides a ‘functional readiness’ for burst and coast swimming after this stage. At a SL of about 5 mm, notochord flexion occurs, which precedes both a drastic increase in overall growth rate (0.18 to 0.36) and the start of the pterygiolarval phase, both of which occur around 5.5 mm SL. These factors indeed indicate a shift in the components of the locomotory apparatus which may be related to an onset of a change in swimming mode at 5 mm SL. Such a shift toward burst and coast swimming is believed to enlarge the dispersal and foraging range and improve predator avoidance (Gisbert, 1999). Therefore, based on these findings, I expect a similar change in swimming mode to occur at that point but a kinematic analysis of the ontogeny of larval swimming in C. aeneus will have to confirm this. Apart from predation, a second correlation I predict is based on the fact that starvation becomes a major threat for larval survival once the yolk sac becomes depleted (Bailey & Houde, 1989). Accordingly, the point at which this depletion occurs is critical during larval development (Pedersen et al., 1990; Jardine & Litvak, 2003). After this depletion, exogenous feeding becomes obligatory and the presence of a functional feeding apparatus is required (van Snik et al., 1997; Jardine & Litvak, 2003). Given this, I expected that under 4.4 mm SL (= point of yolk sac depletion) development would, next to predator avoidance, also be focused on the completion of the differentiation of the feeding apparatus to a first functional stage. My observations show that, around this transition in C. 60 4 ONTOGENY – EARLY DEVELOPMENT AND GROWTH aeneus (4.4 mm SL), growth rate in head length is positively allometric (slope = 1.44). During the early life stages, protrusion of upper jaw bones is not yet functional so that high suction forces must be generated in order to catch prey. Higher suction forces can be related to head volume (Osse, 1990), which implies that a fast head growth (as is the case here) would indeed allow an increase in suction efficiency. The fact that, in the earliest stages of development in some teleosts, growth intensities are not distributed uniformely across the body was already observed by Fuiman (1983). For the earliest stages, Fuiman (1983) describes a continuous U-shaped gradient across the body axis in between a high intensity of growth rates at the terminal growth centres (head and tail). Fuiman (1983) also described that in juveniles, this gradient approaches isometry, which probably remains throughout adulthood. As discussed above, in C. aeneus, a higher growth rate in both head and tail are also present in the earliest phases. Fig. 4.8 shows those different growth rates mapped along the body axis for three different phases of development in C. aeneus. The graph also shows that growth turns almost isometric near the start of the juvenile phase, with the exception of the tail region. However, since standard error in the growth rate of the tail region is very high, it seems probable that the model of Fuiman (1983) also applies here. Fuiman (1983) explained this U-shaped profile in early development discussing the need of propulsive force provided by the tail region (to escape predators), and the need of a large head for the elaboration of feeding and sensory mechanisms, as we also already suggested for C. aeneus. Given the properties of the physical environment, like the high viscosity of the aqueous medium and of the biological environment such as the relatively large size of the available prey organisms considered, size acts as an important constraint on development of form and function during early ontogeny. This is a view that gains importance in many recent developmental studies on fish larvae (Strauss, 1984; Buckel et al., 1995; Stern & Elmen, 1999; Adriaens & Verraes, 2002) and can be exemplified here too. Feeding strategies of fish larvae as well as locomotor habits demonstrably show pronounced changes at certain developmental stages and certain size ranges are rather strictly correlated to typical morphogenetic events or differentiation of particular organ systems. In Cyprinus carpio (van Snik et al., 1997), for example, a second change in swimming mode from anguilliform to subcarangiform was found around 8 mm 61 4 ONTOGENY – EARLY DEVELOPMENT AND GROWTH SL. Gisbert (1999) found a change in swimming style to be correlated to the first appearance of caudal fin rays in the chondrostean fish Acipenser baeri. My observations show that at 8 mm SL caudal fin rays start to develop and anal, adipose and pelvic fins separate from the respective finfolds and that, just prior to this, at 7 mm SL, overall growth rate reaches a maximum (0.76). Therefore, a similar change in swimming mode is expected around 8 mm SL in C. aeneus. This altered swimming would also imply a decrease of the head yaw (Osse, 1990), which again results in better aiming skills for prey capture. In this context a decrease of eye and abdomen growth rate (1.19 to 0.8 and 1.89 to 1.02, respectively) shortly after this shift indeed may indicate changes in prey preference and/or visual performance in relation to predation. Conclusively, the analysis of inflexion points in growth curves, in combination with observed changes in morphological development, holds valuable information on changes in functional demands throughout ontogeny. Such an analysis can provide a framework of these shifting functional demands placed on a developing larvae, useful for both the comparison with the results of additional ontogenetic research on the species, as well as for aquacultural applications. 62 4 ONTOGENY – CRANIUM 4.3 THE CRANIUM Modified from: Huysentruyt F, Brunain M & Adriaens D Ontogeny of the chondrocranium in Corydoras aeneus (Gill, 1858) (Callichthyidae, Siluriformes) Journal of Morphology, accepted. and: Huysentruyt F, Geerinckx T, Brunain M & Adriaens D Development of the osteocranium in Corydoras aeneus (Gill, 1858) (Callichthyidae, Siluriformes) Journal of Morphology, submitted. Abstract Callichthyids take a basal position in the loricarioid evolutionary lineage leading up to an algae scraping feeding mechanism in the loricariid family. Therefore, the study of the morphology and development of a callichthyid representative would contribute to a better knowledge on the differences in cranial morphology and their impact on feeding ecology within this superfamily. Therefore, development in the cranium of Corydoras aeneus was studied based on 48 cleared and stained specimens and ten series of serial sections. The latter sections were also digitized and used for 3D-reconstructions. Development overall follows the typical siluriform trends in cranial development. Even the low complexity of the chondrocranium at hatching fits the trend observed in other siluriforms, although other studies showed loricarioid hatchlings to generally show more complex chondrocrania. In contrast to other catfish, in C. aeneus, the notochord was never found to protrude into the hypophyseal fenestra. In addition, also differing from other siluriforms, a commissura lateralis is present, a state also reported for Ancistrus cf. triradiatus (Geerinckx et al., 2005). The splanchnocranium again has the typical siluriform shape during its ontogeny, with the presence of a compound hyosymplecticpterygoquadrate plate, although not fused to the neurocranium or interhyal at any time during ontogeny, a state described earlier for Callichthys callichthys (Hoedeman, 1960b; Howes & Teugels, 1989). The most striking differences found in comparison to other catfishes, however, involves the branchial basket, which arises as a single element with a further differentiation from the middle arches on in both a rostral and caudal direction. As in other studies on siluriform osteocranial formation, ossifications generally appeared as a response to functional demands (Tilney & Hecht, 1993; Vandewalle et al., 1995; 63 4 ONTOGENY – CRANIUM Vandewalle et al., 1997; Adriaens & Verraes, 1998). This way, early development of the skull was found to occur in two distinct phases. In a first phase, several new bony elements, all of dermal origin and related to feeding, appear shortly after yolk depletion (4.4 mm SL). After this, in between 5 and 8 mm SL, developmental priorities seem to shift to size increase of the cartilaginous skull and no new bony elements appear. Finally, a second phase of osteogenesis occurs from 8-18 mm SL, in which all remaining dermal and perichondral bones appear. Brief introduction A thorough knowledge of anatomical ontogeny is of critical importance in understanding the functional trends during early development, since organisms must be functional at each moment, including the early ‘temporary’ stages (Fukuhara, 1992; Koumoundouros et al., 2001a, b; Geerinckx et al., 2005). In this context, the study of the morphology and development of C. aeneus can contribute to a better knowledge on the differences in cranial morphology and their impact on feeding ecology within this superfamily. This is especially the case for the earliest stages in ontogeny, at which point the yolk sac often becomes depleted and feeding becomes obligatory. Therefore, the aim of this study is to provide a complete description of the ontogeny of the cranium in C. aeneus, adding to the wide variety of data on catfish cranial development (Kindred, 1919; Ballantyne, 1930; Bamford, 1948; Srinivasachar, 1957a, b, 1958a, 1959; Hoedeman, 1960b; Vandewalle et al., 1985; Surlemont et al., 1989; Surlemont & Vandewalle, 1991; Adriaens & Verraes, 1994, 1997b; Geerinckx et al., 2005) and providing a basis for further ongoing ontogenetic research on this species. The data generated this way were compared to the general trends in siluriforms and to similar data published earlier by Geerinckx et al. (2005) on Ancistrus cf. triradiatus, a loricariid representative. The comparison of the differences and similarities of both ontogenetic sequences provides additional information on the evolutionary processes leading up to the remarkable niche diversification in the Loricariidae family. This way, I hypothesize that, given the close phylogenetic relationship of the species compared, both ontogenies would exhibit a great amount of 64 4 ONTOGENY – CRANIUM similarity, only showing significant differences in the development of those elements associated with the specialized feeding mechanism in A. cf. triradiatus. Given the fact that the study of the developing cranium in C. aeneus is hereto undescribed, an additional aim of this study is to link the morphological aspects of the developing skull to the species’ most important early life history traits. Therefore, results are compared to data on the species’ early life history (see also 4.2), providing knowledge on morphological trends guided by possible functional constraints and environmental preferences at different developmental stages (Fukuhara, 1992). This way, it is also expected that differences in functionality during the early ontogeny of the skull between the different species studied would be reflected in the timing and sequence of the developing structures associated with these specific functions. Brief materials and methods The specimens examined are presented in table 2.1. For clearing and staining procedures, procedures on serial sectioning and 3D-reconstructions see 2.2. Bone terminology used throughout this paper follows Harrington (1955), Adriaens & Verraes (1998) and Arratia (2003a). In the earliest stages, 3Dreconstructions sometimes yielded some left-right asymmetry, which, however, could very well be an artifact of the digitization, since very low amounts of cartilage are present in these stages. Therefore, and given the fact that, in no other study done on early catfish chondrocrania, as well as in my cleared and stained specimens a left-right asymmetry was found, this asymmetry was treated as an artefact and not further discussed. Results Stage 1: 3.3 mm SL (hatchling) (fig. 4.9A) Neurocranium 65 4 ONTOGENY – CRANIUM At hatching, the neurocranium is rudimentary and chondrification is concentrated around the notochord tip. Bordering this notochord tip laterally are two parachordal cartilages, which interconnect anterior to the notochord through the acrochordal cartilage. On their lateral margins, both parachordal cartilages connect to the basiotic laminae, making up the primordium of the basal plate. Further laterally this plate contacts the dorsocaudally oriented anterior part of the otic cartilages through the anterior basicapsular commissure. This way, two large posterior metotic fissures appear, through which the nervus glossopharyngeus (IX) and vagus (X) pass. At the rostral margin of this basal cartilaginous skull, two polar cartilages are present. Splanchnocranium At about the time of hatching, the splanchnocranium is confined to the suspensorium, which arises as a single chondrification or hyosymplecticpterygoquadrate plate (sensu Arratia, 1992) and articulates dorsally with the neurocranium at the level of the otic cartilage. This hyosymplectic- pterygoquadrate plate does not yet bear a process for articulation with the operculare, nor is the foramen truncus hyomandibularis formed at this stage. Stage 2: 4.0 mm SL (figs. 4.9B; 4.13A; B) Neurocranium At 4.0 mm SL, the trabecular bars are formed as rostral expansions of the polar cartilages. As is to be expected in a platybasic teleost skull, these trabecular bars lie well separated from each other and are slightly curved. Not yet touching rostrally, they leave a wide hypophyseal fissure at the anterior margin of the skull. Since the internal carotid artery still passes through this large fissure (and not through a smaller fissure in the cartilaginous trabecular bars themselves) it is impossible to distinguish the trabecular bars from the polar cartilages at this point. Observations on later stages and other siluriform fish, however, lead to the assumption that the posterior part at this point would correspond to the polar cartilage (Adriaens & Verraes, 1997b; Geerinckx et al., 2005). Splanchnocranium At 4.0 mm SL, the entire hyoid arch is also formed as a single element, incorporating both left and right hypohyals and ceratohyals. In addition, the hyosymplectic-pterygoquadrate plate has broadened and shifted its orientation 66 4 ONTOGENY – CRANIUM from oblique to vertical. At this level, the dorsal and ventral parts of the suspensorium can be distinguished as the hyosymplectic and quadrate part respectively. In the branchial basket, which is still very much compressed dorsoventrally, the presence of chondrocytes, as observed on serial sections, already indicates the onset of branchial arch formation, in which all five arches are present and fused together ventrally. In addition, the first four arches are also fused at the dorsal side of the basket, forming a large dorsal plate. Stage 3: 4.4 mm SL (figs. 4.10; 4.13C; 4.14A; B) Neurocranium At this stage, both trabecular bars contact each other rostrally, forming the ethmoid cartilage, and expanding laterally into the solum nasi. Thus, the fissura hypophysea is closed off, forming the fenestra hypophysea. At the caudal side of this fenestra, the otic cartilages have expanded ventrolaterally, contacting the polar cartilages through the lateral commissures, splitting up the fenestra sphenoidea into anterior and posterior parts. At this stage, serial sections already show the passage of the nervus oculomotorius (III), trochlearis (IV), trigeminus (V), and the mandibular branch of the nervus facialis (VII) through the anterior part and the passage of the ramus hyomandibularis of the nervus facialis (VII) through the posterior part. At the level of the basal plate an additional connection with the otic cartilage is formed through the posterior basicapsular commissure, splitting up the metotic fenestra. Serial sections further indicate that the aperture present between anterior and posterior basicapsular commissure is penetrated by the nervus glossopharyngeus (IX), but not by the nervus vagus (X), which makes this the basicapsular fenestra. The posterior subdivision of the metotic fenestra is not penetrated by any nerves and will here be called foramen “A.” Caudally, two occipital pilae have emerged, which fuse at the dorsal side of the neural tube, forming the tectum posterius. At the caudal margin of this chondrocranium, the lamina basiotica has further expanded, rigidifying the neurocranial floor. At the dorsal side of the chondrocranium, both otic capsules have expanded rostrally forming the taeniae marginales, which already interconnect through the epiphyseal bridge but do not yet protrude further rostrally. 67 4 ONTOGENY – CRANIUM Splanchnocranium At 4.4 mm SL, both the interhyal, Meckel’s cartilage and autopalatine first appear, all as separate elements. The hyosymplectic-pterygoquadrate plate has developed a processus opercularis for the later articulation with the opercular bone. In the branchial basket, the central copula has split up into an anterior and a posterior part, with the former part situated near the hypobranchial parts of the first three branchial arches and the latter part with those of arches IV and V. The medial tip of the hypobranchial part of the third arch is situated in between these two copulae. Furthermore, hypobranchials II and III are already present as separate elements. All arches already have ceratobranchial cartilages and arches I-IV bear epibranchials, none of which has detached from the ceratobranchial parts at this point. One large infrapharyngobranchial element is present as a separate element, with the anterior tip articulating with epibranchials one and two, and the posterior part articulating with epibranchials III and IV. Teeth are already present in both the dorsal and ventral part of the basket, but no ossifications supporting these teeth have been formed at this point. This is also the case for both both lower and upper oral jaw, where teeth primordia are also present without the presence of supportive ossifications. The autopalatine still articulates with the maxillary barbel through the submaxillary cartilage and the cartilaginous hyosymplectic-pterygoquadrate plate has developed an articulation with the opercular bone, the first ossified element to appear. Stage 4: 4.9 mm SL (figs. 4.11; 4.13D; 4.14C; D)) Neurocranium At this stage, the taenia marginalis makes contact with the solum nasi, through the commissura spheno-ethmoidalis, forming the lamina orbitonasalis. Rostrally, the taenia marginalis further expands, forming the onset of the commissura sphenoseptalis. In the skull base, the fenestra basicapsularis is split up by the commissura basivestibularis into an anterior and a posterior part, with the latter accommodating the nervus glossopharyngeus (IX). Splanchnocranium The hyosymplectic-pterygoquadrate plate further differentiates into a more distinct pars quadratum, which articulates with Meckel’s cartilage and a pars hyosymplecticum, articulating with the neurocranium. In the latter, the foramen 68 4 ONTOGENY – CRANIUM truncus hyomandibularis is present from this stage on (not visible in fig. 4.11). Additionally, the onset of a pterygoid process is present. The hyoid arch has started to split up medially and the two lateral halves have begun to differentiate into a more distinct ceratohyal and hypohyal part, with the latter articulating with the interhyal. In the branchial basket, epibranchials II and III have detached from the ceratobranchials but in this specimen, the hypobranchial of the second arch was still loosely connected to the ceratobranchial (even though they were already separated in the specimen used for the previous stage). Cartilage had also disappeared at the level of the articulation between cerato- en epibranchial IV, indicating that both elements probably detach soon after this stage. In addition, the infrapharyngobranchial cartilage has expanded anteriorly and has started to differentiate into a more clear anterior and posterior part, again clearly articulating with, respectively, the first and last two epibranchials. Both a thin layer of dentary bone and the maxillary and premaxillary bones are present, all of which are dermal ossifications supporting teeth. The maxillary also supports the maxillary barbel at this point and articulates through a double dorsal process and two submaxillary cartilages with the autopalatine. The bone further has a medial articulation with the premaxillary, which in turn bears a dorsal process for articulation with the ethmoid cartilage. In addition, the parurohyal bone has started to ossify in the tendon connecting the sternohyoideus muscle to the hyoid. Stage 5: 5.3 mm SL (figs. 4.12; 4.13E; 4.14E; F)) Neurocranium At this stage, the solum nasi has expanded laterally and the commissura sphenoseptalis now is fused to the ethmoid cartilage at the level of the lamina precerebralis. The preorbital base forms another vertical connection between taeniae marginales and trabeculae cranii slightly behind the lamina orbitonasalis, leaving a small orbitonasal foramen, which disappears later during ontogeny. Therefore a large foramen fila olfactoria appears. The cartilage of the skull floor has expanded further, comprising a caudal expansion of the acrochordal cartilage and closing off both the foramen “A” as well as narrowing the foramen for the passage of the nervus vagus (X). 69 4 ONTOGENY – CRANIUM Splanchnocranium The pterygoid process of the hyosymplectic-pterygoquadrate has expanded significantly. In the branchial basket the first epibranchial has detached from the rest of the arch but in the fourth arch, this element is still slightly connected to the ceratobranchial. In middle of the fifth ceratobranchial, along its medial side, a thin bony plate supporting the ventral pharyngeal teeth has started to develop. Along the ventromedial side of infrapharyngobranchials III and IV, a similar ossification was observed, supporting the dorsal pharyngeal teeth. Stage 6: 6.3 mm SL (fig. 4.13F) Neurocranium No additional changes have occurred in the chondrocranium at this point. Splanchnocranium At this point, in the gill arches, all epibranchials have detached from the ceratobranchials and the infrapharyngobranchial part has split up into a posterior and anterior part articulating with the last and first two arches, respectively. In both the fourth and first arch, the hypobranchial part is still fused to the ceratobranchial part. In the first arch, these elements will separate shortly after this stage, while, in the fourth arch, both elements never detach. 8.2 mm SL (figs. 4.15; 4.20A; 4.21A) Neurocranium In the skull roof, three ossification centres are present. Rostrally, the dorsal perichondral supraethmoid bone expands over the ethmoid cartilage. Next to this, on the dorsomedial side of the taeniae marginales, just behind the epiphyseal bridge, the frontal bones appear. On the caudal margin of the neurocranium, the parieto-supraoccipital bone has started to ossify at the level of the tectum synoticum. The skull floor is made up of the ventral hypoethmoid bone (not illustrated), caudally contacting the prevomeral bone, which is flanked by two palatal splints at is posterior margin. Caudal to this ossification, the borders of the hypophyseal fenestra have started to ossify, forming the early parasphenoid bone. Caudolaterally, the early posttemporo-pterotico- supracleithra are present, in which the separate elements composing this complex bone can not be distinguished. 70 4 ONTOGENY – CRANIUM Splanchnocranium Ossification of the palatine has started at its posterior margin and, in the lower jaw, Meckel’s cartilage has also started to ossify. This way, at its anterior margin, the mentomeckelian appears, while posteriorly, the articular and retroarticular bone become apparent, still separated at this point. In the suspensorium, the hyomandibular and quadrate parts show ossification centres, enforcing articulations with the opercular and articular bone respectively. In the hyoid arch, the ventral hypohyal and anterior and posterior ceratohyals ossify, as do the branchiostegal rays. Finally, in the opercular series, the interopercle and suprapreopercle have formed. The opercle has now developed its typical oval shape and oblique dorsoventral orientation. 9.7 mm SL (figs. 4.15B; 4.18A; 4.20B; 4.21B) Neurocranium The supraethmoid bone has expanded caudolaterally and ventrally, it has fused to the hypoethmoid bone, forming the mesethmoid bone. Laterally, this complex shows two expansions of dermal origin, which represent the laterodermethmoid parts. The frontals have expanded in both the anterior and posterior direction, next to a small expansion which covers the epiphyseal bridge and connects both at the midline. At the caudal margin of the neurocranium, anterior expansions of the parieto-supraoccipital bone almost reach the frontal bones. In the skull floor, the hypophyseal fenestra has started to close, due to further ossification of the parasphenoid. Next to this, the orbitosphenoids have started ossifying at the anterior border of the sphenotic fenestra. In the otic region, both the prootics in the skull floor as well as the sphenotics in the skull roof are present. In the latter bones, the dermosphenotic and autosphenotic parts are already fused. At the caudal margin of the otic region, the posttemporo-pterotico-supracleithrum has expanded dorsally, housing the posttemporal branch of the cranial lateral line system. Ventrally, the basioccipitals and exoccipitals are present. Splanchnocranium The pterygoid process has started to ossify at its tip, forming the metapterygoid and, next to this, the preopercular bone is present. 71 4 ONTOGENY – CRANIUM Gill arches Ceratobranchials III-IV, epibranchials I-IV, infrapharyngobranchial III (anterior infrapharyngobranchial) and infrapharyngobranchial IV (posterior infrapharyngobranchial) have all started to ossify in the centre of their respective cartilaginous elements. 11.2 mm SL (figs. 4.16A; 4.20C; 4.21C) Neurocranium All skull roof bones have expanded, rigidifying the entire skull. The lateral ethmoid bone is present, with the prefrontal and parethmoid parts directly appearing as fused elements. Further, the pterosphenoid ossifies at the posterodorsal margin of the fenestra sphenoidea and, in the skull floor, the hypophyseal fenestra has closed entirely. Splanchnocranium In the splanchnocranium, ossification of the interhyal bone has started. Gill arches Cerato- and hypobranchials I-II have began to ossify, although, in contrast to the gill arch bones formed earlier, ossification starts at the rostrolateral side of the cartilaginous hypobranchials. 13.9 mm SL (figs. 4.16B; 4.18B; 4.20D; 4.21D) Neurocranium Bone expansions have further rigidified the skull structure. Splanchnocranium All ossified parts have expanded and in the hyoid arch, with the appearance of the dorsal hypohyal, all bones are present. Gill arches Basibranchials II and III ossify at this stage, completing ossification in the branchial arches. 72 4 ONTOGENY – CRANIUM 16.0 mm SL (figs. 4.17A) Neurocranium At this stage, only the nasal and first infraorbital appear as new ossifications in the neurocranium. 18.0 mm SL (figs. 4.17B; 4.19) With the second infraorbital present, the skull resembles the adult configuration. Next to this, the palatal splints have merged to the prevomer, although additional observations on serial sections, show them to still be present in some adult specimens, albeit highly diminished in size. This suggests that their reported absence in other callichthyid genera like Callichthys, Dianema, Lepthoplosternum, Hoplosternum and Megalechis is probably the result of a fusion or even secondary loss during ontogeny. Discussion of chondrocranial development At hatching the chondrocranium is rudimentary and concentrated around the notochord tip, a situation common in catfishes (Adriaens & Verraes, 1997b). The fact that chondrification has started in both the neurocranium and splanchnocranium at this stage is a state also described in most studies dealing with teleost ontogeny (e.g., de Beer, 1937; Adriaens & Verraes, 1997b; Geerinckx et al., 2005). In contrast, however, the complexity of the chondrocranium at hatching is highly variable in teleosts. Species like Salmo letnica, Salmo trutta fario (Salmoniformes) and Oryzias latipes (Beloniformes) already exhibit a fairly high level of neurocranial chondrification at hatching, while in other teleost (Synbranchiformes), Solea species solea such as Mastacembelus (Pleuronectiformes) and armatus Hepsetus odoe (Characiformes) no chondrification is present at that time (de Beer, 1927; Bhargava, 1958; Bertmar, 1959; Langille & Hall, 1987; Wagemans & Vandewalle, 1999; Ristovska et al., 2006). However, several studies indicate that siluriform chondrocrania are generally limited in complexity at the time of hatching. Species like Chrysichthys auratus (Claroteidae) and Heterobranchus longifilis and Clarias gariepinus (Clariidae) show little to no chondrification at 73 4 ONTOGENY – CRANIUM hatching while the chondrocranium in hatchlings from other species like Heteropneustes fossilis (Heteropneustidae) is only slightly developed (Srinivasachar, 1959; Adriaens & Verraes, 1997b; Vandewalle et al., 1997, 1999). Studies on species of the Loricarioidea superfamily, like Ancistrus cf. triradiatus (Loricariidae) and Callichthys callichthys (Callichthyidae), however, show hatchlings of this group to have a more developed chondrocranium (Hoedeman, 1960b; Geerinckx et al., 2005). As an exception, Hoplosternum littorale (Callichthyidae), only showed localized patches of connective tissue where chondrification was taking place, but without actual cartilage formed at the time of hatching (Ballantyne, 1930). This way, the complexity of the chondrocranium at hatching in Corydoras aeneus seems to fit the general trend found in Siluriformes, rather than that found in the Loricarioidea. In this context, egg size has often been suggested as a key factor influencing larval development and chondrocranium complexity at hatching since this implies an increase in yolk material, frequently related with the duration of the prehatching development (Araujo-Lima, 1994; Adriaens & Vandewalle, 2003). For example, in Galeichthys feliceps (egg size: 15.6 mm) (Ariidae) hatchlings already possess complete bony skulls, whereas in the loricariid A. cf. triradiatus (egg size: 3.1 mm) a highly developed chondrocranium is present at that point (Tilney & Hecht, 1993; Geerinckx et al., 2005). In comparison, the egg size of 1.5 mm in C. aeneus is rather small, indeed resulting in a less developed chondrocranium at hatching (see 4.1). Next to body size and clutch size, egg size is often suggested to be related to parental care (Kolm & Ahnesjö, 2005). Such parental care, observed in many loricariid species (Suzuki et al., 2000), is unreported in Callichthyidae, which could explain the smaller egg size in members of this group. As a result, yolk sac amount is also limited, which speeds up the time of hatching during development. Indeed, C. aeneus hatches three days after fertilization at a SL of 3.5 mm, at which stage the chondrocranium is still primordial, shortening the period between onset of chondrification (hatching) and yolk sac depletion (4.4 mm) to three days. For comparison, the larger egg size in A. cf. triradiatus indeed leads to a prolonged prehatching period of five days, yielding a much more complex neurocranium at hatching. After this, development takes another five days until the yolk sac becomes depleted, this way prolonging the entire period of chondrocranial formation up to the level of yolk sac depletion (and thus functional complexity) up to 10 days. Therefore, although the pathways of 74 4 ONTOGENY – CRANIUM chondrocranial development seem to be more or less rigid within large taxa, the time available for this development could very well play an important role in creating the necessary opportunities for the evolution of innovative structures from the earliest stages on. Neurocranium Skull floor As in other siluriform fish, the C. aeneus skull is of the platybasic type, often related to the reduction in eye size, as is typical for this group (Daget, 1964; Adriaens & Verraes, 1997b). At hatching, the rostral cartilages of the chondrocranial floor do not yet expand beyond the polar cartilages, as could be observed by examination of serial sections revealing that the path of the arteria carotis interna (commonly accepted to mark the border between the trabeculae cranii s.s. and the polar cartilages (Goodrich, 1958; Bertmar, 1959; Adriaens & Verraes, 1997b)) passes in front of them. Examination of sections of later ontogenetic stages showed the trabeculae cranii s.s. to be present only from 4.0 mm SL on, with the arteria carotis interna running medial to this cartilage through the hypophyseal fenestra. Finally, from 4.9 mm SL on, these cartilages expand medially, forming a fissura through which this artery runs. At 5.3 mm SL this fissura closes medially, incorporating the artery in the trabecular bars. Since, however, like in A. cf. triradiatus, no evidence of cartilage resorption is seen, the mechanism suggested for the appearance of this fissure is that of allometric growth of the trabecular bars (Geerinckx et al., 2005). The notochord never protrudes into the hypophyseal fenestra in C. aeneus, a situation different from other catfish like Ariopsis felis and Arius jella (Ariidae), other loricarioids like A. cf. triradiatus and even other callichthyids like C. callichthys (Bamford, 1948; Srinivasachar, 1958a; Hoedeman, 1960b; Geerinckx et al., 2005). Further, at 4.4 mm SL, the trabecular bars fuse with the ethmoid cartilage closing the hypophyseal fenestra. This fusion occurs at the caudal border of the ethmoid plate, as is generally the case, although this fusion has been reported to occur at the dorsal face of the ethmoid plate in C. callichthys and H. littorale (Ballantyne, 1930; Hoedeman, 1960b). The ethmoid cartilage then further expands in both a dorsal and lateral direction, forming the precerebral lamina en solum nasi respectively. Also at 4.4 mm SL the metotic fenestra is split up into an anterior 75 4 ONTOGENY – CRANIUM and posterior part, with the first accommodating the passage of the nervus glossopharyngeus (IX). Since in teleosts no nerves pass through the anterior basicapsular fenestra, the anterior fenestra found here is homologous to the basicapsular fenestra and the commissure dividing both the anterior and posterior fenestra can be designated the commissura basicapsularis posterior (Daget, 1964). As a consequence, the posterior fenestra can not be homologized to any of the described fenestrae in teleost chondrocrania and is therefore here called foramen “A”. At 4.8 mm SL the basicapsular fenestra is again split up, this time providing passage for the glossopharyngeus (IX) nerve in the posterior part, making these openings the fenestra basicapsularis anterior and posterior, split up by the commissura basivestibularis. From a SL of 5.3 mm SL on the foramen “A” disappears and caudally, the lamina basiotica has further expanded filling up the neurocranium floor, leaving the fenestra of the vagus (X) nerve. Skull roof An actual skull roof in C. aeneus is formed from a SL of 4.4 mm SL on. At this point, at the rostral skull roof margin, the otic capsules have formed taeniae marginales which interconnect through the epiphyseal bridge. At its caudal side both halves of the neurocranium also interconnect, forming the tectum posterius as a skull roof element. In other catfish, loricarioids and even callichthyids, the tectum posterius generally closes at a much higher SL. This occurs at a SL of about 7 mm in C. gariepinus (7.1 mm), A. cf. triradiatus (6.8 mm) and C. callichthys (6-7 mm), while in H. fossilis (12 mm) and A. jella (29 mm), this state is first reported at even higher body lengths (Srinivasachar, 1958a, 1959; Hoedeman, 1960b; Adriaens & Verraes, 1997b; Geerinckx et al., 2005). Rostrally, the taeniae marginalis anteriores (in front of the epiphyseal bridge) are significantly reduced in C. aeneus and the bifurcation into an anteriorly directed commissura sphenoseptalis and a laterally directed commissura spheno- ethmoidalis is situated almost at the level of the epiphyseal bridge itself. A reduction of these anterior parts is indeed common in siluriforms (Adriaens & Verraes, 1997b; Geerinckx et al., 2005), but such a significant reduction or a complete absence of these structures is a condition commonly found in loricarioids like the loricariid A. cf. triradiatus and the callichthyids C. callichthys and H. littorale (Ballantyne, 1930; Hoedeman, 1960b; Geerinckx et al., 2005). 76 4 ONTOGENY – CRANIUM Skull wall As in other catfish, the lamina orbitonasalis is the first preotic vertical commissure to develop in C. aeneus, separating the fenestra sphenoidea from what is later to be the foramen fila olfactoria (Geerinckx et al., 2005). Shortly after this, the preorbital base and commissura sphenoseptalis are formed as additional vertical connections between the taeniae marginales and the trabeculae cranii. At the caudal margin of the orbital region, a commissure splits up the fenestra sphenoidea into an anterior and posterior part. Formerly the absence of a lateral commissure was believed to be a typical siluriform feature, but, recently, Geerinckx et al. (2005) showed a true lateral commissure to be present in A. cf. triradiatus. Given this, and given the fact that, as in A. cf. triradiatus, the nervus oculomotorius (III), trochlearis (IV), trigeminus (V) and the mandibular branch of the nervus facialis (VII) pass through the anterior part of the sphenoid fenestra and that the ramus hyomandibularis of the nervus facialis (VII) passes through the posterior part, I also designated this commissure to be a true lateral commissure. Splanchnocranium In the splanchnocranium, no fusion of the “hyosymplectic-pterygoquadrate plate” with the interhyal or neurocranium is observed. Such a fusion has been suggested by Hoedeman (1960b) for C. callichthys and has since been confirmed by Howes & Teugels (1989) for Corydoras paleatus, at least for the earliest stages. Observations on cleared and stained specimens and serial sections, however, clearly indicate that in C. aeneus, the interhyal is formed as a separate element from the earliest stages on and that the “hyosymplectic- pterygoquadrate plate” never fuses to the neurocranium. The palatine also arises as a single element, a state which is indeed typical for siluriform fish (Adriaens & Verraes, 1997b). Hoedeman (1960b) also mentions the fact that both halves of Meckel’s cartilage in C. callichthys are “anteriorly connected by ligamentous tissue, but do not fuse”. In all C. aeneus specimens examined here, both halves of Meckel’s cartilage remain fused during almost the entire chondrocranial development, only separating near the start of ossification. In A. cf. triradiatus, this situation is remarkably different with both halves of Meckel’s cartilage arising separately, after which they fuse and later separate again at the onset of ossification in the lower jaw. This increase in lower jaw mobility has been argued 77 4 ONTOGENY – CRANIUM to be an important structural innovation in the families Astroblepidae and Loricariidae and thus seems to have its basis early in ontogeny. The situation of the early Meckel’s cartilage in A. triradiatus also differs structurally from that found in C. aeneus, since both these cartilages are directed medially in the former species and rostrally in the latter and in siluriform species in general. All elements of the splanchnocranium in C. aeneus are present from a SL of 4.4 mm on, which corresponds to the point of yolk sac depletion. Since, at that point starvation becomes a major threat for larval survival (Bailey & Houde, 1989), this point is critical during larval development (Pedersen et al., 1990; Jardine & Litvak, 2003). From this stage on exogenous feeding becomes obligatory and the presence of a functional feeding apparatus is required (van Snik et al., 1997; Jardine & Litvak, 2003). It has also been suggested that a fusion between lower jaw and hyosymplectic-pterygoquadrate plate is crucial in the passive mouth opening mechanisms acting during early ontogeny of fishes, but again, such a fusion was not present in any of the stages examined here (Surlemont et al., 1989; Adriaens et al., 2001). Branchial arches The development of the branchial arches in C. aeneus differs from the situation described in siluriform fish up to this point. Generally, in Siluriformes, chondrification in each arch starts with the ceratobranchials, followed by the hypobranchials and pharyngobranchials basibranchials, (Srinivasachar, and 1959; eventually Adriaens & the epi- Verraes, and 1997b; Vandewalle et al., 1997). In C. aeneus, however, I found that all different elements arose simultaneously and continuously with each other. Since I did not notice this initial single formation of the branchial basket in the cleared and stained specimens, these observations are probably the result of the detailed observation method used in this study. In this study, serial sections for 3-D reconstructions of these early branchial baskets were also used, in contrast to only the cleared in stained specimens used in former studies on branchial arch ontogeny of fish. However, Adriaens & Verraes (1997b), also mention that the general ontogenetic sequence in siluriform branchial arches involves a differentiation in an antero-posterior direction. In C. aeneus, differentiation started with the separation of an infrapharyngobranchial plate, which, based on its position and articulation with the epibranchials, probably resembles a 78 4 ONTOGENY – CRANIUM continuous infrapharyngobranchial I-IV. At the same moment of this differentiation, the central copula detaches from all arches and splits up into the typical anterior and posterior copula. Several authors have argued that the anterior part is the fusion of the first three basibranchials, with the posterior copula being a fused basibranchial IV and V (Srinivasachar, 1959; Adriaens & Verraes, 1997b; Vandewalle et al., 1997). At the same time of the formation of both copulae, the hypobranchials of arches II and III separate as well. However, in the specimen of stage 3 (4.4 mm SL) examined here, I found the basibranchial of arches II an III to be separate, while in the stage 4 specimen (4.9 mm SL), the basibranchial was still loosely connected to the ceratobranchial part. At this point, however, the epibranchials of arches II and III have also separated, indicating that both arches differentiate rapidly and almost simultaneously, prior to arches I and IV. After this, the epibranchial of arch I first detaches at stage 5 (5.3 mm SL), shortly after which the epibranchial of arch IV also separates from the ceratobranchial part. At this point, the infrapharyngobranchial plate has also split up into an anterior and posterior part. The last element to detach is the first hypobranchial, which separates at around 8 mm SL. The configuration of the gill arch basket at the end of chondrocranial differentiation does resemble the typical siluriform state. The basket has five arches with an anterior and posterior copula, separate hypobranchials in the first three arches and separate epibranchials in the first four (Adriaens & Verraes, 1997b). Adriaens & Verraes (1997b), however, also suggested the presence of a remnant cartilaginous infrapharyngobranchial I and II, articulating with the epibranchials of these two arches in adult C. gariepinus. In adult C. aeneus such a small cartilaginous mass is also observed, and was placed in homology with a fused infrapharyngobranchial I and II (see 3.1). The ontogenetic sequence observed here, however, suggests otherwise. In stage 4, an anterior projection of the cartilaginous infrapharyngobranchial plate is indeed observed, articulating with the epibranchial parts of these first two arches. In the later stages, however, this projection degenerates, with even a loss of contact between the epibranchial of the first arch and the infrapharyngobranchial plate in stage 6. Therefore, the cartilaginous mass seen in adult C. aeneus can not be homologous to these infrapharyngobranchials and probably resembles the cartilaginous tips of epibranchials I and II, which have come in close contact to each other or have fused. 79 4 ONTOGENY – CRANIUM Conclusions In general, the chondrocranium in C. aeneus follows the typical siluriform trends in chondrocranial development as described by Adriaens & Verraes (1997b). The skull is of the platybasic type, as described for all Siluriformes and the notochord does not protrude into the hypophyseal fenestra due to the presence of the acrochordal cartilage (Adriaens & Verraes, 1997b; Vandewalle et al., 1999). In contrast to the general situation found in catfishes, but similar to the situation described for A. cf. triradiatus by Geerinckx et al. (2005) a commissura lateralis was found. The splanchnocranium again has the typical siluriform shape during its ontogeny, with the presence of a compound hyosymplectic-pterygoquadrate plate present, although not fused to the neurocranium or interhyal as described for C. callichthys by Hoedeman (1960b) and Howes & Teugels (1989). In general, as hypothesized, development of the chondrocranium was very similar to that of A. cf. triradiatus, with one of the main differences found in the development of the lower jaw. This resulted in an increase in jaw mobility in the loricariid species, used for algae scraping. In comparison to all siluriforms, including A. cf. triradiatus, however, the most striking differences found involved the ontogeny of the branchial basket. Whereas, generally, chondrification occurs in a ventrodorsal and antero-posterior direction, in C. aeneus, the branchial basket arises as a single element and differentiation first starts dorsally, followed by arches II and III and then proceeds in both an anterior and posterior direction. Since I did not notice this initial single formation of the branchial basket in the cleared and stained specimens, these observations are probably the result of the detailed observation method used in this study. To find out whether or not this pattern is indeed more common in siluriforms, further histological comparative research on branchial basket development is needed. Discussion of osteocranial development Ossification sequence in teleosts is related to functional demands arising in developing organisms (Adriaens & Verraes, 1998). In this context, the moment of yolk sac depletion is critical, since, from that point on, exogenous feeding becomes obligatory and starvation becomes a major threat for larval survival 80 4 ONTOGENY – CRANIUM (Bailey & Houde, 1989; Pedersen et al., 1990; Jardine & Litvak, 2003). After this transition, larvae are faced with high functional demands on the structures associated with feeding (van Snik et al., 1997; Jardine & Litvak, 2003). In addition, in most teleost larvae the common feeding method is suction feeding, resulting in additional functional demands on the structures necessary to create a negative buccal cavity pressure (Wagemans & Vandewalle, 2001). Therefore, it is expected that development in both the chondro- and osteocranium reflects these demands and that priorities during early cranial ontogeny would be focused on the completion of a functional feeding apparatus, prior to neurocranial fortification. It has also been noted that the first bones that appear during teleostean development always seem to be of dermal origin (Wagemans et al., 1998). Indeed, throughout studies dealing with osteological development in teleosts, onset of ossification generally involves the maxilla, premaxilla, dentary, upper and lower pharyngeal jaws, branchiostegal rays and opercular elements, most of which are of dermal origin and all of which are associated with jaw movement and the allied act of respiration (Weisel, 1967; Kobayakawa, 1992; Tilney & Hecht, 1993; Mabee & Trendler, 1996; Adriaens & Verraes, 1998; Faustino & Power, 2001; Vandewalle et al., 2005; Geerinckx et al., 2007). As expected, in Corydoras aeneus all elements of the cartilaginous splanchnocranium are present at the time of yolk sac depletion (4.4 mm SL). At this point also, the opercle is the first ossified element to appear and teeth are already present on both lower and upper oral and pharyngeal jaws, although not yet erupted and not yet supported by ossified elements. Buccal teeth appearing prior to ossification of their supporting elements is a situation uncommon in teleost fishes, where teeth usually appear at the same time or after premaxillary and dentary bones (Vandewalle et al., 2005). In contrast to this, pharyngeal teeth do generally appear separated from their respective supportive skeletal elements (Vandewalle et al., 2005), as is also the case here. In general, teeth appearing prior to the formation of their supporting ossified elements seems plausible, since it has been argued that bone morphogenesis is possibly influenced by the presence or absence of teeth (Huysseune, 1989). Shortly after yolk depletion, at 4.9 mm SL, the appearance of the parurohyal may indicate a response to mechanical stress resistance due to activity of the sternohyoideus muscle (Adriaens & Verraes, 1998), which implies an increase in lower jaw mobility shortly after this dietary shift. After such a functional and 81 4 ONTOGENY – CRANIUM ossified feeding and breathing apparatus has developed, in between 5 and 8 mm SL, developmental priorities seem to shift to size increase of the cartilaginous skull. After this, cartilage resorption and further ossification begins, accomplishing a rigidification of the skull. Such a fortification becomes necessary since, from the moment of exogenous feeding on, the overlying brain has to be protected from the particles passing below (Adriaens & Verraes, 1998). Next to this, it has also been hypothesized that respiratory movements and buccal pressure would also generate mechanical loadings inducing skull floor ossification (Mabee & Trendler, 1996; Geerinckx et al., 2007). This explains why ossificiations in the skull floor also generally ossify early in the ontogeny of teleost fish (Vandewalle et al., 1995, 1997; Adriaens & Verraes, 1998). In C. aeneus, these ossification indeed occur early, however not directly after the shift from endo- to exogenous feeding, as would be expected. In this case, skull floor bones like the ventral mesethmoid part, the prevomer and parasphenoid only start to ossify from 8 mm SL on, well after this shift in feeding style (which occurs at 4.4 mm SL). At this stage, the prevomer also clearly shows two lateral cornua or palatal splints. The presence of such structures in adult callichthyids has been documented in the past for species of Corydoras and Aspidoras (Reis, 1998), but was not observed in my earlier study on the osteology of adult C. aeneus (see 3.1). Observations now show these splints to fuse to the prevomeral bone, although additional observations on serial sections show them to be still present in some adult C. aeneus-specimens, but, in that case, highly diminished in size. This suggests that the reported absence of such splints in other callichthyid genera like Callichthys, Dianema, Lepthoplosternum, Hoplosternum and Megalechis is probably the result of a secondary loss (through fusion or resorption) during ontogeny. The prevomeral bone in C. aeneus is toothless, in contrast to the general siluriform state. Also unlike other catfish, the bone arises as un unpaired element, but since it has been argued that the prevomeral bone generally consists out of an unpaired dermal toothless bone fused to paired autogenous toothplates, its seems that only the latter are absent in C. aeneus development (Adriaens & Verraes, 1998). At 8 mm SL, the neurocranium also has started to ossify at its dorsal side, since such reinforcements become necessary due to skull growth (Adriaens & Verraes, 1998). Ossification simultaneously starts at both the level of the frontal bones as well as the level of the parieto-supraoccipital bone. The latter bone arises as a single ossification and 82 4 ONTOGENY – CRANIUM a separate parietal and supraoccipital part could not be discerned at any stages. Hoedeman (1960c), in his studies on the callichthyids Callichthys and Hoplosternum also did not find ontogenetic evidence for the compound nature of this bone. However, various authors have observed such a fusion between both the dermal parietal and perichondral supraoccipital in various Siluriphysi and have argued this state to be typical for the group (Bamford, 1948; Lundberg, 1975; Fink & Fink, 1996). Therefore, the bone found here was homologized with this fused parieto-supraoccipital. Another example of such a compound bone arising at this stage is the posttemporo-pterotico-supracleithrum. Again, during ontogeny, no signs of fusion of the separate elements constructing this compound bone were observed, but, nonetheless, the compound nature of this bone was accepted based on the arguments provided under 3.1. From 8 mm SL on, ossification in the splanchnocranium drastically increases, with various centers of ossification in the autopalatine, opercular series, lower jaw, suspensorial and hyoid arch. Also at this point all articulatory facets have started ossifying, rigidifying the entire splanchnocranium structure. In the lower jaw, the mentomeckelian, dentary, articular and retroarticular bones, which will later make up the compound dentary bone (s.l.), are still present as separate elements. As in all callichthyids and various other loricarioids like loricariids, astroblepids and most trichomycterids, no coronomeckelian bone is present (Mo, 1991; de Pinna, 1993; Geerinckx et al., 2007). At this point in the opercular series, a suprapreopercular bone has also developed. The presence of this bone was already mentioned in adult C. aeneus specimens as a condition uncommon for Callichthyidae and it was suggested that further ontogenetic research on the ontogeny of this bone would have to confirm its hypothesized homology (see also 3.1). Given the fact that the data here shows the bone to develop in close contact to the preopercular canal and given its position, the small bone shown in the figures would indeed appear to be a suprapreopercular bone. However, also as shown in the figures, the bone described as the suprapreoperculare in the ontogenetic series fuses to the hyomandibula and the preopercular canal does not protrude into it. Next to this, the bone develops early in ontogeny, while in a related species like A. triradiatus, it arises much later in ontogeny (Geerinckx et al., 2007) and does not resemble the bone described as suprapreopercle in the adult specimens of C. aeneus. In addition, study of additional adult specimens shows the bone to be variably present in these adults. Therefore it is 83 4 ONTOGENY – CRANIUM questionable whether or not the bone described in early ontogeny as the suprapreopercle is homologous to the bone seen in some adult specimens. Nonetheless, since both bones, when present, are in close contact with the preopercular canal and are situated above the preopercular bone, the use of name suprapreopercle seem justifiable in both cases. In contrast to the 8 mm SL-stage, at around 10 mm SL, ossification appears more focused on the neurocranium, with additional ossifications appearing in the skull roof, wall and floor. In the splanchnocranium, however, only the preopercular and metapterygoid bone appear at this stage. The catfish metapterygoid has been described as a fusion of ecto-, ento- and metapterygoid by Howes & Teugels (1989). Other authors, however, reported the ecto- and entopterygoid to be absent, which I believe to be the case in this species as well (Regan, 1911; Arratia, 1990, 1992; Adriaens & Verraes, 1998; Reis, 1998). Also at 10 mm SL, most elements of the branchial basket start to ossify, all of which start in the center of the respective cartilaginous elements. Shortly after this, at 11 mm SL, ceratobranchials and hypobranchials I-II also start to ossify, although, in this case, ossification starts at the rostrolateral side of the cartilaginous hypobranchials. Finally, at 14 mm SL, the basibranchials are the last bony elements to develop. This way, the entire branchial basket, apart from the tooth bearing elements, has completely ossified within the short range of 1014 mm SL and together with the hyoid arch, it is the last large functional unit to complete ossification. Next to this, the interhyal first ossifies at this point. The late emergene of a bony interhyal indicates that the fact that it is lost in several siluriform catfishes like Clarias gariepinus, Loricariidae and Scoloplacidae (Adriaens & Verraes, 1998; Bailey & Baskin, 1976) could be the result of a truncation in development. This could also explain the variability in the number of bones in the infraorbital series, found throughout siluriform phylogeny. Only three additional ossifications occur after this point. At 16 mm SL, the nasal and first infraorbital bone are the last bones to appear in the development of the C. aeneus-cranium. Conclusively, overall ossification sequence in C. aeneus follows the general trends observed in siluriform development, with ossifications appearing as a response to functional demands (Tilney & Hecht 1993; Vandewalle et al., 1995, 1997; Adriaens & Verraes 1998). This way, early development of the skull is 84 4 ONTOGENY – CRANIUM focused on the completion of a functional feeding and respiratory apparatus, prior to rigidification and growth. 85 4 ONTOGENY – CRANIUM 86 4 ONTOGENY – CRANIAL MYOLOGY 4.4 THE CRANIAL MYOLOGY Modified from: Huysentruyt F, Brunain M & Adriaens D Ontogeny of the cranial musculature in Corydoras aeneus (Gill, 1858) (Callichthyidae, Siluriformes) Journal of Fish Biology, submitted. Abstract Very few studies have dealt with the ontogeny of the cranial musculature in teleosts in general and Siluriformes in particular and the ones that have often fail to describe the earliest stages of ontogeny. Nevertheless, the study of these earliest stages could yield valuable information on the evolutionary-anatomical response to functional needs placed on the developing organism; information on the presence of muscular tissue and sites of origin and insertion would be indicative of functionality. To fill this gap, a complete study of the early ontogeny of the cranial muscles of Corydoras aeneus (Callichthyidae) was done. This species belongs to the Loricarioidea, a superfamily in which an evolutionary trend has been observed that has led to the formation of a suckermouth in two lineages (Astroblepidae and Loricariidae) and an additional highly specialized feeding mechanism (i.e. algae-scraping) in the Loricariidae. Results found here were compared to those found by Geerinckx et al. (in press) for Ancistrus cf. triradiatus, a loricariid representative. As expected, this comparison revealed a high degree of similarity in the ontogeny of both species’ cranial muscles. This way, both species lack a m. protractor hyoidei, and the m. intermandibularis posterior is divided into two different parts which have partly obtained a novel function in A. cf. triradiatus. A similar increase in complexity in this species is found in the dorsal constrictor of the hyoid muscle plate. This constrictor gives rise to the same muscles as in C. aeneus, but, in A. cf. triradiatus, the m. levator operculi later hypertrophies. In addition, in A. cf. triradiatus, the m. extensor tentaculi further differentiates into two separate bundles (as opposed to a single muscle diverging posteriorly in C. aeneus) and a loricariid neoformation is present called the m. levator tentaculi (Geerinckx et al., in press). 87 4 ONTOGENY – CRANIAL MYOLOGY Brief introduction In order to detect functional changes as they occur throughout ontogeny, a good starting point of study are the higher levels of organization, linked to these changes (Simonovic et al., 1999). The development of the cartilaginous as well as the bony skull has been the subject of various studies on siluriforms in general (Kindred, 1919; Bamford, 1948; Srinivasachar, 1957a, b, 1958a, b, 1959; Vandewalle et al., 1985, 1997; Tilney & Hecht, 1993; Adriaens & Verraes, 1994, 1997b, 1998; Geerinckx et al., 2005) and callichthyids in particular (Ballantyne, 1930; Hoedeman, 1960a, b, c). However, only in the cases of Clarias gariepinus and Ancistrus cf. triradiatus has the focus of ontogenetic research been extended to the level of the development of the muscular system (Surlemont et al., 1989; Surlemont & Vandewalle, 1991; Adriaens & Verraes, 1996, 1997b, c, d; Geerinckx & Adriaens, 2007, in press; Geerinckx et al., in press). As far as teleosts in general are concerned, a similar lack of studies dealing with the issue in general is found and the ones that have, often fail to describe the very earliest stages of ontogeny. Nevertheless, the study of this system up to the earliest stages of development could yield valuable information on the evolutionaryanatomical response to functional needs placed on the developing organism since, although morphogenesis and differentiation are very intense during early life stages (van Snik et al., 1997), functional constraints of vital importance act on each of those stages. Information on the presence of muscular tissue and on sites of origin and insertion, necessary for performing movements in various functional units, would therefore be indicative of the functionality of these units throughout ontogeny. To fill this gap, the aim of this study is to provide a complete description of the ontogeny of the cranial muscles of Corydoras aeneus, in continuation of previous studies published on the species’ overall development and the ontogeny of the chondrocranium, osteocranium and postcranial skeleton (see 4.3, 4.4, 4.5, 4.7). Additionally, results found here will also be compared to the results found in A. cf. triradiatus by Geerinckx et al. (2005). 88 4 ONTOGENY – CRANIAL MYOLOGY Brief materials and methods The specimens examined are presented in table 2.1. For clearing and staining procedures, procedures on serial sectioning and 3D-reconstructions see 2.2. Terminology on muscles and their development follows Winterbottom (1974) and Jarvik (1980). For a discussion on homology issues and some aspects of the innervation pattern, I refer to 3.2. Results Stage 1: 3.3 mm SL (hatchling) (fig. 4.22) At hatching, the chondrocranium of Corydoras aeneus is still very rudimentary with the neurocranium concentrated around the notochord tip and the splanchnocranium still confined to the hyosymplectic-pterygoquadrate plate (sensu Arratia, 1992). At this point, a small sheet of muscular tissue, the mandibular muscle plate, is present anterior to this early suspensorium. The sheet is broad dorsally and orientated in a dorsoventral direction with a slight antero-ventral inclination in the lower half, narrowing toward its ventral tip. Stage 2: 4.0 mm SL (fig. 4.23) Shortly after hatching, the neurocranium has expanded rostrally and, at the level of the splanchnocranium, the suspensorium has differentiated into a quadrate and hyosymplectic part and the hyoid arch has been formed. The dorsal part of the mandibular muscle plate has differentiated into an anterior musculus levator arcus palatini, which connects the anterodorsal margin of the hyosymplectic to the anterior margin of the otic capsule, and a posterior m. dilatator operculi. Both of these muscles remain in close contact with each other and even share some fibers in adult specimens (see 3.2). The m. dilatator operculi also originates on the otic capsule but its insertion site, the opercle, is still absent at this point in development. Posterior to the suspensorium, the constrictor dorsalis and constrictor ventralis of the hyoid muscle plate are present with the former associated with the otic capsule and lying posteromedial to the hyosymplectic. The ventral constrictor (or m. interhyoideus posterior, see 89 4 ONTOGENY – CRANIAL MYOLOGY discussion), on the other hand, is situated just posterior to the early hyoid. At the posterior margin of the neurocranium the m. protractor pectoralis is present, originating at the ventroposterior side of the otic capsule. Stage 3: 4.4 mm SL (fig. 4.24) At this stage, the neurocranium has further differentiated and its roof has closed anteriorly at the level of the ethmoid cartilage and posteriorly where the otic capsules and neurocranial floor have expanded significantly. In the splanchnocranium, a lower jaw is present and ventral to it, the anterior part of the mandibular muscle plate has differentiated into a m. intermandibularis anterior and posterior. The anterior part is situated and inserts on the anterior margin of the still unpaired cartilaginous lower jaw, while the posterior part originates on both ceratohyals, but does not yet insert on the lower jaw. In the larger middle section of the mandibular muscle plate, the dorsal fibers have differentiated into a distinct m. retractor tentaculi, originating on the anterior side of the suspensorium and inserting on the primordial ligament, near the autopalatine (not shown in fig. 4.24). The remaining part of the mandibular muscle plate forms the still undifferentiated m. adductor mandibulae complex which also originates on the suspensorium but which, next to an insertion on the primordial ligament, also inserts on the lower jaw. The constrictor dorsalis of the hyoid muscle plate has formed a m. adductor arcus palatini, originating on the trabeculae cranii and inserting on the medial side of the supensorium. More anteriorly, fibers of this hyoid muscle plate have differentiated further into a m. extensor tentaculi, which also originates on the trabeculae cranii and inserts on the posterior margin of the autopalatine. In addition, the constrictor dorsalis of the hyoid muscle plate has differentiated into two parts, which both originate on the ventral side of the otic capsule. The anterior part can be discerned as the m. adductor hyomandibulae, and the posterior part as the remaining constrictor dorsalis of the hyoid muscle plate. This part, although it already inserts on the dorsomedial margin of the operculum, is not yet differentiated into a m. adductor and m. levator operculi. Ventrally, the m. sternohyoideus is now present, connecting the hyoid arch with the pectoral girdle. 90 4 ONTOGENY – CRANIAL MYOLOGY Stage 4: 4.9 mm SL (fig. 4.25; 4.26A) At 4.9 mm SL the chondrocranium is further rigidified anterodorsally through the taeniae marginales and all parts of the splanchnocranium have become more differentiated. At the level of the cranial muscles, the m. extensor tentaculi has enlarged and inserts on the solum nasi. The m. retractor tentaculi connects to the primordial ligament (not shown in fig. 4.25) and the m. adductor mandibulae has not yet differentiated into an A2 and A1OST-A3’ part (fig. 4.26A). Further, at the dorsal margin of the opercle the muscle sheet has become fully differentiated and a separate m. levator operculi and m. adductor operculi are now present. Stage 5: 5.3 mm SL (fig. 4.26B; C) At the level of the jaw musculature, the m. adductor mandibulae has now fully differentiated into an A2 and A1OST-A3’ part (fig. 4.26B) and, at this stage, the last cranial muscles to develop are the m. hyohyoideus abductor and adductores (fig. 4.26C). Discussion As mentioned, very few studies have dealt with the ontogeny of the cranial musculature in teleosts in general and Siluriformes in particular and the ones that have, often fail to describe the very earliest stages of ontogeny. However, a detailed overview on the developmental pattern of cranial muscle development in Amia calva was given by Jarvik (1980), based on work done by Allis (1897) and Edgeworth (1928). Given the phylogenetic position of the Amiiformes as a sister group to the Teleostei (Hurley et al., 2007), and given the lack of other studies dealing with the subject, the pattern described by Jarvik (1980) has often been regarded as the plesiomorph teleostean pattern. In this pattern, all of the cranial muscles originate from an ontogenetic primordium of the visceral musculature, with the exception of the eye musculature and the m. sternohyoideus, which originate from somatic musculature (Jarvik, 1980; Adriaens & Verraes, 1996). Of the three parts of visceral muscle primordium (mandibular, hyoid and branchial muscle plates) only the first two are dealt with in this paper on Corydoras 91 4 ONTOGENY – CRANIAL MYOLOGY aeneus. In addition, results are compared to those on Ancistrus cf. triradiatus (Geerinckx & Adriaens, in press; Geerinckx et al., in press). Mandibular muscle plate The fact that, at hatching in C. aeneus, the small sheet of muscular tissue present is situated anterior to the suspensorium, suggests this muscle sheet to be the early mandibular muscle plate. This is confirmed by the fact that, at 4.9 mm SL, the different muscular units developing from this plate are all innervated by parts of the mandibular trunk of the trigeminal nerve (V), a discriminative character of the mandibular muscle plate (Edgeworth, 1935; Jarvik, 1980). Shortly after hatching, at 4.0 mm SL, a dorsal section splits off, forming the m. dilatator operculi and m. levator arcus palatini. This is in accordance with the constrictor dorsalis of the mandibular muscle plate as described by Winterbottom (1974) and Jarvik (1980). In A. calva, the remaining part of the mandibular plate further gives rise to the m. palatomandibularis, a compound m. adductor mandibulae and a ventral m. intermandibularis (Jarvik, 1980). Afterwards, in A. calva, the m. palatomandibularis forms the m. nasalis, preorbitalis and parabasalis, all of which are absent in teleosts, with the exception of a preorbital muscle described in two species of cobitids (Takahasi, 1925; Winterbottom, 1974; Jarvik, 1980). Takahasi (1925) has further proposed that at least part of this m. palatomandibularis has been incorporated into the A1 part of the adductor mandibulae complex in teleosts. Given this, I believe a primitive m. palatomandibularis to be absent in teleosts and have designated the ventral part of the muscle plate at this point as constrictor ventralis. At 4.4 mm SL, this constrictor ventralis has further differentiated into a m. intermandibularis anterior and posterior (Winterbottom, 1974; Edgeworth, 1935). The latter muscle is generally believed to have fused to the m. interhyoideus anterior forming the compound m. protractor hyoidei in teleosts, based on an observed double innervation of the muscle by both the mandibular trunk of the trigeminal nerve (V) and the hyoid trunk of the facial nerve (VII) (Winterbottom, 1974). However, since serial sections revealed this muscle in C. aeneus to be innervated only by the inferior mandibular nerve branch of the trigeminal nerve (V) and since no myocomma was present, it, in this case, merely consists of the m. intermandibularis posterior and thus no m. protractor hyoidei part is present (see 3.2). This is also the case in A. cf. triradiatus and possibly in all Siluriformes 92 4 ONTOGENY – CRANIAL MYOLOGY (Geerinckx & Adriaens, 2007). The m. intermandibularis posterior of other adult non-loricarioid siluriforms generally has a secondary subdivision, with some bundles serving the mandibular barbel base (Diogo & Chardon, 2000). In adult C. aeneus, such a subdivision is also present, although no insertion on the mandibular barbels is found (see 3.2). In A. cf. triradiatus, these subdivisions are also found, but, given the rotation of the lower jaw in A. cf. triradiatus and the subsequent orientation of these bundles in combination with the absence of mandibular barbels in loricariids, Geerinckx & Adriaens (2007) have termed them the labial and dental parts. Nonetheless, is seems plausible that they are homologous to the subdivisions found in C. aeneus, which have partly obtained a novel function in the Loricariidae. At 4.4 mm SL, the remaining part of the adductor division of the mandibular muscle plate differentiates into a dorsal m. retractor tentaculi, inserting only on the primordial ligament, and a more ventral compound m. adductor mandibulae. At 4.9 mm SL, the latter splits up further into a dorsomedial A2 part, which also inserts on the primordial ligament and a compound A1OST-A3’, diverging anteriorly into an A1OST bundle inserting on the dorsolateral side of the lower jaw and an A3’ bundle inserting on the dorsomedial side. The fact that the m. retractor tentaculi is also derived from the adductor division of the mandibular plate, adds evidence to the hypothesized homology of this muscle to part of the A3”, as already suggested by various authors (Lubosch, 1938; Alexander, 1965; Howes, 1983; Adriaens & Verraes 1996, 1997a; Diogo & Chardon, 2000; Diogo, 2005). In A. cf. triradiatus, the adductor mandibulae complex consists of the same bundles, with the addition of a m. retractor veli (Geerinckx et al., subm.). The bundle homologous to the m. retractor tentaculi, on the other hand, is present but its insertion point has shifted from the primordial ligament directly onto the premaxilla. Given this, and since in various studies on loricariids the term m. retractor tentaculi was already mistakenly used for what Geerinckx et al. (subm.) called the m. levator tentaculi (Howes, 1983; Schaefer & Lauder, 1986, 1996; Diogo & Vandewalle, 2003), Geerinckx et al. (subm.) proposed the term m. retractor premaxillae for this muscle. 93 4 ONTOGENY – CRANIAL MYOLOGY Hyoid muscle plate In C. aeneus, the hyoid muscle plate starts to develop shortly after the appearance of the mandibular muscle plate. In the specimens examined, at its time of first appearance, the plate is already split up into a constrictor dorsalis and constrictor ventralis. In A. calva, the constrictor ventralis is described to differentiate into a m. interhyoideus anterior and m. interhyoideus posterior in later development (Jarvik, 1980). Since the m. interhyoideus anterior is then believed to further differentiate into part of the m. protractor hyoidei (Edgeworth, 1935; Winterbottom, 1974), which is absent in C. aeneus (see 3.2), the ventral part present here can be homologized with the m. interhyoideus posterior of A. calva. This is also confirmed by the fact that this muscle plate further differentiates into the m. hyohyoideus inferior, the m. hyohyoideus abductor and the mm. hyhyoidei adductores, all of which are believed to differentiate from the m. interhyoideus posterior (Winterbottom, 1974, Jarvik, 1980). Jarvik (1980) further describes the constrictor dorsalis to differentiate into a single anterior muscle, the m. adductor hyomandibulae and two posterior muscles, the m. adductor operculi and m. levator operculi. At 4.4 mm SL, however, only two muscle bundles were found. Given its position and given the fact that the anterior part inserts on the medial side of the hyosymplectic, it was homologized with the m. adductor hyomandibulae. This way, the posterior part was still homologous to the posterior portion of the constrictor dorsalis, which had not yet differentiated into a m. adductor operculi and m. levator operculi. In addition, the anterior part of this muscle plate is described to further differentiate, giving rise to the m. adductor arcus palatini, which, in turn, anteriorly gives rise to m. extensor tentaculi (Winterbottom, 1974; Diogo & Vandewalle, 2003). It has also been argued that the m. adductor arcus palatini has evolved from a shifted position of the m. adductor hyomandibulae (Winterbottom, 1974). In several teleosts, however, both a m. adductor arcus palatini as well as a m. adductor hyomandibulae are frequently found (Winterbottom, 1974), as is the case here in C. aeneus. Given this, Winterbottom (1974) has argued that in these cases, the m. adductor hyomandibulae is a secondary derivative of the anterior fibers of the m. adductor operculi or of the posterior fibers of the m. adductor arcus palatini. Diogo & Vandewalle (2003), in their overview on siluriform cranial muscles, accept the former possibility, but provide no arguments for this choice. The data on C. aeneus now shows that at 94 4 ONTOGENY – CRANIAL MYOLOGY 4.4 mm SL, the dorsal constrictor is divided into a m. extensor tentaculi, a m. adductor arcus palatini, a m. adductor hyomandibulae and a dorsal muscle associated with the opercle, which later differentiates into a m. levator operculi and m. adductor operculi part. Therefore, I reject the possibility of the m. adductor hyomandibulae being a secondary derivate of the m. adductor operculi in this case, since the former differentiates earlier in ontogeny than the latter. The other option, where the m. adductor hyomandibulae is derived from the posterior part of the m. adductor arcus palatini would therefore seem more plausible here. In comparison to A. cf. triradiatus, a similar pattern of development in the hyoid muscle plate is found, again with an increase in complexity, as is the case in the mandibular muscle plate. This way, the dorsal constrictor gives rise to the same muscles as in C. aeneus. The m. levator operculi, in A. cf. triradiatus, however, later hypertrophies, resulting in the formation of large myodomes in the neurcranial roof (Geerinckx & Adriaens, in press). Additionally, in A. cf. triradiatus and some other loricariids, the m. extensor tentaculi is differentiated into two separate bundles (as opposed to a single muscle diverging posteriorly in C. aeneus) and a loricariid neoformation is present called the m. levator tentaculi (Diogo, 2005; Geerinckx et al., in press). This fits the evolutionary trend of an increased differentiation of the constrictor dorsalis of the hyoid muscle plate in teleosts, in which anterior fibers of the m. adductor arcus palatini give rise to new muscles in various groups. While, in this case, the m. extensor tentaculi arises, later splitting up into two separate bundles with a subsequent differentiation of the m. levator tentaculi in Loricariidae, other examples exist in teleosts. These examples comprise the formation of a m. retractor arcus palatini in Acanthuridae, Balistidae and Ostraciidae and of a m. retractor palatini in Balistes, both of which are also anterior derivations of the m. adductor arcus palatini. The ontogeny of the m. adductor hyomandibulae, however, differs substantially from the situation found in C. aeneus. In A. cf. triradiatus, this muscle appears in a later stage of ontogeny than the m. adductor arcus palatini and m. adductor operculi. In addition, it appears in close contact with the latter muscle (Geerinckx et al., in press), which, in this case, would confirm the hypothesis of Winterbottom (1974) and Diogo & Vandewalle (2003), where it is a secondary derivative of the m. adductor operculi. This would, however, imply a non- 95 4 ONTOGENY – CRANIAL MYOLOGY homology with the m. adductor hyomandibulae, as described in more primitive species like A. calva. Given the results found in both C. aeneus and A. cf. triradiatus, and given the fact that, in more primitive lineages like A. calva, a m. adductor hyomandibulae and no m. adductor operculi are present, it would seem that indeed multiple pathways toward the formation of a m. adductor hyomandibulae exist in teleosts. However, given the function of this muscle and its consistent origin on the otic region of the neurocranium and insertion on the dorsomedial side of the suspensorium, the use of the name m. adductor hyomandibulae seems highly justifiable throughout the different actinopterygian groups, as long as a true homology is unknown. 96 4 ONTOGENY – POSTCRANIAL SKELETON 4.5 THE POSTCRANIAL SKELETON Abstract Most ontogenetic research on teleosts focuses on the cranium. A detailed study of the ontogeny of the postcranial skeleton would, however, help understanding function in the developing larva since development of the body axis provides firmness, influenced by increasing forces acting on it during ontogeny, and since fin development also relates to changes in swimming mode, swimming velocity and feeding techniques and preferences (Fukuhara, 1992; Koumoundouros et al., 1999; Koumoundouros et al., 2001b). Therefore, such a study was performed on Corydoras aeneus, and results were compared to some of the data presented in 4.2. Indeed results confirmed that swimming mode in this species changed at two points during ontogeny. Observations on the ontogeny of the caudal fin skeleton showed that all cartilaginous elements in this fin had developed between a SL of ca. 5–6 mm, at which point also notochord flexion occurred. This corresponds to the suggested change to burst swimming (rapid generation of trust propelling the animal away) at 5 mm SL (Weihs, 1980; Verhagen, 2004)(see also 4.2). In the analysis of early development a second shift in swimming mode from anguilliform (with trust generated by the entire body resulting in high undulatory amplitudes along the body surface) to subcarangiform swimming (with trust being generated mostly posteriorly in the body and low amplitudes of anterior body undulations) was believed to be present at around 8 mm SL (see 4.2). Indeed, at 8 mm SL, the anal, adipose and pelvic fins were found to separate from the larval finfold and ossification in the vertebral column and caudal skeleton had started. Also, development of the anal and pelvic fin had started and ossification of the scapulocoracoid plate and development of the proximal radials and fin spine in the pectoral fin also took place. Brief introduction In C. aeneus, general early development as well as the ontogeny of the chondrocranium and osteocranium (see 4.2; 4.3; 4.4) has been studied, but no data exist on postcranial ontogeny. Such data would, however, help understanding function in the developing larva since development of the body axis provides firmness, influenced by increasing forces acting on it during 97 4 ONTOGENY – POSTCRANIAL SKELETON ontogeny, and since fin development also relates to changes in swimming mode, swimming velocity and feeding techniques and preferences (Fukuhara, 1992; Koumoundouros et al., 1999; Koumoundouros et al., 2001b). In addition, ontogenetic data could also help solve some homology issues. This way, e.g. in the caudal skeleton of adult specimens, the dorsal plate is believed to be a fusion of the urostyl and the dorsal hypurals III, IV and V while the ventral plate would comprise the parhypural and hypurals I and II (Lundberg & Baskin, 1969) (see also 3.1). In the case of the Callichthyidae, however, little ontogenetic data support this. Therefore, the objective of this study is to describe the normal pattern of postcranial development in C. aeneus and link it to possible changes in function of the locomotory and feeding apparatus as well as to determine the homology of the different skeletal elements involved. In this study, the development of the Weberian apparatus was not included, since a thorough description of this was already given by Coburn & Grubach (1998) for C. paleatus. Brief materials and methods The specimens examined are presented in table 2.1. For clearing and staining procedures, see 2.2. Bone terminology used throughout this paper follows Arratia (2003b), Lundberg & Baskin (1969) and Grandel & Schulte-Merker (1998). Results Axial skeleton The vertebral column (fig. 4.27) Development in the vertebral column of Corydoras aeneus takes place over a narrow range in standard length. The first signs of vertebral development were present from a SL of 5.2 mm on, but until 6.2 mm SL, the vertebral column shows no signs of ossification. Shortly after, at 7.2-7.7 mm SL (see also fig. 4.28F), the first ossifications appear and by 8.2 mm SL, all vertebrae have fully 98 4 ONTOGENY – POSTCRANIAL SKELETON developed and all neural and haemal processes have ossified. At this point, the parapophyses of the sixth vertebral centre already support a large hollow rib, which will later contact the first ventrolateral bony scutes behind the pectoral girdle. Further, the remaining precaudal vertebrae (7-12) each bear a thin rib at this point (not shown in figure 4.27). Behind these, the haemal spines of vertebrae 13-17 are gradually longer in a posterior direction, making space for the abdominal cavity. In addition to these five caudal vertebrae, nine more are present and an additional one is incorporated in the caudal skeleton, which totals the number of vertebrae to 28 (see also 3.1). The caudal skeleton (figs. 4.27; 4.28) The caudal skeleton is of the pleurostyl type and consists of two bony plates. The ontogenetic series studied here revealed the fusion of the urostyl to the dorsal hypurals, as well as the development of a plate-like lamina on the epural (the neural spine of the first preural centre). An actual fusion of the hypurals, however, was not detected. At 4.8 mm SL a ventral hypural cartilage is present, soon after which, at 5.2 mm SL, the dorsal hypural cartilage appears. At 5.4 mm SL, notochord flexion starts and the first five principal fin rays appear. Of these, four articulate with the dorsal and one with the ventral hypural plate. At 5.7 mm SL, the cartilaginous haemal arch of the second preural centre is present and three additional principal fin rays appear that articulate with the ventral hypural plate. In addition, ventrally, the parhypural cartilage also appears and is already fused to the ventral hypural plate. At 6.2 mm SL, six principal rays and one procurrent ray are present ventrally and the epural cartilage has appeared dorsally. Shortly after this, at 7.2 mm SL, all principal rays are present together with two procurrent rays on both the dorsal and ventral side. At this stage, ossification of the vertebrae has started, clearly showing ossification centres on the anterior and posterior margin of the second preural vertebra and in its neural and haemal spines. The first preural vertebral centre, which will later fuse to the tail skeleton, only shows ossification at its anterior margin. After this, at 7.7 mm SL, the haemal spine of the second preural centre also ossifies and at 8.2 mm SL almost all parts of the caudal skeleton become completely ossified. After this stage, the epural further ossifies and both hypural plates partially fuse followed by the second preural haemal arch, which is the last part to ossify. 99 4 ONTOGENY – POSTCRANIAL SKELETON The dorsal fin (figs. 4.27; 4.29) Ossification in the dorsal fin starts around 7.7 mm SL in C. aeneus, with the ossification of the first two fin rays. Before this, from a SL of 5.7 mm on, eight cartilaginous pterygiophores are present supporting seven actinotrichia. At 8.2 mm SL, an additional double actinotrichium appears at the level of the 8th pterygiophore. During ontogeny, it is clear that both the second fin spine and the first fin ray appear in relation to the second pterygiophore. Finally, in adult specimens of C. aeneus, the second dorsal fin spine will also develop a lateral process through which it articulates with the transverse process on the second pterygiophore. Further in ontogeny, pterygiophores II-VI all develop a proximal and distal part. In the second pterygiophore, articulation with the second fin spine occurs at the level of the proximal part, while the first fin ray articulates with the cartilaginous distal part. In all other fin rays, articulation occurs only through the distal part. In addition, the first pterygiophore also fuses to the nuchal plate forming a large rigid bone proceeding the dorsal fin. Therefore, this bone was not recognized as a pterygiophore and the transverse process was mistakenly reported to be present on the first pterygiophore in my description of the species’ adult osteology (see 3.1). The anal fin (figs. 4.27; 4.30) At 7.7 mm SL, the first three proximal pterygiophores of the anal fin become apparent, shortly after which, at 8.2 mm SL, all six pterygiophores are present. Ossification in the fin rays starts around 11 mm SL, at which time the onset of ossification in the first pterygiophore can also be observed and the last, double fin ray first appears. The first two fin rays both develop in relation to the first pterygiophore and the distal parts of all pterygiophores remain unossified. Appendicular skeleton The pectoral girdle (fig. 4.31) At 6.2 mm SL, a thin cleithral bone is present, caudally bordered by an endoskeletal cartilaginous disc which articulates with five actinotrichia. At this point, the anterior dorsoventrally orientated cartilaginous rod, corresponding to the scapulocoracoid cartilage, is still connected to this rostro-caudally orientated endoskeletal disc. At 6.6 mm SL, the fin plate splits into three lobes, and three cartilaginous distal radials individualize at this point. The upper lobe disappears 100 4 ONTOGENY – POSTCRANIAL SKELETON during further ontogeny (or fuses with the scapulo-coracoid cartilage), while the lower two ossify as proximal radials. From 8 mm SL on, the first fin ray, later forming the pectoral fin spine, starts to ossify, while the remaining fin rays ossify much later during ontogeny (after 12.5 mm SL). Also around 8 mm SL, the dorsal part of the pectoral girdle, the posttemporo-pterotico-supracleithrum, is incorporated into the neurocranium and also starts to ossify (see 4.4; fig. 4.15A). At 9.7 mm SL, the scapulocoracoid cartilage ossifies, with both left and right bones extending much further ventrally than the cartilage itself and contacting each other at the midline. After this stage, all bones further ossify, and at 12.5 mm SL, both scapulocoracoid and cleithrum have developed a median process, contacting each other. At this stage, the first serrations on the pectoral fin spine appear and eight additional ossified fin rays are present. Pelvic girdle (fig. 4.32) The first signs of a pelvic girdle are observed in specimens with a SL of around 8.2 mm, in which two cartilaginous basipterygia articulate with six actinotrichia, already approaching the definite number of pelvic fin rays. These rays become ossified at 9.7 mm SL. The cartilaginous basipterygium further develops in an anterior direction, forming the cartilaginous precursor anterior external and internal processes, which later fuse. Around 16 mm SL, ossification starts, through which a lateral process and dorsal and ventral ischiac processes develop. Discussion Data on postcranial siluriform morphology is mostly dedicated to the Weberian apparatus (for a review, see Chardon et al., 2002), although some recent data is present on the pectoral girdle (Diogo et al., 2001) and caudal skeleton (Lundberg & Baskin, 1969; Arratia, 1983; Lakshmi & Srinivasa Rao, 1989; Fujita, 1992; Arratia, 2003b; de Pinna & Ng, 2004; De Schepper et al., 2004). In the specific case of loricarioids, a detailed description on the pectoral girdle of Nematogenys inermis has been provided by Diogo et al. (2006) and the postcranial osteology of Corydoras aeneus is dealt with under 3.1. In addition, some literature on the ontogeny and ossification of loricarioids and even callichthyids exists, with a study on the Weberian apparatus in Corydoras paleatus by Coburn & Grubach 101 4 ONTOGENY – POSTCRANIAL SKELETON (1998), and a description of the postcranial ontogeny in Trichomycterus areolatus and Diplomystes nahuelbutaensis (Arratia, 2003b) and Callichthys callichthys and Megalechis thoracata (Hoedeman, 1960d). Ossification of the vertebral column in C. aeneus occurs much faster than has been described for some marine teleosts like Hippoglossus hippoglossus (Lewis & Lall, 2006), Diplodus sargus (Koumoundouros et al., 2001b), Anarhichas lupus (Pavlov & Moksness, 1997), Dentex dentex (Koumoundouros et al., 1999) and Pagellus erythrinus (Sfakianakis et al., 2004). In the freshwater siluriform Ictalurus punctatus (Arratia, 2003b), this development also occurs slower than in C. aeneus but Arratia (2003b) does mention a higher speed of development in the more closely related Trichomycterus areolatus. While the primitive teleost condition is to have two separate ural centra posterior to the centrum that bears the last haemal arch (i.e. first preural centrum), in catfishes and many other advanced groups of fishes it is generally believed that both the first and second ural centra are co-ossified with the first preural centrum (Lundberg & Baskin, 1969). Next to the evidence found in primitive teleostean lineages, this hypothesis is partially confirmed for siluriforms by the discovery of a separate second ural centrum in certain siluriform groups (Lundberg & Baskin, 1969; De Pinna & Ng, 2004). However, throughout the ontogeny of C. aeneus, in none of the stages examined, separate ural centra were found. This may indeed imply that fusion of these elements already occurs at the very early stages of development, as already suggested by Lundberg & Baskin (1969). In the same study, Lundberg & Baskin (1969) described the dorsal plate of C. aeneus as a fusion of the urostyl and hypurals III, IV and V and the ventral plate as the fusion of parhypural and hypurals I and II. Although the data show no separate hypurals at any point during ontogeny, the fusions between the cartilaginous parhypural and the ventral plate, and between the ossifying urostyl and dorsal plate were indeed clearly observed. The fact that not all hypurals develop as individual elements is a state which has only occasionaly been reported in teleosts (e.g. Pavlov & Moksness, 1997). In catfishes, however, not much data is present on the ontogeny of the caudal fin skeleton. Fujita (1992), however, in his paper on caudal fin ontogeny in Clarias batrachus also reports a first hypural that is fused to the second hypural from the moment it appears. Given the high prevalence of hypural fusions in catfish (Lundberg & 102 4 ONTOGENY – POSTCRANIAL SKELETON Baskin, 1969), this could indicate a different ontogenetic pattern in Siluriformes, but additional data is required to confirm this. Literature on the ontogeny of the pectoral girdle in Siluriformes is limited to a single publication by Arratia (2003b) based on an ontogenetic series of Ictalurus punctatus; the results found are highly comparable to those of C. aeneus. Both species exhibit a chondral endoskeletal disc that splits up distally, with two proximal radials ossifying from this during later ontogeny. Again, in the case of the pelvic girdle, literature is limited to the studies done by Arratia (2003b) on Ictalurus punctatus, Diplomystes nahuelbutaensis and Trichomycterus areolatus. These data clearly show that the presence of the external and internal anterior process is determined early in ontogeny and visible on the cartilaginous basipterygium (Arratia, 2003b). This implies that the two cartilaginous processes visible in the early ontogeny of C. aeneus represent both the internal and external anterior processes, which fuse during later ontogeny. As a result, the external anterior process, as described under 3.1, actually represents a lateral process. The homology of these anterior processes has already been questioned by Shelden (1937), and as a result both were referred to as ‘projections’. However, no obvious motivation was given to support this idea, and therefore both were considered as the internal and external process under 3.1. Evidence now thus shows the external process to be a true lateral process, which might very well also be the case for Callichthys callichthys as described by Reis (1998). Posteriorly, in callichthyids, two bony ischiac processes are present which also show no chondral precursor in early ontogenetic stages. According to Arratia (2003b), these processes, together with the lateral processes, are therefore not homologous to the anterior and ischiac processes, and should be considered as mere projections of the bony surface of the basipterygial plate. The fact that the situation found in the callichthyid pelvic girdle differs from that found in other siluriforms has already been argued to relate to reproductive strategies (Reis, 1998), since during spawning female callichthyids carry their eggs in a ‘shell-shaped’ pouch made by the pelvic fins (Kohda et al., 1995). The study of the changes in growth and allometries during ontogeny (see 4.2) seemed to confirm that around ca. 5 mm SL a shift in swimming mode toward a burst and coast swimming technique was present. Both Weihs (1980) and Verhagen (2004) indeed found that from this SL on, burst swimming became the 103 4 ONTOGENY – POSTCRANIAL SKELETON most effective swimming mode in fish larvae. Given that propulsive forces in this swimming mode are obtained by tail beats, development in the caudal fin was expected to be at a peak at this time. At this point in development (up to 6.2 mm SL), growth in the caudal fin indeed was at a maximum reaching a growth coefficient of 4.72 (see 4.2; fig. 4.6D). Observations on the ontogeny of the caudal fin skeleton now also show that all cartilaginous elements in this fin develop between a SL of ca. 5–6 mm, at which point also notochord flexion occurs. After this, between 6.2 and 7.2 mm SL, no major changes in the development of the caudal fin skeleton occurred. In the analysis of early development a second shift in swimming mode from anguilliform to subcarangiform swimming was believed to be present at around 8 mm SL. Indeed, at this point, anal, adipose and pelvic fins were found to separate from the larval finfold and abdominal and postanal growth increased at around this SL (see 4.2). The results here again confirm this, with ossification starting in the vertebral column and caudal skeleton in between 7 and 8 mm SL. In addition, development of the anal and pelvic fin starts around 8 mm SL and ossification of the scapulocoracoid plate and development of the proximal radials and fin spine in the pectoral fin also concentrate around 8–10 mm SL. In conclusion, all elements presented here seem to confirm that during ontogeny in C. aeneus, two separate shifts in swimming style occur, making the presence of a functional locomotory apparatus crucial at those points in development. As suggested by Weihs (1980) and Verhagen (2004), these shifts are caused by the differences in the physical properties of a longer larval body in an aqueous environment and therefore relate to developmental state. 104 the absolute length of the larvae rather than its 4 ONTOGENY – MINIATURIZATION 4.6 ONTOGENY OF CORYDORAS PYGMAEUS, A MINIATURIZED CONGENERIC Abstract The ontogenetic pattern of a miniaturized Corydoras species, Corydoras pygmaeus, was studied and results on growth, external morphology and development of the chondrocranium, osteocranium and postcranial skeleton were compared to comparable data on C. aeneus, a non-miniaturized Corydoras species. Given the low number of miniaturized Corydoras species, miniaturization was believed to have been secondary derived in this genus through the occurrence of paedomorphic events. Results showed high similarities in the overall pattern of development in both species with some difference in the timing of onset of ossification of different elements. In conclusion, miniaturization in the body of C. pygmaeus was indeed found to be the result of paedomorphic events, including a lower growth rate and subsequent postdisplacement of several traits. These paedomorphic events have, however, besides the smaller size, not led to the retention of a large number of paedomorphic traits. Introduction The highly diverse catfish genus Corydoras comprises 152 species with species ranging in maximal standard length (SL) between 20 mm (Corydoras habrosus) and 88 mm (C. britskii) (Reis, 2003; Froese & Pauly, 2006; Ferraris, 2007). The smallest species, C. habrosus, C. boehlkei, C. gracilis, C. hastatus and C. pygmaeus even qualify as miniaturized, since in freshwater teleostean fishes, a maximal SL of 25-26 mm is commonly set as the threshold for miniature species (Hanken & Wake, 1993; Weitzman & Vari, 1988). In studies by Reis (1998) and Britto (2003), morphological data provided a phylogenetic framework for both the family (Callichthyidae), and subfamily (Corydoradinae) to which this genus belongs and in a study by Shimabukuro-Dias et al. (2004) molecular data provided this for the Callichthyidae. The phylogeny of the genus itself, however, remains largely unresolved. This way, the dwarf species C. hastatus and C. 105 4 ONTOGENY – MINIATURIZATION pygmaeus were first described as belonging to the separate genus Microcorydoras by Myers (1953), which was later put into synonymy with Corydoras. In 1970, Nijssen also suggested a close relationship between C. hastatus and C. pygmaeus based on their minute size and suggested the resurrection of the Microcorydoras genus. In their revision of the genus in 1980 however, Nijssen & Isbrücker rejected this hypothesis and put both these dwarf species together with C. gracilis and other non-pygmy species into the “elegans”group of the Corydoras-genus. In this revision, the other dwarf species, C. habrosus, was even placed into an entirely different group, the “aeneus”-group. In 1985, Strauss performed a morphometric analysis of some representatives of the genus and he also found that there seemed to be: “no justification on morphometric grounds for isolating the dwarf taxa (C. hastatus, C. australe and C. pygmaeus) into a separate genus”. He found that C. hastatus and C. pygmaeus lie well within the normal patterns of variation for the genus but did mention that C. australe differed considerably from all other taxa in sizeindependent form (Strauss, 1985). The latter statement thus questioned the earlier hypothesis of Nijssen & Isbrücker (1980), in which they synonymized C. australe with C. hastatus, a synonymy which since has been generally validated (Eschmeyer, 1998). However, Schaefer et al. (1989) again proposed to put the two dwarf species C. hastatus and C. pygmaeus into one monophyletic group based on their minute size, color pattern and reduced dorsal fin ray number (six), but they were unable to retain any additional synapomorphies (besides small size) uniting all miniature Corydoras species. This led them to conclude that they were: “unable to reject the hypothesis that miniaturization arose more than once in the genus” (Schaefer et al., 1989). Finally, in the most recent study done on this group which included dwarf species, Britto (2003) placed dwarf species like C. habrosus and C. gracilis throughout the family’s phylogeny. They did, however, find four synapomorphies to place both C. pygmaeus and C. hastatus into one monophyletic group. Of these four characters, however, two characters are based on pigmentation and one on ethology. This leaves the reduction of the suture between metapterygoid and hyomandibula as a single morphological synapomorphy to group both dwarf species. This confirms the hypothesis of Schaefer et al. (1989) that C. pygmaeus and C. hastatus are indeed sister species, but it also confirms their assumption that miniaturization arose more than once in the genus. In this context, heterochronic events are 106 4 ONTOGENY – MINIATURIZATION often suggested as key factors in the evolution of miniaturized species. Also, of the few synapomorphies indicating a sister-species relationship between C. pygmaeus and C. hastatus (next to body size itself) the two main characters (reduced dorsal fin ray number (Schaefer et al., 1989) and reduction of the suture between metapterygoid and hyomandibula (Britto, 2003)) are both reductive characters, which also indicates the importance of heterochronic events in this miniaturization process. Given this and given the fact that miniaturization has often been postulated as a key feature in the evolution of high level taxa (Hanken, 1993; Hanken & Wake, 1993), the question whether miniaturization within the Callichthyidae is also the result of heterochronic events and the extent of their role in callichthyid evolution becomes apparent. The adult morphology of the family has been studied thoroughly in a phylogenetic context (Reis, 1998) and an extensive study on the adult osteology and myology of C. aeneus has been done recently (see chapters 3.1, 3.2), but very little is known about the ontogeny of the genus’ representatives. In chapters 4.2, 4.3, 4.4, 4.5 and 4.6, the ontogeny of the cranial muscles and skeleton of C. aeneus is dealt with in detail, but ontogenetic data on a miniaturized congeneric is still lacking. Therefore, the aim of this study is to describe the early ontogenetic pattern of a miniaturized Corydoras species, C. pygmaeus, and compare this pattern to the pattern found in C. aeneus. Doing this, I hypothesize that the miniaturized body in C. pygmaeus is the result of paedomorphic processes. In addition, a comparison of developmental sequences of related taxa within a phylogenetic framework, would serve the purpose of identifying such possible heterochronic events (Mabee & Trendler, 1996). Although the phylogenetic framework of the Corydoras genus, as mentioned, is lacking, it seems plausible to assume that the non-miniaturized form is the plesiomorphic condition of the genus. This assumption is based on the fact that only five of the 152 species of Corydoras and only one of the 24 species in the closely related genera Aspidoras and Scleromystax are miniaturized (Reis, 2003; Ferraris, 2007). In a first stage of this research, the focus was limited to an analysis of overall growth, external morphology and of the skeletal cranial and postcranial elements. This was done through the study of an ontogenetic series of cleared and stained specimens of C. pygmaeus. 107 4 ONTOGENY – MINIATURIZATION Brief materials and methods The specimens examined are presented in table 2.2. For clearing and staining procedures, see 2.2. Bone terminology used throughout this paper follows Arratia 2003b), Lundberg & Baskin (1969) and Grandel & Schulte-Merker (1998). For details on the analysis of growth, I refer to 4.2. Results Growth In the growth curve of Corydoras pygmaeus, ranging from 0 until 50 dph, only a single inflexion point was calculated where growth rate (slope) changed (fig. 4.33A). This inflexion is located at 5 dph (which corresponds to a fitted value of 5.17 mm SL), at which the estimated slope increases from 0.12 to 0.37 (p<0.001). External morphology At hatching, C. pygmaeus larvae have an average SL of 3.7 mm and a yolk sac remained present up to 4.4 mm SL. During this eleutherembryonic phase (fig. 4.34A) (for the terminology of the different phases I refer to Balon (1975); see also 4.2), pigmentation was limited to small patches in the dorsal finfold and on the dorsal and lateral sides of the head. At 4.4 mm SL, the anus, mouth and opercular cavity open and the protopterygiolarval phase starts (fig. 4.34B). At this point, body pigmentation in the dorsal finfold has expanded, while pigmentation on the head is present as a double line running from the buccal area, over the eye, to the back of the head. The median finfold further remains undifferentiated until 6.2 mm SL, at which time the pterygiolarval phase starts (fig. 4.34C). This pterygiolarval phase is characterized by a substantial expansion of the body pigmentation and the onset of fin ray formation in all fins. At 8.9 mm SL all fins have fully detached from the finfold and the juvenile period starts (fig. 4.34D). At this point, pigmentation on the body has already began to concentrate into a median lateral line expanding into the basis of the caudal fin, as is the case in adult C. pygmaeus. 108 4 ONTOGENY – MINIATURIZATION Skeleton Stage 1: 3.7 mm SL (hatching) (Fig. 4.35A; B; 4.40A) At hatching in C. pygmaeus, the polar and otic cartilages are already interconnected through the lateral commissure and the basiotic laminae contact the otic cartilages through the anterior basicapsular commissure. At the caudal edge of the neurocranium floor, the parachordal cartilages connect to the otic cartilage through the occipital pilae, leaving two large fenestrae metotica. Both trabecular bars are present, running forward, merging at the level of the ethmoid cartilage, bordering the fenestra hypophysea. Dorsally, both taeniae marginales posteriores are present and interconnect through the epiphyseal bridge, but do not yet contact the ethmoid cartilage. Dorsocaudally, both sides of the neurocranium are extended mediocaudally, forming the onset of the tectum posterius. Most parts of the cartilaginous splanchnocranium have been formed by the time of hatching. In the hyoid arch, the hypohyal, ceratohyal and interhyal parts are present and in the hyosymplectic-pterygoquadrate plate, distinguishable hyosymplectic and quadrate parts are present. In the lower jaw, Meckel’s cartilage is present, rostrally bearing a small layer of teeth-bearing dentary bone. In the premandibular arch, the autopalatines and both teeth-bearing premaxillary bones are visible. A ventral and dorsal hypural are present in the caudal skeleton, but no notochord flexion was observed yet. Stage 2: 4.3 mm SL (Fig. 4.35C) The neurocranium floor has further expanded splitting up the metotic fenestra. Given the situation described under 4.2, in C. pygmaeus, the anterior fenestra was also designated as the fenestra basicapsularis anterior, while the posterior aperture was called foramen “A”. This is corroborated by the fact that, at the caudal edge of the neurocranium, as in C. aeneus, two foramina for the vagus (X) nerve are present. Stage 3: 4.8 mm SL (Fig. 4.36) On the dorsocaudal side of the neurocranium, the otic capsules have expanded medially and the tectum posterius has closed dorsocaudally. Rostrally, the taenia marginalis contact the solum nasi and ethmoid plate through the commissura spheno-ethmoidalis and commissura sphenoseptalis respectively. 109 4 ONTOGENY – MINIATURIZATION Stage 4: 5.6 mm SL (fig. 4.40B) In the skeleton of the caudal fin, both the dorsal and ventral hypural plates have expanded substantially. Stage 5: 6.6 mm SL (Fig. 4.37; 4.40C; 4.42A) In specimens with a SL of 6 to 7 mm, the otic capsules and neurocranium floor have both expanded. At the rostral end of the neurocranium the ethmoid region has broadened by lateral expansions of both solum nasi. The suspensorium has become more rigid and the hyosymplectic has developed a distinct articulatory facet for future articulation with the opercle. A pterygoid process is present at the ventrorostral margin of the suspensorium and the parurohyal bone has appeared in the hyoid arch. In the postcranial skeleton, notochord flexion has now occurred and vertebral centra start to develop. In the caudal fin, a cartilaginous haemal spine has been formed on the second preural vertebral centre. In the cartilaginous pectoral girdle, a dorsal scapular and ventral coracoid process are present and matrix decomposition has started in the endoskeletal disc forming the first distal radials. In addition, in the pelvic girdle, two small cartilaginous basipterygia have been formed. Stage 6: 7.5 mm SL (fig. 4.40D; 4.41A) From this stage on, the epural cartilage is present in the caudal skeleton. In the anal fin, fin rays have appeared, supported by pterygiophores that are not yet split up into proximal and distal radials. Cartilaginous pterygiophores have also appeared in the dorsal fin and are already split up into proximal and distal radials. Stage 7: 7.9 mm SL (fig. 4.38A) In the splanchnocranium, additional to the maxillary, premaxillary, opercular and parurohyal bone, ossifications have appeared in the autopalatine and anterior ceratohyal. Stage 8: 8.5 mm SL (fig. 4.40E) No additional ossifications are present in the neuro- and splanchnocranium and in the caudal fin, the first signs of early ossification (not yet stained) are seen. 110 4 ONTOGENY – MINIATURIZATION Stage 9: 9-9.5 mm SL (fig. 4.38B; 4.40F; 4.41B; 4.42B) At this stage, ossification has started throughout the entire neurocranium. In the skull roof, the mesethmoid, frontal, sphenotic, parieto-supraoccipital and posttemporo-pterotico-supracleithrum have formed, while in the skull floor the prevomer, parasphenoid, prootics and occipital complex are present. At the level of the skull wall, the pterosphenoid and orbitosphenoid are still confined to small layers of bone surrounding the fenestra sphenoidea. With the appearance of the preopercle and interopercle and the ossification of the quadrate, interhyal, hyomandibula, metapterygoid, articular complex and posterior ceratohyal, almost all elements of the bony splanchnocranium are also present from this stage on. All vertebrae and their neural and haemal spines have started ossifying at this stage, as is also the case for the elements of the caudal, dorsal and pectoral fin skeleton, although in the latter, no ossification was observed at the level of the fin rays yet (only at the level of the fin spine). In the pelvic girdle and anal fin, on the other hand, ossification of the fin rays had started but no ossification was present in the supportive skeletal structures yet. Stage 10: 10.0-11 mm SL (4.39A; 4.40G; 4.41C; 4.42C) At the level of the skull roof, the lateral ethmoid has appeared while all other bones now contact each other. In the skull wall, orbito- and pterosphenoid have enlarged substantially. In the splanchnocranium, all bones have ossified further, contacting each other and rigidifying the entire structure. Ossification in the postcranial skeleton is completed in all fins. Stage 11: 16 mm SL (4.39B) At this point the two tubular infraorbital bones are the last bones to appear. Discussion Growth In comparison to the five different inflexions found in the growth curve of Corydoras aeneus (see 4.2), only a single inflexion was found in the analysis of growth in C. pygmaeus (fig. 4.33B). This can partially be related to the lower 111 4 ONTOGENY – MINIATURIZATION number of specimens used in this analysis (26 compared to 39 used for C. aeneus). In addition, differences in growth could also be an effect of differences in dietary requirements in both species. Since the nutritional requirements of both species are unknown, they were fed a similar diet throughout their ontogeny (see 2.1), this way attempting to minimize the impact of nutrition. A thorough analysis of the optimal dietary conditions could however, prove usefull in eliminating this possible effect on differences in growth. Nevertheless, the SL at which this inflexion point is situated (5.2 mm) does correspond to the second inflexion found in the growth of C. aeneus (5.4 mm). In C. aeneus the occurrence of this inflexion was linked to a change in swimming mode around this SL, combined with the onset of finfold differentiation (start of pterygiolarval phase) and notochord flexion. Weihs (1980), in his study on Engraulis mordax, found a continuous swimming mode to be more efficient in larvae under 5 mm, while in larvae with a length between 5 and 10 mm burstand-coast swimming became more effective. Blaxter (1986), on the other hand, mentions that a change in swimming mode occurs as the caudal fin develops and inertial forces start to play a bigger role in locomotion. Although in C. pygmaeus, no notochord flexion is observed, development in the caudal fin skeleton is ongoing at that point and both finfold differentiation as well as notochord flexion occur only slightly later in ontogeny (at 6.2 and 6.6 mm SL respectively). This all indicates that the single inflexion point observed in the growth of C. pygmaeus may also be related to a change in swimming mode similar to the one suggested for C. aeneus. External morphology At hatching, both species are of comparable size (3.5 mm SL in C. aeneus; 3.7 mm SL in C. pygmaeus), although duration of prehatching development in C. aeneus (3.3 days ± 0.5; n=120 (MEAN ± S.D.)) was significantly shorter (p= 4.3E-06) than in C. pygmaeus (4.1 days ± 0.3; n=12). After this, the yolk sac becomes depleted in both species at a SL of 4.4 mm, respectively at 1 dph (5 dpf) in C. pygmaeus and 3 dph (6 dpf) in C. aeneus. Although no information on egg size in C. pygmaeus was recorded, the delay in time seems to be result of a decreased rate in the development of C. pygmaeus since yolk sac size in hatchlings was highly comparable (fig. 4.4A; 4.34A). After this point, the SL at which transitions between the different developmental phases occur starts to 112 4 ONTOGENY – MINIATURIZATION differ between both species. While the shift from protopterygiolarval to pterygiolarval phase is somewhat comparable between both (6.2 mm SL in C. pygmaeus and 5.7 in C. aeneus), the start of the juvenile period in C. pygmaeus occurs at 8.9 mm SL, which is substantially lower than the 14 mm SL at which this shift takes place in C. aeneus. These shifts, however, do correspond to a comparable age of 22 days in C. pygmaeus and an age of 23 days in C. aeneus (fig. 4.43A; B). Chondrocranium Generally, during development of the chondrocranium, the cranial morphology of C. aeneus and C. pygmaeus is comparable and no major differences are noted. The fact that both halves of the Meckel’s cartilage are medially fused in the earliest stage described for C. pygmaeus is one of such few differences but this could also be the result of staining of the ligamentous tissue between both halves at this point, subsequently masking the separation of both halves (no histological sections were available to verify this). The fact that, from that point on, in no other stages a similar fusion has been found, seems to confirm this. However, the level of development in the chondrocranium of both investigated species at the time of hatching, is remarkably different, which could relate to the longer prehatching stage in C. pygmaeus. The level of chondrification in C. pygmaeus is very high, whereas in C. aeneus only the onset of chondrification is present in hatchlings. In general, the complexity of the chondrocranium at hatching, although highly variable in teleosts, is generally limited in Siluriformes, while studies on species of the Loricarioidea group, however, tend to show a more developed chondrocranium at that point (see 4.3). The fact that, in both congeneric species investigated here, the degree of chondrification of the neurocranium also differed remarkably, confirms the high variability of the chondrification state at hatching in general. Again, the difference observed probably relates to the longer prehatching stage in C. pygmaeus. Fig. 4.43 indeed shows that development of the chondrocranium takes place over a similar age-range in both species, with a delay of hatching in C. pygmaeus. At the time of yolk depletion, the level of development of the chondrocranium in both species has become highly comparable (figs. 4.9; 4.36). This condition evidently may relate to the fact that, from this stage on, exogenous feeding becomes obligatory 113 4 ONTOGENY – MINIATURIZATION and the presence of a functional feeding apparatus is required (van Snik et al., 1997; Jardine & Litvak, 2003). Osteocranium In both species the onset of ossification in the neurocranium was situated around 8-9 mm SL. This size, however, corresponded to an age of around 10 dph in C. aeneus, but an age of around 20 dph in C. pygmaeus. In both species ossification of most elements continues up to the age of 35-40 days post hatching, but in C. pygmaeus, as mentioned earlier, the onset of ossification is at 20 days, while this occurs at 10 days in C. aeneus. During this period, growth in the skull of C. aeneus continues, while in C. pygmaeus it almost stops. In C. pygmaeus, however, the last elements, the infraorbital bones, appear substantially later, at 105 dph. The result of this is that, at the end of cranial ossification, next to the skull being smaller in C. pygmaeus, tubular infraorbital bones, typically regarded as typical paedomorphic features (Weitzman & Vari, 1988) are also retained. Postcranial skeleton While initial formation of the cartilaginous postcranial skeleton starts at a lower SL (and age) in C. pygmaeus, ossification in most parts, as is the case in the cranium, starts at 8-9 mm SL in both species. This corresponds to a substantial time delay in C. pygmaeus. Again, this could help to explain the smaller size of the adult, as well as the lower number of dorsal fin rays, both considered typical paedomorphic, or reductive, traits (Weitzman & Vari, 1988). Heterochronies Because heterochrony deals with changes in the rates and timing of developmental processes, the most straightforward way to study it is to compare the actual curves depicting measures of size as a function of developmental time (Klingenberg, 1998). In this case, the analysis of both growth curves points toward neoteny as the main paedomorphic mechanism in the miniaturization of C. pygmaeus, since, in the formalism of Alberch et al. (1979), neoteny is defined as a decrease in the rate of development. In both C. pygmaeus and C. aeneus, growth rate is comparable in specimens under 10 dph (SL of around 6.5 mm), after which, growth speeds up in C. aeneus, but not in C. pygmaeus (fig. 4.33B). 114 4 ONTOGENY – MINIATURIZATION In the light of specific traits, next to rate, shifts in the point of onset and offset can also produce paedomorphisms. In this case, the clearest example is the postdisplacement of onset of ossification in the cranium in C. pygmaeus. The results also show this onset in ossification to be size-related (8 mm SL), as is also the case for offset of development in the chondrocranium (5-6 mm SL). On the other hand, offset of ossification appears to be age-related (35-40 dph). Overall, the ontogenetic pattern in these species appears to be very conservative with both age- and size related aspects. This way, growth rate and the subsequent postdisplacement of size-related traits (like the onset of ossification at 8 mm SL) appear to be the main cause of a miniaturized body in C. pygmaeus. In the context of the identification of heterochronic events through interspecific comparisons, Fuiman et al. (1998) suggested the use of an ontogenetic index. This index OL gives a measure of the relative timing (in size) of an ontogenetic event in relation to the size at metamorphosis. These measures are put on the log scale and the relation is mutiplied by 100, placing these events on a scale between 0 and 100. In this study, calculations are based on SL, and this way, for example, the ontogenetic index of the moment of hatching is calculated as: log (SLhatch)/log(SLmetamorph.)*100. These indexes were calculated for hatching, yolk depletion and finfold differentiation and results are shown in fig. 4.44. This shows a clear neotenic mechanism behind the miniaturization in C. pygmaeus since similar ontogenetic events have been delayed in time in this species. In conclusion, miniaturization in the body of C. pygmaeus is indeed the result of paedomorphic events, which include a lower growth rate and subsequent postdisplacement of several traits. These paedomorphic events have, however, besides the smaller size, not led to the retention of a large number of paedomorphic traits. 115 4 ONTOGENY – MINIATURIZATION 116 Chapter 5 General Discussion 5 GENERAL DISCUSSION 5.1 THE ONTOGENY OFCORYDORAS AENEUS The theory of saltatory ontogeny Studies on the early development of fish have viewed ontogeny either as a continuous, inconspicuous accumulation of small changes (gradual ontogeny), or as a series of steps and thresholds (saltatory ontogeny) (Vilizzi & Walker, 1999). The concept of saltation in ontogeny was first proposed by Vasnetsov (1953) and Kryzhanovsky et al. (1953) who proposed a ‘step theory’ in which they recognized ontogeny to be a sequence of ‘etapes’ of quantative morphogenesis and growth, separated by a combination of brief but distinct qualitative changes (‘leaps’) in a developmental process. Balon (1959, 1975, 1979, 1990) later adapted this theory and formulated saltatory ontogeny as a sequence of longer stabilized states, or steps, punctuated by rapid changes in ‘integrative actions’, which form ‘thresholds to survival’. Such thresholds are considered abrupt functional changes in ontogeny that produce relationships to the environment (McElman & Balon, 1980). Balon (1959, 1971) also stated that: “these thresholds could very well be the most important intervals of ontogeny, during which adaptations to the environment express themselves and the future course of ontogeny is determined”. A clear recognition of these thresholds would however, prove to be extremely difficult based on the mere study of adult specimens (Balon, 1979). Therefore, ontogenetic studies aiming to recognize these important thresholds should aim to describe as many features of the developing larvae as possible (Balon, 1979). In my study of Corydoras aeneus, I examined the ontogeny of skeletal and muscular elements during cranial and postcranial development. Although no additional internal organ systems were studied, the data presented here should suffice in recognizing the position of some of these important thresholds during ontogeny. In addition, under 4.2, an analysis of several mensural characters was performed. Mensural characters with inflexion points occurring at similar size are also suggested to be associated with some common ecological, physiological and/or behavioural function and could therefore also prove useful in recognizing thresholds during development (Kováč & Copp, 1999). 117 5 GENERAL DISCUSSION Finally, similar data on Corydoras pygmaeus (see 4.7) and equally extensive comparable data on Ancistrus cf. triradiatus (Geerinckx, 2006) allow for a comparison of these thresholds in all three species thus adding an evolutionary dimension. Given the extensive differences in the development, morphology and early life history between both A. cf. triradiatus and C. aeneus, I expect the position of these thresholds to differ substantially between both ontogenies. Since the objective of this part is to compare the differences in the number and relative position of thresholds related to the differences in the ontogenetic trajectories of all three species, it is sufficient that only a single axis of comparison was used. For this purpose, I chose SL, since most data on all three species describe the occurrence of the different developmental events in a SLbase staging system. The ontogeny of Corydoras aeneus In this part, the results presented under chapter 4 are briefly recapitulated, with the exception of the egg morphology. The beginning and end of the development of different morphological units and the position of other important events in the development are subsequently mapped onto a SL axis depicting development in Corydoras aeneus as was done under 4.7 (Fig. 4.43). This way, those positions in development in which several important events (inflexions in growth, onset and offset of development in certain structures and the transitions between the different developmental phases) co-aligned, were designated as thresholds. As an arbitrary marker, I chose only to describe those points at which at least three events clearly co-aligned as thresholds. This allowed for a more objective comparison with the ontogenies described for C. pymaeus and Ancistrus cf. triradiatus, as presented below. Ontogeny Early development and growth Based on the study of ontogeny of external morphology, four different phases were recognized in the ontogeny of C. aeneus. First an eleutherembryonic phase is present from hatching until 4.4 mm SL, during which specimens are 118 5 GENERAL DISCUSSION unpigmented and the mouth and anal opening are still closed. During the protopterygiolarval phase, at a SL of 4.4 to 5.7 mm, finfold differentiation starts, the yolk sac disappears and the anus, mouth and opercular cavity open. After this, at a SL between 5.7 and 14.0 mm (the pterygiolarval phase), the finfold further differentiates until all fins have fully detached from the larval finfold. In addition to an analysis of external morphology, a separate analysis of change in overall growth yielded five different inflexion points at which growth rate changed significantly. These points existed at 3.9, 5.4, 7.4, 14.5 and 17.9 mm SL. Chondrocranium The development of the chondrocranium starts around the time of hatching at a SL of 3.3 mm. At 5.3 mm SL, almost all elements of the cartilaginous skull are present with the exception of some structures in the branchial basket which are still fused at that point. Osteocranium The first ossifications in the skull of C. aeneus were found at the level of the splanchnocranium in which the opercle and tooth-bearing elements in the oral and pharyngeal jaws develop a thin layer of bone in between 4.9 and 5.3 mm SL. After this, no new centers of ossification are found in the skull until 8.2 mm SL. From that point on until 13.9 mm SL the remaining elements of the splanchnocranium ossify, while ossification of the neurocranium continues until 18 mm SL. Cranial myology The first signs of cranial muscles become apparent at the time of hatching, at 3.3 mm SL. After this, development of the cranial musculatures occurs rather fast, with all different muscles present from a SL of 5.3 mm on. Postcranial skeleton In C. aeneus the development of the postcranial skeleton takes place between 4.8 and 15-16 mm SL. Thresholds Fig. 4.45A shows all elements described above mapped onto a SL axis. As the two most important events during early development, hatching and yolk depletion are also indicated on this axis. Four thresholds were found at those SL 119 5 GENERAL DISCUSSION at which several events co-aligned. The overview clearly shows that development in C. aeneus is discontinuous with hatching co-occurring with a first threshold, at which point development of the chondrocranium and the cranial musculature starts and shortly after which a first inflexion in growth occurs. This is followed by a first stabilized step occurring between 3.3 and 5.3 mm SL in which yolk is depleted, the chondrocranium further develops and all cranial muscles and the first elements of the bony splanchnocranium (the tooth bearing elements) appear. A second threshold occurs at the time at which all cranial muscles, the tooth bearing elements of the splanchnocranium and all parts of the chondrocranium have been formed. At this point, a second inflexion in growth was also found. Just prior to this point, development of the cartilaginous postcranial skeleton has started and shortly after this point finfold differentiation starts. Between 5.3 and around 8 mm SL, a second stabilized state is present in which only in the postcranial cartilaginous skeleton, new elements arise. The third threshold is located at 8 mm SL, shortly after a third inflexion in growth and at the same time of the onset of ossification in the rest of the splanchnocranium, the postcranial skeleton and the osteocranium. After another stabilized period between c. 8 and 14 mm SL, the last threshold is found. At this point, ossification in the splanchnocranium has completed, a fourth inflexion in growth is found and the juvenile period starts. Conclusively, the analysis of the development of both the musculoskeletal system and external development indeed show strong indications of saltatory ontogeny in C. aeneus. Since this can not be determined definitively without a further analysis of additional organ systems beyond the musculoskeletal apparatus, this conclusion appears somewhat preliminary. However, the different morphological traits used for the demarcation of the different developmental phase also include several aspects of the development of the internal organ systems. This way, depletion of yolk and opening of anal and mouth opening indicates a functional digestive tract, whereas fin development is indicative of developing locomotory functions. Therefore, although additional data is needed to confirm the exact points at which thresholds occur in the ontogeny of C. aeneus, it is clear that ontogeny in this species follows a saltatory pathway. During this ontogeny, at least four thresholds are present in which important events co-align. 120 5 GENERAL DISCUSSION Corydoras pygmaeus and Ancistrus cf. triradiatus Corydoras pygmaeus (fig. 4.45B) In the early development of C. pygmaeus also four different phases are found. The eleutherembryonic phase starts at 3.7 mm SL until, at 4.4 mm SL, yolk becomes depleted (fig. 4.34A;B). After this, from 4.4 mm SL until 6.2 mm SL, the protopterygiolarval phase is present. At the end of this phase, finfold differentiation starts and specimens reach the pterygiolarval phase (fig. 4.34C). At 8.9 mm SL all fins are fully detached from the larval finfold and the juvenile period starts (fig. 4.34D). The growth curve of C. pygmaeus only shows a single inflexion point at 5.2 mm SL. Development of the chondrocranium starts during the embryonic phase which leads to a high level of development in the chondrocranium at hatching. By around 4.8 mm SL, all elements of the cartilaginous cranium have been formed. In C. pygmaeus, the first ossifications found in the developing cranium are located at the level of the splanchnocranium (opercle, dentary and premaxilla). These early ossifications arise between hatching and 4.3 mm SL. After this, ossification of the other cranial elements does not start until 8-9 mm SL. At 16 mm SL, the infraorbital bones are the last elements to ossify. Development in the postcranial skeleton of C. pygmaeus starts at hatching and the different elements start to ossify at 8-9 mm SL. Ossification in the postcranial skeleton is completed at around 10-11 mm SL. As in C. aeneus, it is clear that the ontogenetic trajectory fits the definition of saltatory ontogeny as given by Balon (1959, 1975, 1979, 1990). Also similar to C. aeneus, the first threshold is located at hatching, while the second threshold occurs around the end of development of the chondrocranium at a point at which growth rate changed. In contrast to the situation in C. aeneus, however, this point also corresponds to the time of yolk depletion in C. pygmaeus. A third and last inflexion, again similar to C. aeneus, is found at the time at which ossification in the neurocranium and postcranial skeleton starts. In contrast to the situation in C. aeneus, this point does not correspond to the onset of ossification in the non-tooth-bearing elements of the splanchnocranium, which is already ongoing at that point. Another difference with C. aeneus is that this point corresponds to the start of the juvenile phase. 121 5 GENERAL DISCUSSION Ancistrus cf. triradiatus (fig. 4.45C) In the early development of A. cf. triradiatus, only three distinct periods were recognized. Embryos hatch at 6 mm SL, after which the yolk sac remains present until 10 mm SL, which is a distinctly longer period than in C. aeneus. At the moment of yolk sac depletion, specimens of A. cf. triradiatus already have an adult-like appearance and therefore, a true larval stage is absent (cfr. Balon, 1975). Data by Geerinckx (2006) also showed that the development of the chondrocranium in A. cf. triradiatus starts around 5.2 mm SL, well before hatching and that most structures of the cartilaginous skull are formed by 8.9 mm SL. Also in this species, the first elements of the skull to ossify were the tooth-bearing elements and opercle (at 5.6 mm SL), but after this, ossification of the splanchnocranium continues in each of the stages examined. During the ontogeny of the cranial muscles in A. cf. triradiatus, the first elements appear at a SL of 6.1 mm, while the last elements are formed by a SL of 10.2 mm. The method described under 4.2 for the calculation of inflexion points in growth was not applied to the growth data of A. cf. triradiatus and therefore no comparable data in this segment is present, nor is this the case for data on postcranial development. However, it seems clear that the ontogeny of A. cf. triradiatus follows a totally different path than those found in both Corydoras species. The only point at which a possible threshold is found is at the point of hatching, just after the onset of splanchnocranium formation, and co-aligned with the onset of development of the cranial muscles. Due to this limited number of thresholds and the absence of stages in which very little new structures are developed (compare to both Corydoras species), the term ‘saltatory ontogeny’ does not apply to the ontogeny of A. cf. triradiatus. Thresholds in the development of Loricarioidea The results clearly show that the main difference found between the loricariid and callichthyid ontogenies studied here is the fact that the number of thresholds is substantially lower (or even totally zero) in the ontogeny of the loricariid. The only threshold found in all three species was the point of hatching itself, although the importance of hatching as a threshold of ultimate survival value has already 122 5 GENERAL DISCUSSION been questioned (Balon, 1984). Additional thresholds found in the ontogenies of both callichthyids correspond to other important life history events such as the transition to exogenous feeding and the point at which the larvae become juveniles. Since the point at which the yolk sac is depleted in A. cf. triradiatus has been postponed in comparison to both callichthyids, most cranial structures have been formed by that point and the point probably forms a smaller threshold to survival than in both Corydoras species. In addition, it is known that in loricariid fishes a true larval stage is absent (Geerinckx, 2006), again eliminating that point as a threshold in early development. In general, such a reduction of the larval period in fishes is believed to be associated with the development of reproductive hiding strategies (or other forms of parental care), which itself is directly related to an increase in egg size and yolk density (Balon, 1979; Kolm & Ahnesjö, 2005). In general Balon (1984) formulated the relation as: “In fishes with an increased endogenous food supply and parental care (increased reproductive cost per embryo), the embryo develops permanent organs directly, bypassing the larva period with its remodelling of temporary structures”. Indeed, in this case, the differences in egg size seem to have clearly influenced the differences found in the early ontogenies of Corydoras and Ancistrus species. This way, it was already shown that the appearance of a highly developed chondrocranium at hatching in A. cf. triradiatus was probably the direct result of this increase in egg size. The analysis presented here now additionally shows that the prolongation of the embryonic period also seems to affect the occurrence of thresholds and a subsequent saltatory developmental pattern in early ontogeny, with the elimination of a metamorphosis from a larva into a juvenile as the most striking result. The elimination of this larval period is generally accepted as an important ecological and evolutionary phenomenon (Balon, 1984) and may therefore have played a crucial part in the evolution of a specialized feeding apparatus in Loricariidae. The prolonging of the embryonic period would, after all, have provided the time necessary for the development of these structures since active feeding only becomes necessary at a point much later in development. Given the well documented relation between the elimination of the larval period, egg size and parental care, it seems very likely that selection at the level of the reproductive strategies in loricariids has also played an important part in 123 5 GENERAL DISCUSSION the radiation and evolutionary success of this family, but this will be further discussed under 5.2. 124 5 GENERAL DISCUSSION 5.2 THE EVOLUTION OF ALGAE SCRAPING IN LORICARIOIDEA Scaridae, another example of adaptive radiation and algivory in teleosts It has been argued that seemingly different vertebrate radiations follow similar evolutionary trajectories in which groups diverge along axes of habitat, trophic morphology and communication, often in that order. Speciation events in stages one (habitat) and two (trophic morphology) probably follow the rules of ecological selection models, while lineage diversification in stage three (communication) would occur according to sexual selection models (Streelman & Danley, 2003). Seehausen (2004) has also suggested a similar model of radiation, although he did not suggest an order of the different stages, but does state that three main factors influence radiation. These factors include release from competition in an underutilized environment, key evolutionary innovations that enable utilization of resources that were there but could not be utilized and mating systems that are conducive to rapid divergence (Seehausen, 2004). In general, both hypotheses merge the traditional ecological radiation theory with the sexual model alternative, with or without presenting an order in which both effects occur. Streelman & Danley (2003) exemplify their statement, including the order of events, using various examples including the example of the family of the parrotfishes (Scaridae), another teleost group in which algivory has evolved. Although it must be stated that not all loricariids are algivorous, the evolution of a highly specialized feeding apparatus with associated algivory in most species does seem to form the cornerstone of their radiation. Therefore, given the relevance of the similarity in trophic specialization between Scaridae and Loricariidae and the fact that the example is situated within the teleosts, it seems appropriate to discuss it here. In Scaridae, divergence has clearly started along the habitat axis when 40 million years ago, scarids diverged into a ‘reef’ and ‘seagrass’ clade. Second, members of the Scaridae family in the reef habitat diverged into lineages that either scrape or excavate algae from coral and rock 125 5 GENERAL DISCUSSION surfaces (Streelman et al., 2002). These differences in trophic utilization were associated with changes in the craniofacial skeleton and musculature, feeding rate and kinematics (Bellwood & Choat, 1990; Bellwood, 1994; Alfaro & Westneat, 1999). By contrast, the seagrass lineage of parrotfishes does not exhibit secondary divergence in morphological features; nearly all species are browsers (Streelman & Danley, 2003). Finally, within the reef clade, a tertiary radiation is observed which involves a sexual dimorphism at the level of male mating coloration, correlated with behavioral differences (e.g. territoriality and a haremic mating system) (Streelman et al., 2002; Streelman & Danley, 2003). Given this, it is beyond doubt that the radiation enclosing the evolution of algae scraping in Scaridae has evolved along the three axes mentioned before and in that specific order. The question dealt with in the following chapter, however, is whether or not the same model applies for the Loricarioidea. Given the fact that Streelman & Danley (2003) argued that this model applied to different vertebrate radiations, I hypothesize that the model is also applicable here. One can test for evidence supporting these different evolutionary steps in morphology, ontogeny, ethology and biogeography of the extant loricarioids, as well as in the known fossil record of the superfamily. Given the limited record existing on Scoloplacids, the following discussion will be limited to evidence found in Callichthyidae, Astroblepidae and Loricariidae. Three phases of evolution Habitat divergence As habitat divergence and subsequent lineage divergence is believed to follow ecological speciation models, both allopatric and non-allopatric models of divergence can apply to such divergence and examples from both are known (Streelman & Danley, 2003). In this context, allopatric speciation is defined as a process in which a reduction to zero of the number of breeding individuals in a population that are immigrants from other populations, is accomplished by a physical barrier extrinsic to the organism (Futuyama & Mayer, 1980). Nonallopatric (albeit parapatric or sympatric) speciation is then accomplished by 126 5 GENERAL DISCUSSION isolation through biological characters of the organisms themselves (Futuyama & Mayer, 1980). The idea of sympatric and/or parapatric speciation has long been considered highly unlikely (Coyne & Orr, 2004; Seehausen, 2004) but recently, several studies have showed evidence for such sympatric/parapatric speciation models. Since most adaptive radiations occur in geographically narrowly confined regions, sympatry and parapatry during part of the speciation process is invoked in many (Seehausen, 2004). The divergence of ‘reef’ and ‘seagrass’ clades in Scaridae provides an example of such a sympatric process in which the split in habitat might have occurred in the absence of physical barriers that would prevent gene flow (Streelman & Danley, 2003). Also, the cichlid fishes of Lake Barombi Mbo (Cameroon) provide one of the most convincing examples of sympatric speciation in animals (Turner, 2007). Nonetheless, although sympatric speciation is theoretically plausible and supported, there is little evidence that it is common (Coyne & Orr, 2004). More common are examples in which lineages diverge in allopatry, like the Hawaiian honeycreepers and Galápagos finches (Lack, 1947; Grant, 2000). In these cases, lineage splitting is believed to have begun in allopatry, but to have been accelerated greatly following secondary contact (Grant, 2000). Given the more common nature of the allopatric speciation model and the broad geographical range in which the adaptive radiation in Loricarioidea has occurred (in contrast to other known radiations), it would seem plausible that habitat divergence in Loricarioidea has also occurred in allopatry. The geographic separation needed for allopatric speciation can occur through diverse climatic or geological events or the extinction of intermediate populations (Coyne & Orr, 2004). In order to elucidate this, one needs to look at the biogeographical context and fossil evidence of the Loricarioidea. Since loricarioids are generally believed to form the sister group of African amphiliids (de Pinna, 1993; 1998), the superfamily itself could have originated >112 Ma, the point at which both the separation of Africa and South America is completed through the opening of the South Atlantic (Lundberg, 1998; Maisey, 2000). Recently, Sullivan et al. (2006) placed the loricarioid superfamily as a sistergroup to all other catfish based on molecular data. This subsequently led to a estimate of divergence between Loricarioidea and all other catfish to 125 Ma, predating the Gondwana break up (Lundberg et al., 2007). In both cases, these estimates by far predate the hypothesis formulated in the refuge theory by Haffer (1982), 127 5 GENERAL DISCUSSION which has long been one of the most popular theories explaining Neotropic diversity. This theory states that most of the Neotropic divergence occurred recently and that the primary forces were the climatological fluctuations of the Pleistocene (1.8 million to 11.000 years ago), which fragmented and reconnected habitats several times, promoting allopatric speciation (Montoya-Burgos, 2003). Since new morphological as well as molecular data predates this, diversification has been linked to the particularly eventful earlier (< 1.8 Ma) history of Neotropical river systems. Based on this, the so-called ‘hydrogeological’ hypothesis was formulated (Montoya-Burgos, 2003). This states that the formation of drainage divides, shifting courses of rivers and repeated incursions and regressions of marine waters must have produced many vicariant events promoting biotic enrichment (Lundberg, 1998). The fossil evidence found in the loricarioid group also corroborates this hydrogeological hypothesis. The earliest fossil known in this superfamily is a callichthyid described as Corydoras revelatus, from the late Paleocene, ca. 58.5 Ma, found in the Argentinian Maíz Gordo formation (Lundberg, 1998; Lundberg et al., 1998). This indicates that, at least up to the level of Callichthyidae, divergence in loricarioids also long predates the Pleistocene. Based on the remaining loricarioid fossil record, Lundberg (1998) infers that a considerable diversity of modern loricarioids had evolved by at least 13.5 Ma. This is recently confirmed by a molecular analysis of some members of the Hypostomus genus, which dates back loricariid diversification at the genus level to 12 – 4 Ma (Montoya-Burgos, 2003). In addition, recent fossil calibrations estimate the Callichthyidae – (Astroblepidae + Loricariidae) divergence at 110 Ma (Lundberg et al., 2007). This way, it is clear that various vicariance events may have played a crucial role in the early divergence of loricarioid clades, and have given the onset of habitat divergence. Regarding habitat divergence in the case of vicariance, much depends on the habitat both groups are isolated into. The most evident manner in which allopatry can lead to a divergence in habitat use is the case in which both groups get isolated into different habitat types. On the other hand, both groups could get isolated in similar habitat types but other selective forces (like competition) or genetic drift could drive both groups into different directions of habitat use. The former option seems the most plausible in this specific case, since the Maíz Gordo formation, in which the Corydoras revelatus fossil was found, was believed to be a wetland to lacustrine environment (Lundberg et al., 1998), while the 128 5 GENERAL DISCUSSION distribution of extant Astroblepidae (no fossil taxa within this family are known (Ferraris, 2007)) is limited to high elevation streams of the Andes (Schaefer, 2003). This limited distribution could indicate that the distribution of the common ancestor of Astroblepidae + Loricariidae is also located in this narrow geographical range, with a secondary recolonization of the Brazil and Guyana shield by loricariids. Another hypothesis could be a broad distribution in the common ancestor to both families with a secondary limitation in the distribution of Astroblepidae. This would imply that Astroblepidae have evolved from an Andes clade of this common ancestor into a separate family. The divergence between Astroblepidae and Loricariidae indeed dates back to 90 Ma (Lundberg et al., 2007), at which point formation of the Andes started, altering the entire hydrogeography of the South American continent (Lundberg et al., 1998). The fact that the distribution of extant Loricariidae also covers the high Andes would, however, imply a secondary recolonization of the Andes by this group. In addition, Astroblepidae possess several derived features like a suctorial mouth and a specialized pectoral to pelvic fin muscle system, two structures that can be used in an alternating manner, giving astroblepids the capability of climbing vertical surfaces and colonize fast water currents (Shelden, 1937). This matter also partially answers the question posed by Adriaens (2003) and Geerinckx (2006) on the evolutionary order of suckermouth attachment versus scraping feeding mode in loricarioids. The evolution of attachment modes in species colonizing turbid environments is indeed common, albeit often developed from other parts than the mouth (Geerinckx, 2006). Also, the fact that a sister family like the Astroblepidae only possesses a suckermouth without the specialized feeding mechanism indicates an appearance of this suckermouth earlier in evolution. In addition, Lundberg et al. (2007) place the common ancestor to (Scoloplacidae + Astroblepidae + Loricariidae) only at 95 Ma, which places the evolution of a suckermouth in Loricarioids between 95 and 90 Ma, at the same time of the start of Andes formation (Lundberg et al., 1998). All this indicates that, at the basis of loricarioid evolution, several vicariant events took place splitting up the Callichthyidae and the common ancestor to Astroblepidae and Loricariidae into habitats with highly different hydrological characteristics. 129 5 GENERAL DISCUSSION Trophic morphology The fact that a great extent of the divergence of the different families of loricarioids has occurred at the level of their trophic morphology forms the main reason this project was started (see 1.1). This is related to the fact that radiation in Loricarioidea has peaked in the Loricariidae, the family in which, in addition to morphological changes, dietary specializations have also evolved, indicating the importance of a divergence in trophic morphology. In addition, loricariids also greatly exceed all other loricarioids in terms of diversity of jaw and teeth size and shape (Schaefer & Lauder, 1986; Geerinckx & Adriaens, 2006). Therefore, it is generally believed that the specialized trophic morphology of the Loricariidae forms the key-stone of their evolutionary success. In 1981, Lauder proposed a general hypothesis concerning evolution of structural patterns as: “primitive members of a morphologically diverse monophyletic lineage possess functional, structural, or morphogenetic networks which have a greater number of independent (decoupled) elements as compared to similar networks in closely related but less morphologically diverse monophyletic lineages”. Schaefer & Lauder (1986) later referred to this as the ‘decoupling hypothesis’. The definition of decoupling in this case can be stated as the repetition of individual elements as redundant design components, followed by the specialization of one or more of these elements as a mechanistic basis for the evolution of a novel structure/function (Schaefer & Lauder, 1996) (also see 1.1; 3.2). Schaefer & Lauder (1986, 1996) showed that the Loricarioidea are a clear example of a clade showing a pattern of progressive increase in mechanical complexity, in this case, in those structures associated with feeding (Schaefer & Lauder, 1986). In the loricarioids, decoupling has mainly occurred at the level of the jaws. In Callichthyidae, Astroblepidae and Loricariidae (Scoloplacidae were not dealt with in the study by Schaefer & Lauder (1986, 1996)), the premaxillaries have been decoupled from the cranium, increasing their mobility through modifications in the ethmoid hinge joint (see 3.1). In addition, Schaefer & Lauder (1986) mention a shift of a dorsal extension of the m. adductor mandibulae onto a sheet of connective tissue between lower and upper jaw. This muscle has been called m. retractor tentaculi by Howes (1983) and under 3.2 and 4.5 in this thesis. In addition, a subdivision of the m. extensor tentaculi in C. aeneus was not reported by Schaefer & Lauder (1986, 1996), although it also fits the general trend of an increase in morphological complexity through decoupling: 130 5 GENERAL DISCUSSION in loricariids, the muscle becomes completely subdivided, allowing an independent mobility of the maxilla and premaxilla (Geerinckx, 2006) (see 3.2). In Astroblepidae and Loricariidae, upper jaw mobility is further improved through a direct insertion of the m. retractor tentaculi on the premaxillaries and a loss of the connection between the lig. primordium and the maxilla. In addition, the lower jaw has been decoupled from the opercular series, allowing an independent mobility of both upper and lower jaw (Schaefer & Lauder, 1996). Decoupling events have also played a major role at the level of the lower jaws and hyoid musculature. These have become decoupled from their plesiomorphic bilaterally constrained midline attachments and a new linkage between hyoid and mandible was acquired (Schaefer & Lauder, 1996). It is clear that all these decoupling events have facilitated the formation of a functional suckermouth in a first phase with a subsequent algae scraping feeding apparatus in a second phase. This divergence in trophic morphology has this way played a major role in the radiation process of the loricariids. Communication In the model by Streelman & Danley (2003) this stage in vertebrate radiation occurs according to sexual selection models, which explain divergence through competition for mating opportunities. This often involves the development of male secondary sexual characteristics (Streelman & Danley, 2003). Indeed, also in the specific case of loricarioid evolution, a high degree of sexual selection seems to have played a significant role. In Callichthyidae, sexual dimorphism is confined to differences in size and length of the dorsal and pectoral fin spine (Kohda et al., 2002; Pruzsinszky & Ladich, 1998), while many Loricariidae are known for their elaborate sexually dimorphic modifications present in mature males. Some of these sex dimorphisms are pervasive throughout loricariids and are considered generalized traits for the family (Rapp Py-Daniel, 2000). These traits include the presence of large odontodes on head, fins and body on mature males (Rapp Py-Daniel, 2000). Examples of sexual dimorphism in Loricariidae are well documented in literature and include studies on Rineloricaria, Pseudancistrus, Otocinclus, Hypostomus, Farlowella and many more (Aquino, 1994; Mazzoni & Caramaschi, 1995; Retzer & Page, 1997; Rodríguez & Miquelarena, 2005; Lujan et al., 2007). In species of the hypostomine Ancistrinitribe, which is considered the most derived clade in loricariid phylogeny 131 5 GENERAL DISCUSSION (Armbruster, 2004), this sexual dimorphism has even evolved beyond these general traits. In these species an erectile cheek-spine apparatus has evolved as one of the synapomorphies of this tribe (Schaefer, 1987). An example of such sexual dimorphism has been described in Ancistrus cf. triradiatus, in which a tuft of enlarged odontodes anterior to the opercle can be erected with great velocity, acting as a defensive/offensive mechanism and which is, on average, more developed in male fishes (Geerinckx & Adriaens, 2006). The evolution of such a cheek-spine apparatus has again been made possible by a decoupling event. The evolutionary decoupling of the opercle from the lower jaw movements has evidently not only increased lower jaw mobility; it has also altered the function of the opercular apparatus, which has lost its function in expiration (redundancy) and became ‘available’ for the acquisition of a novel function. In combination with the presence of odontodes on bony platelets in the skin, the possibility to develop a series of articulations between the opercle, the cheek spines, the cheek plates anterior to them and the quadrate, and, the possibility of the opercular musculature to expand substantially inside newly evolved skull cavities has led to the evolution of this cheek-spine apparatus (Geerinckx & Adriaens, 2006). The appearance of sexual dimorphism in Loricariidae also appears to coincide with the evolution of parental care and several loricariid species are known to exhibit parental care with great diversity in reproductive strategies. In several Otocinclus species, for example, no parental care is present (Schmidt, 2001). This fits the hypothesis of Isbrücker & Nijssen (1992b) that, in the primitive loricariid subfamilies Neoplecostominae and Hypoptopomatinae, sexual dimorphism is totally absent. Reis & Schaefer (1992) and Aquino (1994), however, did find some extent of sexual dimorphism in Eurycheilichthys pantherinus and Otocinclus flexilis and O. vittatus, but this was very limited. Males of some Hypostominae and Ancistrinae are known to guard their eggs in holes constructed in stream banks (Schmidt, 2001) and in both subfamilies examples of sexual dimorphism have been described (Armbruster, 2004; Geerinckx & Adriaens, 2006). In Loricariini, Isbrücker (1981) described several types of secondary sexual dimorphism and various examples of parental care have been documented in members of this tribe. This way, in Loricaria simillima and L. piracicabae, adhesive eggs are carried on the ventral surface of the body by the males and even by both sexes in L. cataphracta (Schmidt, 2001). In 132 5 GENERAL DISCUSSION Loricariichthys anus, L. maculatus, L. platymetopon and Paraloricaria vetula this attachment is located in an expansion of the lower lip (Taylor, 1983; Schmidt, 2001) and has therefore even been described as a form of mouth brooding in L. platymetopon (Lassala & Renesto, 2007). Males of species of Rineloricaria, Ricola, Sturisoma and Sturisomatichthys are even known to, not only guard (Covain & Fisch-Muller, 2007), but also fan the eggs and remove fungused individuals (Schmidt, 2001). In all these genera, except Ricola, sexual dimorphism included the presence of prominent odontode development in males (Isbrücker, 1981; Covain & Fisch-Muller, 2007). Covain & Fisch-Muller (2007) further described sexual dimorphism in Rineloricaria to include hypertrophied development of the odontodes along the sides of the head, on the pectoral spines and rays, and the predorsal area of mature males. In male specimens of Loricaria they describe hypertrophied development of the pectoral spines, blunt odontodes on the pelvic and anal fin spines, and tooth crowns becoming shortened and rounded, while in Loricariichthys hypertrophied development of the lips was reported (Covain & Fisch-Muller, 2007). In general, in their overview of the Loricariinae, Covain & Fisch-Muller (2007), provide data on the reproductive biology of members of 21 genera, 12 of which exhibit both sexual dimorphism and parental care and two of which exhibit neither. In contrast to the hypothesized relation between parental care and sexual dimorphism three genera exhibit parental care only and in four genera sexual dimorphism is described without parental care (Covain & Fisch-Muller, 2007). Nonetheless, since 14 of the 21 genera described fit this hypothesis, there does seem to be a relation between parental care and sexual dimorphism. In Callichthyidae, the extent of parental care found is substantially lower, as is the case for sexual dimorphism. Only in the species Hoplosternum littorale, Megalechis thoracata and Callichthys callichthys some form of parental care has been observed. These species are known to build and guard foam nests in which eggs are deposited (Pascal et al., 1994; Ramnarine, 1994; 1995; Mol, 1996; Andrade & Abe, 1997; Hostache & Mol, 1998). Next to sexual dimorphism, in those groups of loricariids that have developed parental care, a lower fecundity and higher egg size are also noted (Suzuki et al., 2000; Geerinckx, 2006). Both the inverse relationship between parental care and fecundity as well as the positive correlation with egg size have been reported repeatedly throughout teleosts (Suzuki et al., 2000; Kolm & Ahnesjö, 2005). In 133 5 GENERAL DISCUSSION the context of the relation between egg size and parental care, Kolm & Ahnesjö (2005) have recently summarized four, not mutually exclusive, explanations for this general and positive correlation. A first hypothesis, the ‘safe harbour’ hypothesis states that selection will favor a prolongation of the time spent in the more protected embryo stage, while the ‘eggs require care’ hypothesis presumes that the larger eggs would require more care and that selection would therefore act upon the amount of care provided (Kolm & Ahnesjö, 2005). As a third and fourth hypothesis a mutual relation between both former hypotheses as well as the influence of other independent factors is suggested to select simultaneously for more care and larger eggs (coevolution) (Kolm & Ahnesjö, 2005). Although Kolm & Ahnesjö (2005) were not able to elucidate this matter and suggested more research (they proposed Cichlidae, Syngnathidae and Gasterosteiformes as suitable case studies, but the examples clearly indicate the suitability of the Loricariidae), the fact whether or not larger eggs preceded care of vice versa, is beyond the scope of this analysis. The fact remains that, in general, a relation between egg size and parental care exists and that this has influenced not only the matter of selection placed on the radiation process (from ecological to sexual selection), but also the ontogenetic trajectories that generate the morphological diversity that co-occurs with it. It is known that species with the least developed parental care retain a longer vestigial larval period during and immediately after the transitory interval of mixed feeding, which separates the embryonic period of endogenous nutrition and the juvenile period in which all final adult structures have appeared (Balon, 1979). In fishes with more advanced hiding strategies this transitory vestige of the larval period does not exist and, in these, all juvenile characters develop while feeding is still endogenous (Balon, 1979). With the advancement of hiding strategies, the transition from endo- to exogenous feeding is more sudden, and the interval of mixed feeding is very brief or nonexistent (Balon, 1979). It seems plausible that all these differences occur as a direct result of the increase in egg size, since inadequate provision for embryonic nutrition is the primary factor governing the occurrence of a larval stage (Orton, 1953), or as Balon (1986) puts it: ‘the endogenous food supply is the most important, for it determines and changes the entire remaining life history’. This is clearly exemplified in my comparison of ontogeny in Corydoras aeneus and Ancistrus cf. triradiatus under 5.1. 134 5 GENERAL DISCUSSION In conclusion, in comparison to other loricarioid families, parental care and sexual dimorphism are clearly more common in the Loricariidae, which confirms that a substantial degree of ‘communicative divergence’ with subsequent sexual selection has occurred in loricariid radiation. In addition to this, the increase in egg size has led to a prolongation of the embryonic period and a loss of the larval stage in, at least some loricariiids, altering the entire ontogenetic pathway and lowering the number of crucial thresholds throughout ontogeny (see 5.1). Three stages All elements presented above clearly show that, as suggested by Streelman & Danley (2003) for all vertebrate radiations, radiation in this example has also followed an evolutionary trajectory along axes of habitat divergence, trophic morphology and communication. It has also become apparent that the part of this divergence which follows models of sexual selection through sexual dimorphism, parental care and the probable co-evolution of egg size has played a much more crucial role in the radiation of the Loricariidae than expected. The order in which all three phase occur, however, remains unclear. In the cases of trophic morphology and communicative divergence, both paths seem to have coevolved. On the one hand parental care and the suggested co-evolution of egg size would have delayed yolk depletion and eliminated the larval stage, allowing the necessary time for the development of new structures related to feeding morphology. On the other hand, the decoupling of various structures related to feeding has also led to a decoupling of the opercular apparatus, resulting in a sexually dimorphic mechanism (used for nest guarding), which, in turn would have led to the evolution of extensive parental care and larger egg size. Therefore, I believe that the evolutionary basis of radiation in the Loricariidae has been established by various events of vicariance throughout the South American history but that a co-evolution of decoupling events in trophic morphology and sexual dimorphism with related parental care and increase in egg size have ultimately led to the extreme diversity found in only the loricariid family. 135 5 GENERAL DISCUSSION 136 Chapter 6 Summary & Samenvatting 6 SUMMARY & SAMENVATTING 6.1 SUMMARY This doctoral thesis provides a detailed description of the adult morphology of Corydoras aeneus combined with several aspects of the species’ ontogeny. For comparison, similar results were gathered from a miniaturized congeneric, Corydoras pygmaeus and results were further compared to a description of the ontogeny of the loricariid Ancistrus cf. triradiatus (Geerinckx, 2006). All results are subsequently discussed in the context of the evolution of algae scraping in the loricarioid superfamily and adaptive radiation in the Loricariidae. In this summary I list a brief recapitulation of the main results and conclusions per chapter. 1. In the introduction to this dissertation, the example of adaptive radiation in the family of the Loricariidae is presented. These Loricariidae belong to the superfamily of the Loricarioidea, a superfamily in which the various families that belong to it exhibit a trend of increasing morphological complexity in which various elements have been decoupled, creating opportunities for the evolution of new structures and functions (Schaefer & Lauder, 1986, 1996). This way, in members of the more basal loricarioid family of the Callichthyidae a suckermouth or algae scraping apparatus is still lacking; but the mouth already has a ventral position. In the family of the Astroblepidae, such a suckermouth is present, but no algivory is known and finally, in the family of the Loricariidae, both a suckermouth and algae scraping feeding apparatus are often found. In this context, the aims of this research were to study the morphology and early ontogeny of a basal representative of this evolutionary lineage and compare the results found to those of a highly specialized species. 2. In the second chapter, the materials and methods used throughout this dissertation are listed. First, the choice of C. aeneus as the main research object is discussed, giving arguments that the family the species belongs to takes a basal position within the lineage leading up to the highly specialized morphology in the Loricariidae. Within the Callichthyidae, the choice of genus and species is explained in relation to the amount of data already known from 137 6 SUMMARY & SAMENVATTING existing literature. Further, breeding and collection protocols and lists of all specimens collected this way are presented. In a second part of this chapter, the different methods use are given and, when appropriate, detailed protocols are included. The different methods explained include: in toto clearing and staining, dissecting, serial sectioning and the generation of 3D-reconstructions. 3. The third chapter describes the adult morphology of C. aeneus and is divided into two main parts. A first part deals with the osteology of the species in which results are compared to the condition described for other loricarioid fishes. Most results are found to correspond with the expected pattern, although some differences are observed. One of these differences is the presence of a suprapreopercular bone, a bone that was previously unmentioned in any callichthyid species. Further, several homologies are resolved and mainly confirm existing hypotheses. This way, for example, the compound dorsolateral bone of the otic region is identified as the posttemporo-pterotico-supracleithrum while the compound bone forming the neurocranial floor in the occipital region is recognized as a fusion of both basi- and exoccipital bones. This part of the study, however, still fails to resolve the homology of the lacrymo-antorbital and suprapreopercular bones. In a second part, the musculature of the species is described and discussed. Given the importance of the jaw and opercular muscles in the context of decoupling and the acquisition of new functions in relation to the specialized feeding apparatus of Loricariidae, homologies of these bundles are discussed in detail. In the case of the jaw musculature of adult C. aeneus, an A1OST, A2, and A3’ section is identified and arguments supporting the homology of the m. retractor tentaculi with the A3” are presented. In the opercular region, as is the case in A. cf. triradiatus, the m. dilatator operculi is enlarged and the direction of the operculo-hyomandibular articulation has shifted. In addition, in both species, the m. hyohyoidei abductor has also shifted its orientation, acting as an adductor. 4. The third chapter deals with several aspects of the ontogeny of both C. aeneus and C. pygmaeus. 138 6 SUMMARY & SAMENVATTING In the first part, the surface structure of the eggs of C. aeneus is described. This showed to be a unique pattern among teleosts. The surface is covered with small protuberances, which resemble attaching-filaments of teleost eggs. In addition, eggs were also found to be very adhesive, which probably relates to the fact that the species is known to inhabit turbid waters. The second part of this chapter presents an overview of the early development and the growth pattern of C. aeneus. Based on external morphology, the different stages in early development, as described by Balon (1975), are identified. After hatching, from a SL of 3.5 mm, an eleutherembryonic phase is present, followed by the protopterygiolarval phase (4.4-5.7 mm SL), the pterygiolarval phase (5.7-14.0 mm SL) and the juvenile period. In addition, an overall growth curve is presented in which inflexion points were determined, applying a piecewise linear regression method, which follows the ideas of regression spline smoothing procedures. This way, the growth curve was divided into six different stages of growth rate. Initially, the slope is 0.05 until 0.7 dph, then increasing to 0.18 until 4 dph, and 0.36 until 10 dph. After this, growth rate reaches a maximum of 0.76 until 24 dph, slows down to 0.47 until 37 dph and then finally again slows down to 0.36. A similar analysis was also done on the data of growth in the different body parts. The inflexion points found this way match the different key-events known in teleost early development. Such events are: the transition from endo- to exogenous feeding, the transition to the pterygiolarval phase, when priorities shift toward locomotory needs and the transition to a carangiform swimming mode at approximately 8 mm SL. In a third part of this chapter on ontogeny, the development of the cranium is described. For the study of the chondrocranium, serial sections were digitized and used for 3D-reconstructions. Development overall follows the typical siluriform trends in chondrocranial development. Even the low complexity of the chondrocranium at hatching fits the trend observed in other siluriforms, although other studies show loricarioid hatchlings to generally have more complex chondrocrania. As in A. cf. triradiatus, a true commissura lateralis is present, which differs from the situation described for other siluriforms. The most striking differences found in comparison to other catfishes, however, involve the branchial basket, which arises as a single element in C. aeneus, with a further differentiation from the middle arches on in both a rostral and caudal direction. With the description of development of the cranial bones, this part continues to 139 6 SUMMARY & SAMENVATTING describe the development of the cranium. Development of this osteocranium is found to occur in two distinct phases. In a first phase, several new bony elements, all of dermal origin and related to feeding, appear shortly after yolk depletion (4.4 mm SL). After this, in between 5 and 8 mm SL, developmental priorities shift to size increase of the cartilaginous skull and no new bony elements appear. Finally, a second phase of osteogenesis is present from 8-18 mm SL, in which all remaining dermal and perichondral bones appear. In the fourth part, ontogeny of the cranial muscle system is dealt with. Comparison of these results to those of A. cf. triradiatus reveals a high degree of similarity. This way, both species lack a m. protractor hyoidei, and the m. intermandibularis posterior is divided into two different parts which have partly obtained a novel function in A. cf. triradiatus. A similar increase in complexity in this species is found in the dorsal constrictor of the hyoid muscle plate. This constrictor gives rise to the same muscles as in C. aeneus, but, in A. cf. triradiatus, the m. levator and dilatator operculi later become hypertrophied. In addition, in A. cf. triradiatus, the m. extensor tentaculi further differentiates into two separate bundles (as opposed to a single muscle diverging posteriorly in C. aeneus) and a loricariid neoformation is present called the m. levator tentaculi (Geerinckx et al., subm.). In contrast to the former three parts in this chapter, the fifth part does not describe cranial ontogeny in C. aeneus, but focuses on the development of the postcranial skeleton. Results indicate a possible shift in swimming mode at two points during ontogeny. Observations on the ontogeny of the caudal fin skeleton show that all of the cartilaginous elements in this fin develop between a SL of c. 5–6 mm, at which point also notochord flexion occurs, which corresponds to a suggested change to burst swimming at 5 mm SL (Weihs, 1980; Verhagen, 2004). At 8 mm SL, the anal, adipose and pelvic fins separate from the larval finfold and ossification in the vertebral column and caudal skeleton starts. Also, development of the anal and pelvic fin starts and ossification of the scapulocoracoid plate and development of the proximal radials and fin spine in the pectoral fin takes place. This all corresponds to a second shift in swimming mode suggested to occur at that point, based on the analysis of early development. The sixth and final part of this chapter gives a brief overview of similar results in a miniaturized congeneric species, Corydoras pygmaeus. Results show high 140 6 SUMMARY & SAMENVATTING similarities in the overall pattern of development in both species with some difference in the timing of appearance of ossification of different elements. In conclusion, miniaturization in the body of C. pygmaeus was found to be the result of paedomorphic events, including a lower growth rate and subsequent postdisplacement of several traits. These paedomorphic events have, however, besides the smaller size, not led to the retention of a large number of paedomorphic traits. 5. In the fifth chapter, the results of this dissertation are discussed in relation to the evolution within the Loricarioidea. In the first part, the ontogenies of C. aeneus, C. pygmaeus and A. cf. triradiatus are briefly recapitulated and the theory of saltatory ontogeny, as presented by Balon (1959, 1975, 1979, 1990) is applied to these results. This shows that the ontogenies of both Corydoras species follow a saltatory path, altering stabilized periods of quantitative morphogenesis and growth with brief but distinct qualitative changes. The situation in A. cf. triradiatus, however, substantially differs from this, showing no saltatory development. This was related to the larger egg size (and the evolution of parental care) and a subsequently longer prehatching period, which also probably led to the skipping of a true larval stage during the species’ ontogeny. All this is hypothesized to have ultimately facilitated the evolution of a highly specialized feeding apparatus, given the increase in the time available for development of this specialized morphology through a delay in yolk depletion. In the second part, results presented in this dissertation and additional results from literature are combined to apply the theory of Streelman & Danley (2003) to the evolution of algae scraping and subsequent (?) adaptive radiation in Loricarioidea in general and Loricariidae specifically. This leads to the conclusion that the evolutionary basis of radiation in the Loricariidae was established by various events of vicariance throughout the South American geological history. In addition, a co-evolution of decoupling events in trophic morphology and sexual dimorphism (with related parental care and increase in egg size) have ultimately led to the extreme diversity found in the loricariid family only. 141 6 SUMMARY & SAMENVATTING 142 6 SUMMARY & SAMENVATTING 6.2 SAMENVATTING Deze doctoraatsthesis geeft een gedetailleerde beschrijving van zowel de adulte morfologie als een aantal aspecten van de ontogenie van Corydoras aeneus. Ter vergelijking werd een summiere vergelijkbare studie van een geminiaturiseerde soort uit hetzelfde genus, Corydoras pygmaeus, uitgevoerd. De gevonden resultaten werden verder vergeleken met resultaten uit een gelijkaardige studie uitgevoerd op de loricariide Ancistrus cf. triradiatus door Geerinckx (2006). Vervolgens werden alle resultaten bediscussieerd in de context van de evolutie van het algenschrapen in de superfamilie van de Loricarioidea en van de adaptieve radiatie binnen de Loricariidae. In deze samenvatting wordt een korte recapitulatie van de belangrijkste resultaten en conclusies per hoofdstuk gepresenteerd. 1. In de inleiding van deze verhandeling wordt het voorbeeld van de adaptieve radiatie binnen de familie van de Loricariidae voorgesteld. Deze Loricariidae behoren tot de superfamilie van de Loricarioidea, een superfamilie waarin de verschillende families die er deel van uitmaken een trend naar een stijgende morfologische complexiteit vertonen, waarbij verschillende elementen werden ontkoppeld, daarbij de gelegenheid tot de evolutie van nieuwe structuren en functies creërend (Schaefer & Lauder, 1986, 1996). In vertegenwoordigers van de meer basale loricarioide familie Callichthyidae is een zuigmond of algenschrapend voedingsmechanisme afwezig; maar de mond bevindt zich wel reeds ventraal. Bij de Astroblepidae is zo’n zuigmond wel aanwezig is, hoewel geen algivorie gekend is en uiteindelijk zijn bij vertegenwoordigers van de Loricariidae zowel een zuigmond als vaak ook algivorie aanwezig. In deze context was het doel van deze studie het bestuderen van de morfologie en vroege ontogenie van een basale vertegenwoordiger van deze evolutieve lijn en om de resultaten te vergelijken met deze van een zeer gespecialiseerde soort. 2. In het tweede hoofdstuk worden het materiaal en de gebruikte methoden opgesomd. In eerste instantie wordt de keuze van C. aeneus als studieobject beargumenteerd, waarbij wordt aangehaald dat de familie waartoe deze soort behoort een basale positie inneemt binnen de lijn naar een 143 6 SUMMARY & SAMENVATTING gespecialiseerde morfologie binnen de Loricarioidea. Binnen deze familie wordt de keuze van het genus en de soort geplaatst binnen de context van de mate waarin literatuurgegevens voorhanden zijn. Vervolgens worden de verschillende protocols m.b.t. het kweken en verzamelen van specimens weergegeven, aangevuld met een lijst van de gebruikte specimens. In een tweede deel van dit hoofdstuk worden de verschillende methoden uiteengezet, en, waar aangewezen, worden gedetailleerd protocols meegegeven. De verschillende methoden gebruikt zijn: opheldering en kleuring, dissecties en het maken van seriële coupereeksen en 3D-reconstructies. Het derde hoofdstuk beschrijft de adulte morfologie van C. aeneus 3. en is onderverdeeld in twee delen. Een eerste deel behandelt de osteologie van de soort en resultaten hiervan worden vergeleken met deze van andere loricarioide vissen. Het merendeel van deze resultaten beantwoordt aan het verwachte patroon, hoewel enkele verschillen werden gevonden. Een van deze verschillen was de aanwezigheid van een suprapreoperculair been, voorheen onbeschreven in Callichthyidae. Verder worden verschillende bestaande hypothesen homologieproblemen worden bevestigd. opgehelderd, Zo wordt, waarbij vaak bijvoorbeeld, de het samengesteld dorsolateraal been van de otische regio geïdentificeerd als het posttemporo-pterotico-supracleithrum, terwijl het samengesteld been dat de bodem van het neurocranium in de occipitale regio vormt wordt herkend als een fusie van zowel de basi- als exoccipitale beenderen. De homologie van het veronderstelde lacrimo-antorbitale en suprapreoperculare konden daarentegen in dit deel van de studie niet worden bevestigd. In een tweede deel wordt de musculatuur van de soort beschreven en bediscussieerd. Gezien het belang van kaak- en operculaire spieren in de context van ontkoppeling en het verwerven van nieuwe functies in relatie tot het gespecialiseerde voedingsapparaat bij Loricariidae, werden homologieën van deze spieren in detail bestudeerd. Bij C. aeneus worden zo binnen het kaakspiercomplex A1OST, A2 en A3’ secties herkend en worden argumenten aangereikt die de homologie van de m. retractor tentaculi met de A3’’ ondersteunen. Binnen het operculaire spiercomplex wordt, net zoals bij A. cf. triradiatus, een sterk vergrote m. dilatator operculi aangetroffen en was de oriëntatie van het operculo-hyomandibulair ligament verschoven. Daarnaast 144 6 SUMMARY & SAMENVATTING werd bij beide soorten ook een verschuiving van de m. hyohyoidei abductor vastgesteld, die daardoor als een adductor fungeerde. 4. Het vierde hoofdstuk behandelt verschillende aspecten van de ontogenie van zowel C. aeneus als C. pygmaeus. In het eerste deel wordt de oppervlaktestructuur van de eitjes van C. aeneus beschreven. Daarbij wordt een patroon aangetroffen dat uniek blijkt bij teleosten. Het oppervlak was bezet met kleine uitsteeksels, sterk gelijkend op filamenten die voor de vasthechting zorgen bij de eitjes van andere teleosten. Daarnaast vertonen de eitjes ook een sterk vermogen tot vasthechting, waarschijnlijk gerelateerd aan het feit dat van de soort gekend is dat ze in snelstromend water leeft. Het tweede deel van dit hoofdstuk geeft een overzicht van de vroege ontwikkeling en de groeipatronen bij C. aeneus. Gebaseerd op externe morfologie worden de verschillende stadia, zoals beschreven door Balon (1975), in de vroege ontwikkeling geïdentificeerd. Na kippen, vanaf een SL van 3.5 mm, is een vrijlevende embryonale fase aanwezig, gevolgd door de proto- pterygiolarvale fase (4.4-5.7 mm SL), de pterygiolarvale fase (5.7-14 mm SL) en de juveniele periode. Daarnaast wordt een groeicurve gepresenteerd waarin verschillende inflexiepunten werden gezocht, daarbij gebruik makend van een stapsgewijze lineaire regressiemethode volgens de procedures van ‘regression spline smoothing’. Op deze manier kon de groeicurve in zes fases met verschillende groeisnelheid worden onderverdeeld. Bij aanvang is de hellingshoek in de curve 0.05 tot aan 0.7 dph, waarna deze toeneemt tot 0.18 tot aan 4 dph, en 0.36 tot 10 dph. Hierna wordt een maximum in groeisnelheid van 0.76 bereikt tot aan de leeftijd van 24 dph, die daarna afneemt tot 0.47 tot 37 dph en uiteindelijk nogmaals tot 0.36 daarna. Een gelijkaardige analyse werd ook uitgevoerd op de groei van verschillende lichaamsdelen. De inflexiepunten die op deze manier werden gevonden komen overeen met de verschillende gekende ‘key-events’ in de vroege ontwikkeling van teleosten. Voorbeelden daarvan zijn: de transitie van endo- naar exogene voeding, de overgang naar de pterygiolarvale fase, waarbij prioriteiten verschuiven naar locomotie toe en de transitie naar een carangiforme zwemstijl rond c. 8 mm SL. In een derde deel van dit hoofdstuk over ontogenie wordt de ontwikkeling van het cranium beschreven. Voor de studie van het chondrocranium werden seriële 145 6 SUMMARY & SAMENVATTING coupereeksen gedigitaliseerd en voor 3D-reconstructies gebruikt. De ontwikkeling volgt de typisch siluriforme trends in chondrocranium ontwikkeling. Zelfs de geringe complexiteit van het chondrocranium op het moment van hatching past in de algemene trend die bij katvissen wordt waargenomen, alhoewel andere studies bij loricarioide hatchlings in de regel een meer complex chondrocranium aantonen. Zoals bij A. cf. triradiatus wordt ook hier een echte commissura lateralis gevonden, wat verschilt van de situatie zoals die bij andere Siluriformes wordt beschreven. Het meest markante verschil in vergelijking met andere katvissen, daarentegen, heeft betrekking op de kieuwkorf, die als een enkelvoudig element ontstaat bij C. aeneus, met een verdere differentiatie vanuit de middelste bogen in zowel een rostrale als caudale richting. Met de beschrijving van de ontwikkeling van de craniale beenderen, vervolgt dit hoofdstuk de beschrijving van de ontwikkeling van het cranium. De ontwikkeling van dit osteocranium blijkt in twee verschillende fasen te verlopen. In een eerste fase ontwikkelden zich, kort na het verdwijnen van de dooier (4.4 mm SL), verschillende beenderen die allen van dermale origine en aan voeding gerelateerd zijn. Na dit, tussen 5 en 8 mm SL, vershuiven de prioriteiten in de ontwikkeling in de richting van groei van de kraakbeenschedel en worden geen nieuwe benige elementen gevormd. Uiteindelijk is een tweede fase van osteogenese aanwezig in specimens tussen 8 en 18 mm SL, waarbij alle overige dermale en perichondrale beenderen worden gevormd. In het vierde deel wordt de ontogenie van de craniale spieren behandeld. Een vergelijking van deze resultaten met die van A. cf. triradiatus toont een grote mate van gelijkenis. Zo ontbreekt bij beide soorten de m. protractor hyoidei en is de m. intermandibularis posterior onderverdeeld in twee verschillende delen die in A. cf. triradiatus deels een nieuwe functie hebben aangenomen. Een vergelijkbare verhoging van de complexiteit in deze soort wordt ook teruggevonden in de dorsale constrictor van de hyoid spierplaat. Uit deze constrictor worden dezelfde spieren gevormd als bij C. aeneus, waarbij, in een latere fase bij A. cf. triradiatus, de m. levator en dilatator operculi hypertrofieert. Daarnaast differentieert de m. extensor tentaculi bij A. cf. triradiatus verder in twee aparte bundels (in tegenstelling tot een enkelvoudige spier die achteraan vertakt bij C. aeneus) en is ook een loricariide neoformatie aanwezig die m. levator tentaculi wordt genoemd (Geerinckx et al., subm.). 146 6 SUMMARY & SAMENVATTING In tegenstelling tot de vorige drie delen uit dit hoofdstuk worden in het vijfde deel geen elementen van de craniale ontogenie van C. aeneus behandeld, maar wordt er gefocused op de ontwikkeling van het postcraniale skelet. Resultaten hiervan wijzen in de richting van een mogelijke verschuiving in zwemstijl op twee punten gedurende de ontogenie. Observaties met betrekking tot de ontogenie van het staartskelet tonen aan dat alle kraakbenige elementen in deze fin zich ontwikkelen bij een SL van c. 5-6 mm, hetzelfde punt als waarbij flexie in het notochord optreedt, wat overeenkomt met een gesuggereerde verandering naar ‘burst swimming’ bij 5 mm SL (Weihs, 1980; Verhagen, 2004). Bij 8 mm SL, scheiden zowel de anale, vet- en pelvische vinnen zich af van de larvale vin en start de verbening in de wervelkolom. Daarnaast worden de eerste tekenen van verbening gevonden in de scapulocoracoidplaat en beginnen de proximale radialen en vinstekel in de borstvin zich te ontwikkelen. Dit alles komt overeen met een tweede verschuiving in zwemstijl rond dit punt, die wordt gesuggereerd op basis van de analyse van de vroege ontwikkeling. Het zesde en laatste deel van dit hoofdstuk behandelt een kort overzicht van een vergelijkbare studie bij een miniature soort uit hetzelfde genus, C. pygmaeus. Resultaten hiervan tonen een hoge mate van overeenkomst in het ontwikkelingspatroon van beide soorten, met enkele verschillen in de timing van het verschijnen van verschillende elementen. Als besluit kan worden gesteld dat miniaturisatie in het lichaam van C. pygmaeus het resultaat is van paedomorfe processen, waaronder een displacement van lagere groeisnelheid verschillende kenmerken. en daaruitvolgende post- Deze paedomorfiën hebben weliswaar, behalve de kleinere afmetingen, niet geleid tot de retentie van een hoog aantal paedomorfe kenmerken. 5. In het vijfde hoofdstuk worden de resultaten van deze verhandeling bediscussieerd in relatie tot de evolutie binnen Loricarioidea. In een eerste deel worden de ontogenetische patronen van zowel C. aeneus, C. pygmaeus en A. cf. triradiatus kort hernomen en wordt de theorie van de ‘saltatory ontogeny’ van Balon (1959, 1975, 1979, 1990) toegepast op deze resultaten. Dit toont aan dat een dergelijke ‘saltatory ontogeny’ inderdaad aanwezig was bij beide Corydoras-soorten, waarbij stabiele periodes van kwantitatieve morfogenese worden afgewisseld met korte maar distincte kwalitatieve veranderingen. De situatie in A. cf. triradiatus, daarentegen, 147 6 SUMMARY & SAMENVATTING verschilt hier in grote mate van, gezien hier geen ‘saltatory’ ontogenie wordt waargenomen. Dit wordt gerelateerd aan een groter ei (en de evolutie van broedzorg) en een daaruitvolgende langere periode voorafgaand aan kippen, wat daarnaast ook leidde tot het overslaan van een echt larvaal stadium bij de soort. Hieruit volgt de hypothese dat al deze verschillen uiteindelijk de evolutie van een zeer gespecialiseerd voedingsapparaat hebben gefaciliteerd, gezien de verhoging van de tijd die beschikbaar is voor de ontwikkeling van deze structuren door het uitstellen van de dooierabsorptie. In een tweede deel worden de resultaten van deze verhandeling en bijkomende resultaten uit de literatuur gebundeld in een toepassing van de theorie van Streelman & Danley (2003) op de evolutie van het algenschrapen en daaruitvolgende (?) adaptieve radiatie bij Loricarioidea in het algemeen en meer specifiek bij de Loricariidae. De algemene conclusie hierbij is dat de evolutieve basis van de radiatie bij de Loricariidae moet worden gezocht in de verschillende gebeurtenissen die tot periodes van vicariantie hebben geleid gedurende de Zuid-Amerikaanse geologische geschiedenis. Daarbij komt dat een co-evolutie van ontkoppeling van structuren gerelateerd aan de voedingsmorfologie en het seksueel dimorfisme (met de daaraan gerelateerde broedzorg en grotere dooieromvang) uiteindelijk tot de extreme diversiteit, die enkel bij Loricariidae wordt gevonden, hebben geleid. 148 7 References 7 REFERENCES 7. REFERENCES Adriaens D. 2003. Feeding mechanisms. In: Arratia G, Kapoor AS, Chardon M, Diogo R, editors. Catfishes, vol. 1. Enfield, NH: Science Publishers. p 221-248. Adriaens D, Verraes W. 1994. On the functional significance of the loss of the interhyal during ontogeny in Clarias gariepinus Burchell, 1822 (Teleostei, Siluroidei). Belg J Zool 124:139–155. Adriaens D, Verraes W. 1996. Ontogeny of cranial musculature in Clarias gariepinus (Siluroidei: Clariidae): The adductor mandibulae complex. J Morph 229:255-269. Adriaens D, Verraes W. 1997a. Ontogeny of the maxillary barbel muscles in Clarias gariepinus (Siluroidei: Clariidae), with some notes on the palatine-maxillary mechanism. J Zool 241:117-133. Adriaens D, Verraes W. 1997b. The ontogeny of the chondrocranium in Clarias gariepinus: trends in siluroids. J Fish Biol 50:1221–1257. Adriaens D, Verraes W. 1997c. Ontogeny of the hyoid musculature in the African catfish, Clarias gariepinus (Burchell, 1822) (Siluroidei: Clariidae). Zool J Linn Soc 121:105– 128. Adriaens D, Verraes W. 1997d. Ontogeny of the suspensorial and opercular muscles in Clarias gariepinus (Siluroidei: Clariidae), and the consequences for respiratory movements. Neth J Zool 47:61-89. Adriaens D, Verraes W. 1998. Ontogeny of the Osteocranium in the African Catfish, Clarias gariepinus Burchell (1822) (Siluriformes : Clariidae): Ossification sequence as a response to functional demands. J Morph 235:183–237. Adriaens D, Verraes W. 2002. An empirical approach to study the relation between ontogeny, size and age using geometric morphometrics. In: Aerts P, D’Août K, Herrel A, Van Damme R, editors. Topics in functional and ecological vertebrate morphology. Maastricht: Shaker Publishing. p 293-324. Adriaens D, Vandewalle P. 2003. Embryonic and larval development in catfishes. In: Arratia G, Kapoor AS, Chardon M, Diogo R, editors. Catfishes, vol. 1. Enfield, NH: Science Publishers. p 639–666. Adriaens D, Verraes W, Taverne L. 1997. The cranial lateral-line system in Clarias gariepinus (Burchell, 1822) (Siluroidei: Clariidae): morphology and development of canal related bones. Eur J Morphol 35:181-208. Adriaens D, Aerts P, Verraes W. 2001. Ontogenetic shift in mouth opening mechanisms in a catfish (Clariidae, Siluriformes): a response to increasing functional demands. J Morphol 247:197-216. Alberch P, Gouls SJ, Oster GF, Wake DB. 1979. Size and shape in ontogeny and phylogeny. Paleobiology 5:296-317. Albertson RC, Streelman JT, Kocher TD. 2003. Directional selection has shaped the oral jaws of lake malawi cichlid fishes. Proc Natl Acad Sci USA 100:5252-5257. Albertson RC, Streelman JT, Kocher TD, Yelick PC. 2005. Integration and evolution of the cichlid mandible: the molecular basis of alternate feeding strategies. Proc Natl Acad Sci USA 102:16287-16292. Alexander RM. 1965. Structure and function in the catfish. J Zool 148:88-152. 149 7 REFERENCES Alfaro M, Westneat MW. 1999. Motor patterns of herbivorous feeding: electromyographic analysis of biting in the parrotfishes Cetoscarus bicolor and Scarus iserti. Brain Behav Evol 54:205–222 Allis E. 1897. The cranial muscles and cranial and first spinal nerves in Amia calva. J Morph 12:487-809. Andrade DV, Abe AS. 1997. Foam nest production in the armoured catfish. J Fish Biol 50:665-667. Aquino AE. 1994. Secondary sexual dimorphism of the dermal skeleton in two species of the hypoptopomatine genus Otocinclus (Siluriformes: Loricariidae). Ichthyol Explor Freshw 5:217-222. Aquino AE, Schaefer SA. 2002. Revision of Oxyropsis Eigenmann and Eigenmann, 1889 (Siluriformes, Loricariidae). Copeia 2002:374-390. Aranha JMR, Takeuti DF, Yoshimura TM. 1998. Habitat use and food partitioning of the fishes in a coastal stream of Atlantic forest, Brazil. Rev Biol Trop 46:951-959. Araujo-Lima CARM. 1994. Egg size and larval development in Central Amazonian fish. J Fish Biol 44:371-389. Armbruster JW. 2004. Phylogenetic relationships of the suckermouth armoured catfishes (Loricariidae) with emphasis on the Hypostominae and the Ancistrinae. Zool J Linn Soc 141:1-80. Arratia G. 1983. The caudal skeleton of ostariophysan fishes (Teleostei): intraspecific variation in Trichomycteridae (Siluriformes). J Morph 177:213-229. Arratia G. 1987. Description of the primitive family Diplomystidae (Siluriformes, Teleostei, Pisces): morphology, taxonomy and phylogenetic implications. Bonn Zool Monogr 24:11-20. Arratia G. 1990. Development and diversity of the suspensorium of trichomycterids and comparison with loricarioids (Teleostei: Siluriformes). J Morphol 205:193-218. Arratia G. 1992. Development and variation of the suspensorium of the primitive catfishes (Teleostei: Ostariophysi) and their phylogenetic relationships. Bonn Zool Monogr 32:1-148. Arratia G. 2003a. Catfish head skeleton: an overview. In: Kapoor AS, Arratia G, Chardon M, Diogo R, editors. Catfishes. Enfield, NH: Science Publishers. p 3-46. Arratia G. 2003b. The Siluriform postcranial skeleton: an overview. In: Kapoor AS, Arratia G, Chardon M, Diogo R, editors. Catfishes. Enfield, NH: Science Publishers. p 121-157. Arratia G, Schultze H-P. 1990. The urohyal: development and homology within osteichthyans. J Morphol 203:247-282. Arratia G, Huaquín L. 1995. Morphology of the lateral line system and of the skin of diplomystid and certain primitive loricarioid catfishes and systematic and ecological considerations. Bonn Zool Monogr 36:5-110. Bailey RM, Baskin JN. 1976. Scoloplax dicra, a new armored catfish from the Bolivian Amazon. Occ Pap Mus Zool, University of Michigan 674. Bailey K, Houde ED. 1989. Predation on eggs and larvae of marine fish and the recruitment problem. Adv Mar Biol 25:1–83. Ballantyne FM. 1930. Notes on the development of Callichthys littoralis. Trans R Soc Edinb 56:437–466. Balon EK. 1959. Die embryonale und larvale Entwicklung der Donauzope (Abramis ballerus subsp.). Biologické práce 5:1-87. 150 7 REFERENCES Balon EK. 1971. The intervals of fish development and their terminology (a review and proposals). Věst Čs Spol Zool 35:1-8. Balon EK. 1975. Terminology of intervals in fish development. J Fish Res Board Can 32:1663-1670. Balon EK. 1979. The theory of saltation and its application in the ontogeny of fishes: steps and thresholds. Environ Biol Fish 4:97-101. Balon EK. 1984. Reflections of some decisive events in the early life of fishes. T Am Fish Soc 113:178-185. Balon EK. 1986. Types of feeding in the ontogeny of fishes and the life-history model. Environ Biol Fish 16:11-24. Balon EK. 1990. Epigenesis of an epigeneticist: the development of some alternative concepts on the early ontogeny and evolution of fishes. Guelph Ichthyol Rev 1:1-42. Balon EK. 1999. Alternative ways to become a juvenile or a definitive phenotype (and on some persisting linguistic offences). Environ Biol Fish 56:17-38. Bamford TW. 1948. Cranial development of Galeichthys felis. Proc Zool Soc Lond 118:364-391. Baskin JN. 1972. Structure and relationships of the Trichomycteridae. New York: Unpublished PhD thesis. Bellwood DR. 1994. A phylogenetic study of the parrotfishes, family Scaridae (Pisces: Labroidei), with a revision of genera. Rec Aust Mus 20:1–86. Bellwood DR, Choat JH. 1990. A functional analysis of grazing in parrotfishes (family Scaridae): the ecological implications. Environ Biol Fish 28:189-214. Berra TM. 2001. Freshwater fish distribution. London: Academic Press. Bertmar G. 1959. On the ontogeny of the chondral skull in Characidae, with a discussion on the chondrocranial base and the visceral chondrocranium in fishes. Acta Zool 40:203–364. Bhargava HN. 1958. The development of the chondrocranium of Mastacembelus armatus (Cuv. and Val.). J Morph 102:401–426. Blaxter JHS. 1986. The development of sense organs and behavior in teleost larvae with special reference to feeding and predator avoidance. T Am Fish Soc 115:98-114. Bock WJ, Shear CR. 1972. A staining method for gross dissection of vertebrate muscles. Anat Anz 130:222-227. Britto MR. 1998. Two new species of the genus Aspidoras (Siluriformes: Callichthyidae) from Central Brazil. Ichthyol Explor Freshw 8:359-368. Britto MR. 2000. Aspidoras depinnai (Siluriformes: Callichthyidae): a new species from northeastern Brazil. Copeia 2000:1048-1055. Britto MR. 2003. Phylogeny of the subfamily Corydoradinae Hoedeman, 1952 (Siluriformes: Callichthyidae), with a definition of its genera. Proc Acad Nat Sci Phil 153:119-154. Britto MR, Castro RMC. 2002. New corydoradine catfish (Siluriformes: Callichthyidae) from the Upper Paraná and Săo Francisco: the sister group of Brochis and most of Corydoras species. Copeia 2002:1006-1015. Buckel JA, Steinberg ND, Conover DO. 1995. Effects of temperature, salinity, and fish size on growth and consumption of juvenile bluefish. J Fish Biol 47:697-706. Burgess WE, Quinn J. 1992. Colored atlas of miniature catfish. New Yersey: T.F.H. Publications. 151 7 REFERENCES Chardon M, Parmentier E, Vandewalle P. 2002. Morphology, development and evolution of the Weberian apparatus in catfish. In: Kapoor AS, Arratia G, Chardon M, Diogo R, editors. Catfishes. Enfield, NH: Science Publishers. p 71–119. Coburn MM, Grubach PG. 1998. Ontogeny of the Weberian apparatus in the armored catfish (Siluriformes: Callichthyidae). Copeia 1998:301-311. Covain R, Fisch-Muller S. 2007. The genera of the Neotropical armored catfish subfamily Loricariinae (Siluriformes: Loricariidae): a practical key and synopsis. Zootaxa 1462:1-40. Coyne JA, Orr HA. 2004. Speciation. Sunderland, MA: Sinauer Associates. Daget J. 1964. Le crâne des téléostéens, I. Origine et mise en place des tissus squelettogènes chez l’embryon. Mésomésenchyme et ectomésenchyme. Mém Mus Natl Hist Nat Sér A Zool 31:164-340. Darwin C. 1859. On the origin of species by means of natural selection, or preservation fo favored races in the struggle for life. London: John Murray. Darwin C. 1871. The descent of man and selection in relation to sex. New York: Random House, The Modern Library. de Beer GR. 1927. The early development of the chondrocranium of Salmo fario. Q J Microsc Sci 71:259–312. de Beer GR. 1937. Studies on the vertebrate head. Oxford: Clarendon Press. Delariva RL, Agostinho AA. 2001. Relationship between morphology and diets of six neotropical loricariids. J Fish Biol 58:832-847. de Pinna MCC. 1993. Higher-level phylogeny of Siluriformes, with a new classification of the Order (Teleostei, Ostariophysi). New York: Unpublished PhD thesis. de Pinna MCC. 1998. Phylogenetic relationships of Neotropical Siluriformes (Teleostei: Ostariophysi): historical overview and synthesis of hypotheses. In: Malabarba LR, Reis RE, Vari RP, Lucena ZM, Lucena CAS, editors. Phylogeny and classification of Neotropical fishes. Porto Alegre: Edipucrs. p 279-330. de Pinna MCC. 2003. Family Nematogenyidae. In: Reis RE, Kullander SO, Ferraris CJ, editors. Check list of the freshwater fishes of South and Central America. Porto Alegre: Edipucrs. p 268-269. de Pinna MCC, Wosiacki W. 2003. Family Trichomycteridae. In: Reis RE, Kullander SO, Ferraris CJ, editors. Check list of the freshwater fishes of South and Central America. Porto Alegre: Edipucrs. p 270-290. de Pinna MCC, Ng HH. 2004. The second ural centrum in siluriformes and its implication for the monophyly of superfamily sisoroidea (Teleostei, Ostariophysi). Am Mus Novit 3437:1–23. De Schepper N, Adriaens D, Teugels GG, Devaere S, Verraes W. 2004. Intraspecific variation in the postcranial skeleton morphology in African clariids: a case study of extreme phenotypic plasticity. Zool J Linn Soc 140:437–446. Diogo R. 2005. Morphological evolution, aptations, homoplasies, constraints and evolutionary trends. Catfishes as a case study on general phylogeny and macroevolution. Plymouth: Science Publishers. Diogo R, Chardon M. 2000. Homologies among different adductor mandibulae sections of teleostean fishes, with special regard to catfishes (Teleostei: Siluriformes). J Morph 243:193-208. Diogo R, Vandewalle P. 2003. Catfishes as a case study for discussions on general evolution: the importance of functional uncouplings in morphological macroevolution. Eur J Morph 41:139-148. 152 7 REFERENCES Diogo R, Oliveira C, Chardon M. 2001. On the osteology and myology of catfish pectoral girdle, with a reflection on catfish (Teleostei: Siluriformes) plesiomorphies. J Morphol 249:100-125. Diogo R, Chardon M, Vandewalle P. 2006. On the osteology and myology of the cephalic region and pectoral girdle of Nematogenys inermis (Ghichenot, 1848), with comments on the autapomorphies and phylogenetic relationships of the Nematogenyidae (Teleostei : Siluriformes). Belg J Zool 136:15-24. Edgeworth FH. 1928. The development of some of the cranial muscles of ganoid fishes. Phil Trans R Soc (B) 217:39-89. Edgeworth FH. 1935. The cranial muscles of vertebrates. Cambridge: University Press. Eschmeyer WN. 1998. Catalog of fishes. San Francisco: California Academy of Sciences. Faustino M, Power DM. 2001. Osteologic development of the viscerocranial skeleton in sea bream: alternative ossification strategies in teleost fish. J Fish Biol 58:537–572. Ferraris CJ. 2007. Checklist of catfishes, recent and fossil (Osteichthyes: Siluriformes), and catalogue of siluriform primary types. Auckland: Magnolia Press. Fink SV, Fink WL. 1981. Interrelationships of the ostariophysan fishes (Teleostei). Zool J Linn Soc 72:297-353. Fink SV, Fink WL. 1996. Interrelationships of ostariophysan fishes (Teleostei). In: Stiassny MLJ, Parenti LR, Johnson GD, editors. Interrelationships of fishes. London: Academic Press. p 209-249. Friedman JH, Silverman BW. 1989. Flexible parsimonious smoothing and additive modeling. Technometrics 31:3-21. Friel JP, Vigliotta TR. 2006. Synodontis acanthoperca, a new species from the Ogôoué river system, Gabon with comments on spiny ornamentation and sexual dimorphism in mochokid catfishes (Siluriformes: Mochokidae). Zootaxa 1125:45-56. Froese R, Pauly D. 2006. Fishbase. www.fishbase.org. Fuiman LA. 1983. Growth gradients in fish larvae. J Fish Biol 23:117-123. Fuiman LA, Poling KR, Higgs DM. 1998. Quantifying developmental progress for comparative studies of larval fishes. Copeia 1998:602-611. Fujita K. 1992. Ontogeny of the caudal skeleton in the clariid catfish Clarias batrachus. Jpn J Ichthyol 38:430–432. Fukuhara O. 1992. Study on the development of functional morphology and behaviour of the larvae of eight commercially valuable teleost fishes. Contrib Fish Res Japan Sea Block 25:1–122. Fuller IAM. 2001. Breeding corydoradine catfishes. Kidderminster: Ian Fuller Enterprises. Futuyma DJ, Mayer GC. 1980. Non-allopatric speciation in animals. Syst Zool 29:254271. Gavaia PJ, Sarasquete C, Cancele ML. 2000. Detection of mineralized structures in early stages of development of marine Teleostei using a modified alcian blue-alizarin red double staining technique for bone and cartilage. Biotech Histochem 75:79-84. Geerinckx T. 2006. Ontogeny and functional morphology of a highly specialized trophic apparatus: a case study of neotropical suckermouth armoured catfishes (Loricariidae, Siluriformes). Unpublished PhD-thesis, Ghent University. Geerinckx T, Adriaens D. 2006. The erectile cheek-spine apparatus in the bristlenose catfish Ancistrus (Loricariidae, Siluriformes), and its relation to the formation of a secondary skull roof. Zoology 109:287-299. Geerinckx T, Adriaens D. 2007. Ontogeny of the intermandibular and hyoid musculature 153 7 REFERENCES in the suckermouth armoured catfish Siluriformes). Anim Biol 57:339-357. Ancistrus cf. triradiatus (Loricariidae, Geerinckx T, Adriaens D. Ontogeny of the suspensorial and opercular musculature in the suckermouth armoured catfish Ancistrus cf. triradiatus (Loricariidae, Siluriformes). Zoomorphology in press. Geerinckx T, Brunain M, Adriaens D. 2005. Development of the chondrocranium in the suckermouth armored catfish Ancistrus cf. triradiatus (Loricariidae, Siluriformes). J Morphol 266:331–355. Geerinckx T, Brunain M, Adriaens D. 2007. Development of the osteocranium in the suckermouth armored catfish Ancistrus cf. triradiatus (Loricariidae, Siluriformes). J Morph 268:254-274. Geerinckx T, Huysentruyt F, Adriaens D. Ontogeny of the jaw and maxillary barbel musculature in the armoured catfish families Loricariidae and Callichthyidae (Loricarioidea, Siluriformes), with a discussion on muscle homologies. Zool J Linn Soc in press. Gill TN. 1858. Synopsis of the fresh water fishes of the western portion of the island of Trinidad, W I. Ann Lyc Nat Hist N Y 363-430 Gisbert E. 1999. Early development and allometric growth patterns in Siberian sturgeon and their ecological significance. J Fish Biol 54:852-862. Goodrich ES. 1958. The skull. In: Goodrich ES, editor. Studies on the structure and development of vertebrates. New York: Dover Publications. p 230–395. Gosline WA. 1940. A revision of the neotropical catfishes of the family Callichthyidae. Stanf Ichthyol Bull 2:1-29. Gosline WA. 1975. The cyprinid dermosphenotic and the subfamily Rasborinae. Occas Pap Mus Zool, Univ Michig 673:1-13. Gosline WA. 1986. Jaw muscle configuration in some higher teleostean fishes. Copeia 1986:705-713. Gosline WA. 1989. Two patterns of differentiation in the jaw musculature of teleostean fishes. J Zool 218:649-661. Gosline WA. 1993. A survey of upper jaw musculature in higher teleostean fishes. Occas Pap Mus Zool, Univ Michig 724:1-26. Grande L. 1987. Redescription of †Hypsidoris farsonensis (Teleostei: Siluriformes), with a reassessment of its phylogenetic relationships. J Vert Paleontol 7:24-54. Grande L, de Pinna MCC. 1998. Description of a second species of the catfish †Hypsidoris and a reevaluation of the genus and the family †Hypsidoridae. J Vert Paleontol 18:451-474. Grandel H, Schulte-Merker S. 1998. The development of the paired fins in the Zebrafish (Danio rerio). Mech Develop 79:99-120. Grant PR. 2000. R.C.L. Perkins and evolutionary radiations on islands. Oikos 89:195– 201. Grant PR, Grant BR. 2002. Unpredictable evolution in a 30-year study of Darwin's finches. Science 296:707-711. Haffer J. 1982. General aspects of the refuge theory. In: Prance GT, editor. Biological Diversification in the Tropics. New York, NY: Colombia University Press. p 6-24. Hanken J. 1993. Adaptation of bone growth to miniaturization of body size. In: Hall BK, editor. Bone, Vol. 7: Bone growth – B. Boca Raton, FL: CRC Press. p 74-104. 154 7 REFERENCES Hanken J, Wassersug R. 1981. The visible skeleton. A new double-stain technique reveals the native of the "hard" tissues. Funct Photogr 16:22-26. Hanken J, Wake DB. 1993. Miniaturization of body size: organismal consequences and evolutionary significance. Annu Rev Ecol Syst 24:501-519. Harrington RWJ. 1955. The osteocranium of the American cyprinid fish, Notropis bifrenatus, with an annotated synonymy of teleost skull bones. Copeia 4:267–291. Hoedeman JJ. 1960a. Studies on callichthyid fishes (3): Notes on the development of Callichthys (1) (Pisces: Siluriformes). Bull Aquat Biol 1:53-68. Hoedeman JJ. 1960b. Studies on callichthyid fishes: (4) Development of the skull of Callichthys and Hoplosternum (1) (Pisces - Siluriformes). Bull Aquat Biol 1:73–84. Hoedeman JJ. 1960c. Studies on callichthyid fishes: (5) Development of the skull of Callichthys and Hoplosternum (2) (Pisces - Siluriformes). Bull Aquat Biol 2:21-36. Hoedeman JJ. 1960d. Studies on callichthyid fishes (6): The axial skeleton of Callichthys and Hoplosternum (Pisces: Siluriformes). Bull Aquat Biol 2:37-44. Hostache G, Mol JH. 1998. Reproductive biology of the neotropical armoured catfish Hoplosternum littorale (Siluriformes – Callichthyidae): a synthesis stressing the role of the floating bubble nest. Aquat Living Resour 11:173-185. Howes GJ. 1983. The cranial muscles of loricarioid catfishes, their homologies and value as taxonomic characters. Bull Bri Mus Nat Hist 45:309-345. Howes GJ, Teugels GG. 1989. Observations on the ontogeny and homology of the pterygoid bones in Corydoras paleatus and some other catfishes. J Zool 219:441-456. Hurley IA, Mueller RL, Dunn KA, Schmidt EJ, Friedman M, Ho RK, Prince VE, Yang ZH, Thomas MG, Coates MI. 2007. A new time-scale for ray-finned fish evolution. Proc Roy Soc B-Biol Sci 1609:489-498. Huysseune A. 1989. Morphogenetic aspects of the pharyngeal jaws and neurocranial apophysis in postembryonic Astatotilapia elegans (Trewavas, 1933) (Teleostei: Cichlidae). Mededelingen van de Koninklijke Academie voor Wetenschappen, Letteren en Schone Kunsten van België 51:11–35. Huysseune A, Sire J-Y. 1997. Structure and development of teeth in three armoured catfish, Corydoras aeneus, C. arcuatus and Hoplosternum littorale (Siluriformes, Callichthyidae). Acta Zool 78:69-84. Isbrücker IJH. 1981. Revision of Loricaria Linnaeus, 1758 (Pisces, Siluriformes, Loricariidae). Beaufortia 31:51-96. Isbrücker IJH, Nijssen H. 1973. Two new species of the callichthyid catfish genus Corydoras from Brazil (Pisces, Siluriformes, Callichthyidae). Beaufortia 21:1-7. Isbrücker IJH, Nijssen H. 1992a. Corydoras breei, a new species of callichthyid catfish from the Corantijn River basin in Surinam (Pisces, Siluriformes, Callichthyidae). Beaufortia 43:9-14. Isbrücker IJH, Nijssen H. 1992b. Sexualdimorphismus bei Harnischwelsen (Loricariidae). Odontoden, Zähne, Lippen, Tentakel, Genitalpapillen und Flossen. DATZ-Sonderheft (Harnischwelse) 19-33. Jardine D, Litvak MK. 2003. Direct yolk sac volume manipulation of zebrafish embryos and the relationship between offspring size and yolk sac volume. J Fish Biol 63:388– 397. Jarvik E. 1980. Basic structure and evolution of Vertebrates. vol. 1. London: Academic Press. 155 7 REFERENCES Kavanagh KD, Alford RA. 2003. Sensory and skeletal development and growth in relation to the duration of the embryonic and larval stages in damselfishes (Pomacentridae). Biol J Linn Soc 80:187-206. Kim I, Park J. 1996. Adhesive membranes of oocyte in four loaches (Pisces: Cobitidae) of Korea. Korean J Zool 39:198-206. Kindred JE. 1919. The skull of Amiurus. Ill Biol Monogr 5:1–120. Klingenberg CP. 1998. Heterochrony and allometry: the analysis of evolutionary change in ontogeny. Biol Rev 73:79-123. Kobayakawa M. 1992. Comparative morphology and development of bony elements in the head region in three species of Japanese catfishes (Silurus: Siluridae; Siluriformes). Japan J Ichthyol 39:25–36. Kohda M, Tanimura M, Kikue-Nakamura M, Yamagishi S. 1995. Sperm drinking by female catfishes: a novel mode of insemination. Environ Biol Fish 42:1-6. Kohda M, Yonebayashi K, Nakamura M, Ohnishi N, Seki S, Takahashi D, Takeyama T. 2002. Male reproductive success in a promiscuous armoured catfish Corydoras aeneus (Callichthyidae). Environ Biol Fish 63:281-287. Kolm N, Ahnesjö I. 2005. Do egg size and parental care coevolve in fishes? J Fish Biol 66:1499-1515. Koumoundouros G, Divanach P, Kentouri M. 1999. Ontogeny and allometric plasticity of Dentex dentex (Osteichthyes: Sparidae) in rearing conditions. Mar Biol 135:561–572. Koumoundouros G, Divanach P, Kentouri M. 2001a. Osteological development of Dentex dentex (Osteichthyes: Sparidae): dorsal, anal, paired fins and squamation. Mar Biol 138:399-406. Koumoundouros G, Sfakianakis DG, Maingot E, Divanach P, Kentouri M. 2001b. Osteological development of the vertebral column and of the fins in Diplodus sargus (Teleostei: Perciformes: Sparidae). Mar Biol 139:853-862. Kováč V, Copp GH. 1999. Prelude: looking at early development in fishes. Environ Biol Fish 56:7-14. Kramer DL, McClure M. 1980. Aerial respiration in the catfish, Corydoras aeneus (Callichthyidae). Can J Zool 58:1984-1991. Kramer DL, McClure M. 1981. The transit cost of aerial respiration in the catfish Corydoras aeneus (Callichthyidae). Physiol Zool 54:189-194. Kramer DL, Braun EA. 1983. Short-term effects of food availability on air-breathing frequency in the fish Corydoras aeneus (Callichthyidae). Can J Zool 61:1964-1967. Kryzhanovsky SG, Disler NN, Smirnova EN. 1953. Ecomorphological principles of development in percids.Trudy Instituta Morphologii Zhivotnych Severtsova 10:3-138. Laale HW. 1980. The perivitelline space and egg envelopes of bony fishes: a review. Copeia 1980:210-226. Lack D. 1947. Darwin’s Finches. Cambridge: University Press. Lakshmi K, Srinivasa Rao K. 1989. Postcranial skeleton of the marine catfish Arius tenuispinis (Osteichthyes: Siluriformes, Ariidae). J Morph 202:361-377. Langille RM, Hall BK. 1987. Development of the head skeleton of the Japanese Medaka, Oryzas latipes (Teleostei). J Morphol 193:135–187. Lassala MDP, Renesto E. 2007. Reproductive strategies and genetic variability in tropical freshwater fish. Genet Mol Biol 30:690-697. Lauder GV. 1981. Form and function: structural analysis in evolutionary morphology. Paleobiology 7:430-442. 156 7 REFERENCES Lauder GV, Crompton AW, Gans C, Hanken J, Liem KF, Maier WP, Meyer A, Presley R, Rieppel OC, Roth G, Schluter D, Zweers GA. 1989. Group report: Howe are feeding systems integrated and how have evolutionary innovations been introduced? In: Wake DB, Roth G, editors. Complex and organismal functions: integration and evolution in vertebrates. New York: Wiley & Sons. p 97-115. Lee TCM. 2002. On algorithms for ordinary least squares regression spline fitting: a comparative study. J Stat Comput Sim 72:647-663. Lewis LM, Lall SP. 2006. Development of the axial skeleton and skeletal abnormalities of Atlantic halibut (Hippoglossus hippoglossus) from first feeding through metamorphosis. Aquaculture 257:124–135. Liem KF. 1973. Evolutionary strategies and morphological innovations: cichlid pharyngeal jaws. Syst Zool 22:425-441. Losos JB, Leal M, Glor RE, de Queiroz K, Hertz PE, Rodriguez Schettino L, Chamizo Lara A, Jackman TR, Larson A. 2003. Niche lability in the evolution of a Caribbean lizard community. Nature 424:542-545. Lubosch W. 1938. Muskeln des Kopfes: Viscerale Muskulatur (Fortsetzung). B. Teleostier. In: Bolk L, Göppert E, Kallius E, Lubosch W, editors. Handbuch der vergleichenden Anatomie der Wirbeltiere. Berlin & Wien: Urban & Schwarzenberg. p 1011-1024. Lujan NK, Armbruster JW, Sabaj M. 2007. Two new species of Pseudancistrus from southern Venezuela (Siluriformes: Loricariidae). Ichthyol Explor Freshwaters 18:163174. Lundberg JG. 1975. Homologies of the upper shoulder girdle and temporal region bones in catfishes (Order Siluriformes), with comments on the skull of the Helogeneidae. Copeia 1975:66-74. Lundberg JG. 1982. The comparative anatomy of the toothless blindcat, Trogloglanis pattersoni Eigenmann, with a phylogenetic analysis of the ictalurid catfishes. Misc Publ Mus Zool, Univ Michig 163:1-85. Lundberg JG. 1998. The temporal context for the diversification of Neotropical fishes. In: Malabarba LR, Reis RE, Vari RP, Lucena ZM, Lucena CAS, editors. Phylogeny and classification of Neotropical fishes. Porto Alegre: Edipucrs. p 49-68. Lundberg JG, Baskin JN. 1969. The caudal skeleton of the catfishes, order Siluriformes. Am Mus Nov 2398:1-19. Lundberg JG, Marshall LG, Guerrero J, Horton B, Malabarba MCSL, Wesselingh F. 1998. The stage for Neotropical fish diversification: a history of tropical South American rivers. In: Malabarba LR, Reis RE, Vari RP, Lucena ZM, Lucena CAS, editors. Phylogeny and classification of Neotropical fishes. Porto Alegre: Edipucrs. p 13-48. Lundberg JG, Sullivan JP, Rodiles-Hernández R, Hendrickson DA. 2007. Discovery of African roots for the Mesoamerican Chiapas catfish, Lacantunia enigmatica, requires an ancient intercontinental passage. Proc Acad Nat Sci Phila 156:39-53. Mabee PM, Trendler TA. 1996. Development of the cranium and paired fins in Betta splendens (Teleostei: Percomorpha): Intraspecific variation and interspecific comparisons. J Morph 227:249-287. Machado-Allison A, Garcia C. 1986. Food habits and morphological changes during ontogeny in three serrasalmin fish species of the Venezuelan floodplains. Copeia 1986:193-196. Maisey JG. 2000. Continental break up and the distribution of fishes of Western Gondwana during the Early Creataceous. Creataceous Res 21:281-314. Mayr E. 1996. What is a species and what is not? Philos Sci 63:626-277. Mayr E. 2001. What evolution is. New York: Basic Books. 157 7 REFERENCES Mazzoni R, Caramaschi EP. 1995. Length and weight relationship and condition factor of two species of Hypostomus Lacépède, 1803 (Siluroidei, Loricariidae) in the lower Paraíba de Sul River, Rio de Janeiro, Brazil. Arq Biol Tecnol 38:135-140. McElman JF, Balon EK. 1980. Early ontogeny of white sucker, Catostomus commersoni, with steps of saltatory development. Env Biol Fish 5:191-224. Mo T. 1991. Anatomy and systematics of Bagridae (Teleostei), and siluroid phylogeny. Koenigstein: Koeltz Scientific Books. Mol JH. 1996. Reproductive seasonality and nest-stie differentiation in three closely related armoured catfishes (Siluriformes: Callichthyidae). Environ Biol Fish 45:363381. Montoya-Burgos JI. 2003. Historical biogeography of the catfish genus Hypostomus (Siluriformes: Loricariidae), with implications on the diversification of Neotropical ichthyofauna. Mol Ecol 12:1855-1867. Morin RP, Able KW. 1983. Patterns of geographic variation in the egg morphology of the fundulid fish, Fundulus heteroclitus. Copeia 1983:726-740. Myers GS. 1953. Habits and classification of Corydoras hastatus. The Aquarium Journal 24:268-270. Nelson JA. 2002. Metabolism of three species of herbivorous loricariid catfishes: influence of size and diet. J Fish Biol 61:1586-1599. Nelson JS. 2006. Fishes of the world. 4th ed. Hoboken, New Jersey: John Wiley & Sons. Ng HH, Bailey RM. 2006. Chiloglanis productus, a new species of suckermouth catfish (Siluriformes: Mochokidae) from Zambia. Occas Pap Univ Mich Mus Zool 738:1-13. Nijssen H. 1970. Revision of the Surinam catfishes of the genus Corydoras Lacepède, 1803 (Pisces, Siluriformes, Callichthyidae). Beaufortia 18:1-75. Nijssen H, Issbrücker IJH. 1980. A review of the genus Corydoras Lacépède, 1803 (Pisces, Siluriformes, Callichthyidae). Bijdr Dierk 50:190-220. Oliveira C, Almeida-Toledo LF, Mori L, Toledo-Filho SA. 1992. Extensive chromosomal rearrangements and nuclear DNA content changes in the evolution of the armoured catfishes genus Corydoras (Pisces, Siluriformes, Callichthyidae). J Fish Biol 40:419– 431. Oliveira C, Almeida-Toledo LF, Mori L, Toledo-Filho SA. 1993. Cytogenetic and DNA content studies of armoured catfishes of the genus Corydoras (Pisces, Siluriformes, Callichthyidae) from the southeast coast of Brazil. Rev Bras Genét 16:617–629. Orton GL. 1953. The systematics of vertebrate larva. Syst Zool 2:63-75. Osse JWM. 1990. Form changes in fish larvae in relation to changing demands of function. Neth J Zool 40:362-385. Osse JWM, van den Boogaart JGM. 1995. Fish larvae, development, allometric growth, and the aquatic environment. ICES Marine Science Symposia 201:21-34. Osse JWM, van den Boogaart JGM, van Snik GMJ, van der Sluys L. 1997. Priorities during early growth of fish larvae. Aquaculture 155:249-258. Park J, Kim I. 2001. Fine structure of the oocyte envelopes of three related cobitid species in the genus Ikookimia (Cobitidae). Ichthyol Res 48:71-75. Paugy D, Roberts TR. 1992. Mochokidae. In: Lévêque C, Paugy D, Teugels GG, editors. Faune des poissons d’eaux douces et saumâtres de l’Afrique de l’Ouest. Tervuren: MRAC. Paris: ORSTOM. p 500-563. Pascal M, Hostache G, Tessier C, Vallat P. 1994. Cycle de reproduction et fécondité de l’atipa, Hoplosternum littorale (Siluriforme) en Guyane Française. Aquat Living Resour 158 7 REFERENCES 7:25-37. Patzner RA, Glechner R. 1996. Attaching structures in eggs of native fishes. Limnologica 26:179-182. Pavlov DA, Moksness E. 1997. Development of the axial skeleton in wolffish, Anarhichas lupus (Pisces, Anarhichadidae), at different temperatures. Environ Biol Fish 49:401– 416. Pedersen BH, Ugelstad I, Hjelmeland K. 1990. Effects of a transistory, low food supply in the early life of larval herring (Clupea harengus) on mortality, growth and digestive capacity. Mar Biol 107:61–66. Pruzsinszky I, Ladich F. 1998. Sound production and reproductive behaviour of the armoured catfish Corydoras paleatus (Callichthyidae). Environ Biol Fish 53:183-191. Ramnarine IW. 1994. Larval culture, development and growth of the cascadu, Hoplosternum littorale (Hancock 1828; Callichthyidae). Aquaculture 126:291-298. Ramnarine IW. 1995. Induction of nest building and spawning in Hoplosternum littorale. J Fish Biol 47:555-557. Rapp Py-Daniel L. 2000. Tracking evolutionary trends in Siluriformes (Teleostei; Ostariophysi). In: International congress on the biology of fish, Aberdeen. Evolution Symposium Proceedings. University of Aberdeen: Aberdeen. p 57-72. Regan CT. 1911. The classification of the teleostean fishes of the order Ostariophysi: 2. Siluroidea. Ann Mag Nat Hist 8:553-577. Reis RE. 1996. Callichthyidae. The tree of life web project. www.tolweb.org. Reis RE. 1997. Revision of the neotropical catfish genus Hoplosternum (Ostariophysi: Siluriformes: Callichthyidae), with the description of two new genera and three new species. Ichthyol Explor Freshw 7:299-326. Reis RE. 1998. Anatomy and phylogenetic analysis of the neotropical callichthyid catfishes (Ostariophysi, Siluriformes). Zool J Linn Soc 124:105-168. Reis RE. 2003. Family Callichthyidae. In: Reis RE, Kullander SO, Ferraris CJ, editors. Check list of the freshwater fishes of South and Central America. Edipucrs: Porto Alegre. p 291-309. Reis RE, Schaefer SA. 1992. Eurycheilus pantherinus (Siluroidei: Loricariidae), a new genus and species of Hypoptopomatinae from Southern Brazil. Copeia 1:215-223. Retzer ME, Page LM. 1997. Systematics of the stick catfishes, Farlowella Eigenmann & Eigenmann (Pisces, Loricariidae). Proc Acad Nat Sci Phila 147:33-88. Riehl R. 1996. The ecological significance of the egg envelope in teleosts with special reference to limnic species. Limnologica 26:183-189. Riehl R. 1999. The micropyle of teleost fish eggs: morphological and functional aspects. Proceedings of the 5th Indo-Pacific Fish Conference, Nouméa 589-599. Riehl R, Götting KJ. 1974. Zu Struktur und Vorkommen der Mikropyle an Eizellen und Eiern von Knochenfischen (Teleostei). Arch Hydrobiol 74:393-402. Riehl R, Bless R. 1995. First report on egg deposition and egg morphology of the endangered endemic Romanian perch. J Fish Biol 46:1086-1090. Riehl R, Patzner RA. 1998. The modes of egg attachment in teleost fishes. Ital J Zool 65:415-420. Ristovska M, Karaman B, Verraes W, Adriaens D. 2006. Early development of the chondrocranium in Salmo letnica (Karaman, 1924) (Teleostei: Salmonidae). J Fish Biol 68:458-480. Rizzo E, Sato Y, Barreto BP, Godinho HP. 2002. Adhesiveness and surface patterns of 159 7 REFERENCES eggs in neotropical freshwater teleosts. J Fish Biol 61:615-632. Rodríguez MS, Miquelarena A. 2005. A new species of Rineloricaria (Siluriformes: Loricariidae) from the Paraná and Uruguay River basins, Misiones, Argentina. Zootaxa 945:1-15. Rojo AL. 1991. Dictionary of evolutionary fish osteology. Florida: CRC Press. Schaefer SA. 1987. Osteology of Hypostomus plecostomus (Linnaeus) with a phylogenetic analysis of the loricariid subfamilies (Pisces: Siluroidei). Contrib Sci 394:1-31. Schaefer SA. 1988. Homology and evolution of the opercular series in the loricarioid catfishes (Pisces: Siluroidei). J Zool 214:81-93. Schaefer SA. 1990. Anatomy and relationships of the scoloplacid catfishes. Proc Acad Nat Sci Phila 142:167-210. Schaefer SA. 1997. The neotropical cascudinhos: systematics and biogeography of the Otocinclus catfishes (Siluriformes: Loricariidae). Proc Acad Nat Sci Phila 148:1-120. Schaefer SA. 2003. Family Astroblepidae. In: Reis RE, Kullander SO, Ferraris CJ, editors. Check list of the freshwater fishes of South and Central America. Porto Alegre: Edipucrs. p 312-317. Schaefer SA, Lauder GV. 1986. Historical transformation of functional design: evolutionary morphology of feeding mechanisms in loricarioid catfishes. Syst Zool 35:489-508. Schaefer SA, Lauder GV. 1996. Testing historical hypotheses of morphological change: biomechanical decoupling in loricarioid catfishes. Evolution 50:1661-1675. Schaefer SA, Aquino AE. 2000. Postotic laterosensory canal and pterotic branch homology in catfishes. J Morphol 246:212-227. Schaefer SA, Weitzman SH, Britski HA. 1989. Review of the Neotropical catfish genus Scoloplax (Pisces: Loricarioidea: Scoloplacidae) with comments on reductive characters in phylogenetic analysis. Proc Acad Nat Sci Phila 141:181-211. Schmidt RE. 2001. Loricaria cataphracta: parental care and description of early larvae. Ichthyol Explor Freshwaters 12:235-240. Seehausen O. 2004. Hybridization and adaptive radiation. Trends Ecol Evol 19:198-207. Sfakianakis DG, Koumoundouros G, Divanach P, Kentouri M. 2004. Osteological development of the vertebral column and of the fins in Pagellus erythrinus (L. 1758). Temperature effect on the developmental plasticity and morpho-anatomical abnormalities. Aquaculture 232:407–424. Shelden FF. 1937. Osteology, myology, and probable evolution of the nematognath pevic girdle. Ann NY Acad Sci 37:1-96. Shiba Y, Sasaki Y, Enomoto T, Kanno Y. 1982. Microtubule formation in regenerating terminal buds of the silurid fish, Corydoras aeneus. Dev Growth Differ 24:199-203. Shimabukuro-Dias CK, Oliveira C, Reis RE, Foresti F. 2004. Molecular phylogeny of the armored catfish family Callichthyidae (Ostariophysi, Siluriformes). Mol Phylogenet Evol 32:152-163. Simonović PD, Garner P, Eastwood EA, Kováč V, Copp G. 1999. Correspondence between ontogenetic shifts in morphology and habitat use in minnow Phoxinus phoxinus. Environ Biol Fish 56:117-128. Simpson GG. 1953. The major features of evolution. New York, Columbia University Press. Sire JY, Huysseune A. 1996. Structure and development of the odontodes in an armoured 160 7 REFERENCES catfish, Corydoras aeneus (Siluriformes, Callichthyidae). Acta Zool 77:51-72. Srinivasachar HR. 1957a. Development of the skull in catfishes: Part I - Development of chondrocranium in Silonia, Pangasius and Ailia (Schilbeidae). P Natl Inst Sci India 22B:335–356. Srinivasachar HR. 1957b. Development of the skull in catfishes: Part II - Development of chondrocranium in Mystus and Rita (Bagridae). Gegenbauers Morphol Jahrb 98:244– 261. Srinivasachar HR. 1958a. Development of the skull in catfishes: Part IV - The development of the chondrocranium in Arius jella Day (Ariidae) and Plotosus canius Ham. (Plotosidae) with an account for their interrelationships. Gegenbauers Morphol Jahrb 99:986-1016. Srinivasachar HR. 1958b. Development of the skull in catfishes: Part V - Development of the skull in Heteropneustes fossilis (Bloch). Proc Natl Inst Sci, India 24B:165-190. Srinivasachar HR. 1959. Development of the skull in catfishes: Part III - The Heteropneustes fossilis (Bloch) development of the chondrocranium in (Heteropneustidae) and Clarias batrachus (Linn) (Clariidae). Gegenbauers Morphol Jahrb 101:373–405. Stern DL, Elmen DJ. 1999. The developmental basis for allometry in insects. Development 126:1091-1101. Strauss RE. 1984. Allometry and functional feeding morphology in haplochromine cichlids. In: Echelle AA, Kornfield I, editors. Evolution of Fish Species Flocks. Maine: Orono Press. p 217-229. Strauss RE. 1985. Evolutionary allometry and variation in body form the South American catfish genus Corydoras (Callichthyidae). Syst Zool 34:381-396. Streelman JT, Danley PD. 2003. The stages of vertebrate evolutionary radiation. Trends Ecol Evol 18:126-131. Streelman JT, Alfaro M, Westneat MW, Bellwood DR, Karl SA. 2002. Evolutionary history of the parrotfishes: biogeography, ecomorphology, and comparative diversity. Evolution 56:961–971. Sullivan JP, Lundberg JG, Hardman M. 2006. A phylogenetic analysis of the major groups of catfishes (Teleostei: Siluriformes) using rag1 and rag2 nuclear gene sequences. Mol Phylogenet Evol 41:636-662. Surlemont C, Vandewalle P. 1991. Développement postembryonnaire du squelette et de la musculature de la tête de Clarias gariepinus (Pisces, Siluriformes) depuis l'eclosion jusqu'à 6,8 mm. Can J Zool 69:1094–1103. Surlemont C, Chardon M, Vandewalle P. 1989. Skeleton, muscles and movements of the head of a 5,2 mm fry of Clarias gariepinus (Burchell) (Pisces Siluriformes). In: Splechtna H, Hilgers H, editors. Trends in vertebrate zoology. Stuttgart: Gustav Fischer Verlag. p 459-462. Suzuki HI, Agostinho AA, Winemiller KO. 2000. Relationship between oocyte morphology and reproductive strategy in loricariid catfishes of the Parana River, Brazil. J Fish Biol 57:791-807. Takahasi, N. 1925. On the homology of the cranial muscles of the cypriniform fishes. J Morphol Physiol 40:1-109. Tamaru CS, Cole B, Bailey R, Brown C. 1997. A manual for commercial production of the tiger barb, Capoeta tertrazona, a temporary paired tank spawner. Hawaï: Center for Tropical and Subtropical Aquaculture. Taylor JN. 1983. Field observations on the reproductive ecology of three species of armored catfishes (Loricariidae: Loricariinae) in Paraguay. Copeia 1:254-259. 161 7 REFERENCES Taylor WR, Van Dyke GC. 1985. Revised procedures for staining and clearing small fishes and other vertebrates for bone and cartilage study. Cybium 9:107–119. Turner GF. 2007. Adaptive radiation of cichlid fish. Curr Biol 17:R827-R831. Tilney RL, Hecht T. 1993. Early ontogeny of Galeichthys feliceps from the south east coast of South Africa. J of Fish Biol 43:183-212. Vandewalle P, Surlemont C, Sanna P, Chardon M. 1985. Interprétation fonctionnelle de modifications du splanchnocrâne pendant le développement post-embryonnaire de Clarias gariepinus (Téléostéens, Siluriformes). Zool Jahrb Anat 113:91–100. Vandewalle P, Lalèyé P, Focant B. 1995. Early development of cephalic bony elements in Chrysichthys auratus (Geoffroy Sint-Hilaire, 1808) (Pisces, Siluriformes, Claroteidae). Belg J Zool 125:329–347. Vandewalle P, Gluckman I, Baras E, Huriaux F, Focant B. 1997. Postembryonic development of the cephalic region in Heterobranchus longifilis. J Fish Biol 50:227– 253. Vandewalle P, Chikou A, Lalèyé P. 1999. Early development of the chondrocranium in Chrysichthys auratus. J Fish Biol 55:795–808. Vandewalle P, Germeau G, Besancenet P, Parmentier E, Baras E. 2005. Early development of the head skeleton in Brycon moorei (Pisces, Ostariophysi, Characidae). J Fish Biol 66:996–1024. Van Snik GMJ, Van Den Boogaart JGM, Osse, JWM. 1997. Larval growth patterns in Cyprinus carpio and Clarias gariepinus with attention to the finfold. J Fish Biol 50:1339–1352. Vasnetsov VV. 1953. Etaps in the development of bony fishes. In: Pavlovsky EN, editor. Otcherky po obshtchim voprosam ichtiologii. Moskva-Leningrad: AN SSSR Press. p 207-217. Verhagen JHG. 2004. Hydrodynamics of burst swimming fish larvae; a conceptual model approach. J Theor Biol 229:235–248. Vetter B. 1878. Untersuchungen zur vergleichenden Anatomie der Kiemen und Kiefermusculatur der Fische. Jena Zeits J Naturwiss 12:431-550. Vigliotta TR. 2007. A phylogenetic study of the mochokidae (Teleostei, Siluriformes). Unpublished Master Thesis, Cornell University. Vilizzi L, Walker KF. 1999. The onset of juvenile period in carp, Cyprinus carpio: a literature survey. Environ Biol Fish 56:93-102. Wagemans F, Vandewalle P. 1999. Development of the cartilaginous skull in Solea solea: trends in pleuronectiforms. Ann Sci Nat 1:39–52. Wagemans F, Vandewalle P. 2001. Development of the bony skull in common sole: brief survey of morpho-functional aspects of ossification sequence. J Fish Biol 59:1350– 1369. Wagemans F, Focant B, Vandewalle P. 1998. Early development of the cephalic skeleton in the turbot. J Fish Biol. 52:166–204. Weihs D. 1980. Energetic significance of changes in swimming modes during growth of larval anchovy, Engraulis mordax. Fish Bull 77:597-604. Weisel GF. 1967. Early ossification in the skeleton of the sucker (Catostomus macrocheilus) and the guppy (Poecilia reticulata). J Morph 121:1–18. Weitzman SH, Vari RP. 1988. Miniaturization in South American freshwater fishes; an overview and discussion. Proc Biol Soc Wash 101:444-465. 162 7 REFERENCES West-Eberhard MJ. 2003. Adaptive radiation. In: West-Eberhard MJ, Developmental plasticity and evolution. Oxford: University Press. p 564-597. editor. Winterbottom R. 1974. A descriptive synonymy of the striated muscles of the Teleostei. Proc Acad Nat Sci Phila 125:225-317. Wright JJ, Page LM. 2006. Taxonomic revision of the lake Tanganyikan Synodontis (Siluriformes: Mochokidae). Bull Fla Mus Nat Hist 46:99-154. Wu K-Y, Shen S-C. 2004. Review of the teleostean adductor mandibulae and its significance to the systematic positions of the Polymyxiiformes, Lampridiformes, and Triacanthoidei. Zool Stud 43:712-736. 163 7 REFERENCES 164 8 Publication List 8 PUBLICATION LIST PUBLICATION LIST Huysentruyt F, Adriaens D, Teugels GG, Devaere S, Herrel A, Verraes W, Aerts P. 2004. Diet composition in relation to morphology in some African anguilliform clariid catfishes. Belg J Zool 134(1):25-30. Devaere S, Teugels GG, Adriaens D, Huysentruyt F, Verraes W. 2004. Redescription of Dolichallabes microphthalmus (Poll, 1942) (Siluriformes, Clariidae). Copeia 1:108-115. Huysentruyt F, Adriaens D. 2005. Adhesive structures in the eggs of Corydoras aeneus (Gill, 1858; Callichthyidae). J Fish Biol 66(3):871-876. Huysentruyt F, Adriaens D. 2005. Descriptive osteology of Corydoras aeneus (Gill, 1858) (Siluriformes: Callichthyidae). Cybium 29(3):261-273. Adriaens D, Geerinckx T, Huysentruyt F, Schaefer SA, Herrel A. 2005. Evolution of trophic specialisations in neotropical catfishes: more than a mouthful. Integr Comp Biol 45(6):955-955. Van Wassenbergh S, Herrel A, Adriaens D, Huysentruyt F, Devaere S, Aerts P. 2006. A catfish that can strike its prey on land. Nature 440:881. Huysentruyt F, Geerinckx T, Adriaens D. 2007. A descriptive myology of Corydoras aeneus (Gill, 1858) (Siluriformes: Callichthyidae), with a brief discussion on adductor mandibulae homologies. Anim Biol 57:433-452. Huysentruyt F, Brunain M, Adriaens D. ———. Ontogeny of the chondrocranium in Corydoras aeneus (Gill, 1858) (Callichthyidae, Siluriformes). J Morph DOI: 10.1002/jmor.10603. Geerinckx T, Huysentruyt F, Adriaens D. ———. Ontogeny of the jaw and maxillary barbel musculature in the armoured catfish families Loricariidae and Callichthyidae (Loricarioidea, Siluriformes), with a discussion on muscle homologies. Zool J Linn Soc in press. Huysentruyt F, Geerinckx T, Brunain M, Adriaens D. ———. Development of the osteocranium in Corydoras aeneus (Gill, 1858) (Callichthyidae, Siluriformes). J Morph submitted. Huysentruyt F, Brunain M, Adriaens D. ———. Ontogeny of the cranial musculature in Corydoras aeneus (Gill, 1858) (Callichthyidae, Siluriformes). J Fish Biol submitted. Huysentruyt F, Moerkerke B, Devaere S, Adriaens D. ———. Early development and allometric growth in the armoured catfish Corydoras aeneus. In prep. Huysentruyt F, Adriaens D. ———. Ontogeny of the postcranial skeleton in Corydoras aeneus (Gill, 1858) (Callichthyidae, Siluriformes). In prep. 165 8 PUBLICATION LIST Ide C, De Puysseleir J, Huysentruyt F, Geerinckx T, Adriaens D. ———. Nonalgae scraping vs algae scraping, comparison between Synodontis and Atopochilus (Siluriformes: Mochokidae). In prep. 166