Parental Care in the Climbing Perch (Anabas testudineus): Confusion or Lost Data?

Anabantoid fishes comprise a unique group, most of whose representatives are characterized by diverse forms of parental care. Evolutional reconstruction of the reproductive strategies of anabantoids is challenging, partly due to the lack of reliable evidence of parental care in a number of its representatives, the most famous of which is the climbing perch. Moreover, apparent contradictions between literary sources frequently occur on the subject. This brief overview attempts to analyze the situation and identify the possible causes of disagreement. The further development in understanding the phylogeny of life history strategies in the group, as well as the concept of evolutionary transitions among parental care patterns in fishes in general, depend on the clarification of this situation.


INTRODUCTION
The fish of the suborder Anabantoidei are characterized by a high variety of reproductive strategies, and advanced parental care, provided in various forms, including nest building and mouthbrooding, is typical for most of them. However, there is a small group of anabantoids that lack parental care behavior. One of the most famous of these species is the climbing perch (Anabas testudineus). It is widely distributed in fresh waters of South and Southeast Asia. Its eggs are small and numerous, typical for pelagophiles with no parental care.
Despite the wide distribution of the climbing perch, the natural spawning of these fish has hardly been studied. Although there are no descriptions of parental behavior of the climbing perch in the reported spawnings [1][2][3], a number of fundamental works contain statements (not confirmed by specific data) about the presence of primitive parental care for eggs in this species [4,5].
Phylogenetic analysis does not provide for an unambiguous interpretation of the parental care evolution in anabantoids [6,7]. It is most likely that parental care is either (1) a plesiomorphic trait of this group that was subsequently lost in some clades or (2) it developed three times independently [6]. Climbing perch occupies a basal position in the group in relation to both Ctenopoma species without and Microctenopoma species with parental care. Hence the importance of understanding the reproduction of the climb-ing perch for the further development of phylogenetic reconstructions [8]. In particular, the confirmation of the facultative parental care behavior in this species would be evidence in favor of the first variant of the evolutionary scenario. In turn, for the development of general ideas about the evolution of reproductive strategies in animals, this would be an important and relatively scarce example of the parental care reduction.
Thus, clarification of the situation with conflicting data seems important and relevant. This article presents an attempt to reconstruct the history of reports about the parental care in the climbing perch as well as a brief analysis of the current situation.

PARENTAL CARE IN THE CLIMBING PERCH-HISTORY OF THE IDEA
In the second half of the twentieth century, the famous researcher E.K. Balon proposed and then developed the concept of reproductive guilds in fish [5,[9][10][11]. It has been popular for several decades and is generally based on fairly accurate facts. However, it was the classification that mentioned the parental care in the climbing perch for the first time. In early versions of this classification [9] Balon indicates Anabas spp. as typical representatives of the ecological group of fish that do not build nests and spawn in the water column but care for their eggs with positive buoyancy (ecological group B.1.4, pelagophiles, according to the 1975 version).
One could assume that the author did not mean the climbing perch but some other species since he mentioned Anabas spp., but that would be problematic. The genus Anabas now includes two almost indistinguishable species and the validity of one of them (A. cobojius) is questionable. In the 1970s-1980s some other anabantoids also appeared in literature under this generic name, but most of them were later moved to the genera Ctenopoma and Microctenopoma. Microctenopoma species indeed provide parental care, this is their generic trait. In particular, this is true for the "former Anabas" Microctenopoma congicum and M. lineatum. However, they build typical for many anabantoids bubble nests [12] and therefore do not correspond to the Balon's description of pelagic guarders [9]. In later versions of his classification, Balon not only did not correct this dubious passage, but, on the contrary, indicated a specific species, Anabas testudineus [5].
In 1987, another well-known ichthyologist, R.H. Lowe-McConnell, published a monograph Ecological Studies in Tropical Fish Communities [4]. The book still has a well-deserved reputation among researchers. In one of the chapters, Lowe-McConnell represents the classification of E.K. Balon and also mentioned "some Channa and Anabas species" as an example of fish guarding floating eggs. However, in the other part of this book, the author writes that "very small pelagic eggs are left unguarded in Anabas testudineus and Helostoma temmincki." In other words, the author knows that parental care is unknown in Anabas testudineus but assumes that this type of behavior may exist in some other representatives of this genus.
The situation is similar with the other influential review written by R.L Welcomme [13], in which the author also represents the Balon classification and mentions Anabas spp. as an example of fish guarding floating eggs. This idea spread widely over time in the literature, and got into textbooks [14,15], identification guides [16], and reference books [17,18]. The main known references and even descriptions of the supposed existence of parental care in the climbing perch over the last 50 years are presented in chronological order in Table 1.
Where could Balon get the information about the climbing perch? In the very first version of his classification [9], he cites the work by G.W. Barlow et al. [19], which states that anabantoids exhibit a wide range of forms of parental care and, "In some, such as the genus Anabas testudineus guards eggs none article [18] Anabas testudineus guards eggs at the surface of hypoxic waters [5] assessment of invasive risk [16] Anabas testudineus parents guard the eggs until they hatch none identification guide Anabas, no bubble nest is provided; the eggs simply float to the surface where they disperse." This sentence is rather ambiguous and likely to cause misinterpretation. It's unclear what anabantoid species is mentioned, and, on top of that, since it comes to the diversity of parental care, it is not clear, whether the parents still guard the floating eggs or not. Balon probably interprets this explicitly and writes the following about the climbing perch eggs: "Deposited on the surface of the water in openings between vegetation growing in the shallows, they float in a cluster; and propelled by wind, they often drift into remote places. Both the male and female guard the eggs and larvae" [9]. Thus, he reproduces Barlow's information and even supplements it with details, the origin of which is not reported. Barlow, in turn, also did not apparently see the spawning of the climbing perch and described it with reference "Kühme, personal communication to George Webber Barlow." Most probably, he means the famous German ethologist W. Kühme.
Thus, in brief, the basic sequence of events was obviously as follows: Kühme told Barlow something about some Anabas. Barlow mentioned this in his article on other fish of the genus Badis. Balon read this article and included this data into his reproductive guild classification. A few years later Balon began to write not Anabas spp., but Anabas testudineus. The Balon's classification, together with the supposed parental care of the climbing perch, spread in scientific literature, and finally got into manuals, handbooks, identification guides, and university textbooks in the 21st century.
Getting back to the very beginning of this story: what did Kühme actually observe and what did he tell Barlow? Kühme was a true ethologist; his observational skills and precision are beyond any doubt. He indeed investigated the reproductive behavior of various anabantoids and published several works on this issue, in particular an extremely detailed and thorough comparison of two species of the genus Betta, one of which incubates eggs in the mouth and the other builds a bubble nest [20]. It is unlikely that Kühme gave his colleague information he was not sure about. George Barlow is also a respected expert on fish behavior. Who made a mistake? At what point? And was there a mistake?

METHODOLOGICAL FRAMEWORK AS A POSSIBLE CAUSE OF CONTRADICTIONS
Both the aims and the methods of contemporary studies on the reproduction of the climbing perch differ significantly from those used in the middle of the 20th century. In the first 20 years of the current century, approximately 8000 publications related to the biology of the climbing perch, or at least mentioned have already been published [21]. However, none of the known studies attempted to create conditions for performance of the natural reproductive behavior of these fish. Moreover, behavior in general is rarely the subject of modern research of this kind, and the protocol of these few studies is often unsuitable. For example, it has been shown that such studies can lead to incorrect results because they attempt to investigate the behavior of schooling animals in solitary test conditions [22]. The reproductive behavior of fish is also very sensitive to conditions, and its study in the laboratory can lead to the appearance of various artifacts [23].
A particular problem is the use of hormones for induced spawning. Indeed, breeding climbing perches in aquariums is a specific methodological task and often requires long, time-consuming preparation. However, it seems that the use of hormonal stimulation can make behavioral observations completely meaningless. It has been shown, for example, that even the use of the same synthetic hormone in different breeding sets has a different effect on the reproductive behavior of climbing perch [24]. Only a few descriptions of spawning of these fish carried out in aquariums without the use of artificial stimulation are known [1,3], and there is no published description of its spawning in nature.
One of the reasons for use of most simplified research methods is the widespread but erroneous idea of the reproductive strategy of fish as a conservative species-specific complex of traits. This implies the assumption that spawning and post-spawning behaviors change little under different conditions. In fact, numerous studies have shown that this is not true. The reproductive strategy of many species is characterized by high plasticity, which is the basis for various situationally dependent alternative breeding tactics [25,26]. Even a complete termination of parental care, socalled brood desertion, is a frequent variant of alternative tactics in many animals [27,28]. Finally, for a number of lower vertebrates, including fish, facultative parental care has been confirmed, the providing or absence of which depends on a set of factors [29,30].
The life history traits of the climbing perch in general correspond to the opportunistic strategy [31], with no parental care. However, there are indications that facultative parental care in the climbing perch is quite possible [8]. A similar situation is typical for snakeheads (family Channidae) that belong to the same order Anabantiformes and pursue a related reproductive strategy, including small floating eggs, rapid development, etc. However, these fish almost always provide parental care [32,33], but this behavior is sometimes completely absent [34]. Returning to the problem of hormonal spawning stimulation, it is important to emphasize that it can result in the loss of parental care in snakeheads [34].
Thus, we cannot exclude the possibility that unpublished observations of Kühme allowed him to MOSCOW  ACKNOWLEDGMENTS I thank Đinh Thị H i Y n and Vo Thị Ha for long-term collaboration in the study of the climbing perch and D.S. Pavlov, K.F. Dzerzhinskiy, and an anonymous reviewer for fruitful discussion of certain theses of this manuscript.

FUNDING
This study was performed within the framework of the Ecolan E-3.2 program of the Vietnam-Russian Tropical Research and Technological Centre.

COMPLIANCE WITH ETHICAL STANDARDS
The authors declare that they have no conflict of interest. This article does not contain any studies involving animals or human participants performed by any of the authors.