Botia Gray

Maurice Kottelat, 2004, Botia kubotai, a new species of loach (Teleostei: Cobitidae) from the Ataran River basin (Myanmar), with comments on botiine nomenclature and diagnosis of a new genus., Zootaxa 401, pp. 1-18 : 9-11

[[ Genus Botia Gray ]]

Lineages within Botiinae. The history of the lineages recognised within the subfamily Botiinae has been reviewed by Taki (1972) who concluded in recognising two genera, Leptobotia and Botia , as done previously by Fang (1936) and Nalbant (1963). He also recognised three subgenera within Botia , Botia s.s. , Hymenophysa and Sinibotia .

Nalbant (2002) presented a discussion of botiine relationships. His analysis is marred with semantic, conceptual and nomenclatural shortcomings. It does not seem appropriate to expend further here on the semantic problems, although they probably are the source of a number of misunderstandings. Nevertheless, it seems necessary to explain why I have ignored most of Nalbant's discussion. The nomenclatural problems are dealt with below.

Among the most obvious problems, Nalbant ignores important works on the phylogeny of Cobitoidei (e.g., Sawada, 1982). Cobitoid genera which have unsettled relationships and are potentially very basal in the phylogeny of the suborder are not ( Yunnanilus , Tuberoschistura ) or vaguely mentioned ( Ellopostoma , Serpenticobitis ). Nalbant treats Nemacheilinae as a valid family, but does not say a word of Balitoridae in which it is usually included (and which is a senior synonym). New family-group names appear on plate 7 (Ellopostomatinae, Serpenticobitinae) which are not mentioned in the text.

Nalbant claims to have analysed "more than 30" characters. But he does not make the difference between 'character' and 'character state'. What he lists on pp. 310-311 are at best 38 character states of a much smaller number of characters. On the other hand, the alternative states listed for many characters are not related. The determination of the polarity of the characters is not discussed and in some cases seems axiomatic (see below for Vaillantellinae), which may lead to circularity.

Some characters obviously amalgamate characters/character states which should be analysed independently. For example, 'character' 9 ("both lips simple with no mental barbels ") and 'character' 10 ("both lips papillous or furrowed, with well developed mental lobes, sometimes as button, or a fourth or fifth pair of barbels") include information on the structure of the lip surface (furrowed or papillous), the presence of mental barbels and their development. Does 'simple' [smooth], furrowed or papillous mean three states of the same character or pairs of states of two independent characters (smooth vs. furrowed; smooth vs. papillous)? If the plesiomorphic state is smooth, furrowed and papillous seem to be character states of two different characters (smooth vs. furrowed; smooth vs papillous). The presence or absence of mental lobes "sometimes as buttons or a fourth or fifth pair of barbels" is obviously not part of the same lip structure character(s), but represent one or more additional characters (presence vs. absence of mental barbels; mental barbels reduced to a knob vs. well developed; etc.).

Another lethal flaw is that distribution is used as a phylogenetic character in the analysis. As zoogeography has to be derived from phylogeny, this clearly is circular reasoning. There are no data to support the approximate palaeogeographic scenario spanning pages 312-313.

These comments apply at most levels of the analysis (families, subfamilies, tribes) and the limited examples presented here should be enough to explain why Nalbant's discussion cannot be considered as a base for a systematic discussion and zoogeographic analysis.

Nalbant treats Botiinae as a family including the subfamilies Vaillantellinae and Botiinae. This relationship is based only on symplesiomorphic (thus uninformative) character states. The distribution of apomorphic states is not discussed. The existence of possible synapomorphies linking Vaillantellinae and Balitoridae, or Botiinae and Cobitidae is not discussed. Character states 'defining' Botiidae (sensu Nalbant) in fact are not shared by Vaillantellinae (body is deep vs. very elongated in Vaillantellinae; lateral line complete vs. incomplete; suborbital spine present vs. missing). Nalbant does not list synapomorphies diagnosing Vaillantellinae from Botiinae; he lists some characters for each taxa, but not the same characters for all taxa and he does not indicate the alternative character states; again there is no discussion of these characters in species of Balitoridae, Cobitidae or outgroups. Sawada (1982: 190) discussed relationships of Cobitoidei and showed that Vaillantella shares synapomorphies with Nemacheilinae and none with Botiinae. This work is not mentioned by Nalbant, although some of his figures are based on Sawada's.

Contrary to Nalbant's statements, in Vaillantellinae the rostral barbels are not organised as in Botiinae. As in Nemacheilinae, they are organised in a transverse row; their bases are adjacent but they cannot be described as united; they do not have a conspicuous fleshy muscular support. In Botiinae, the rostral barbels are organised in two rows at the tip of the snout, the two median ones clearly anterior to the lateral ones, separated from them by a groove and with an independent movable, fleshy, muscular support. Vaillantellinae have a conspicuous cephalic lateral line system, as do Nemacheilinae (vs. inconspicuous in Botiinae).

This being said and although they are not supported by clearly formulated apomorphies, the lineages or genera recognised by Taki (1972) and Nalbant (2002) for South and Southeast Asian botiines (Nalbant's Botiini) are intuitively reasonable. The branching, evolutionary theories, etc. may be wrong, but the assemblages of species seem to constitute natural groups.

Taki (1972) had proposed a system of two genera ( Botia , Leptobotia ), with Botia divided into three subgenera, Botia s.s. , Hymenophysa and Sinibotia . Taki went further in recognising two lineages within his Hymenophysa and his text (p. 78) and key (p. 76) make it clear that he was considering B. macracanthus as representing a distinct taxon but he did not name it. Within Hymenophysa , he recognised an hymenophysa group and a modesta group . With the available data, the lineages within Botia (sensu Taki) seem real and supported by the characters listed by Taki. But there is no analysis to support closer relationships between some lineages rathers than others. In Nalbant's 2002 system, Botia (sensu Taki) is divided into four genera: Botia (Taki's Botia s.s. ), Hymenophysa (Taki's hymenophysa group ) and Yasuhikotakia (Taki's modesta group ). The generic position of B. macracanthus is not discussed. Taki's Sinibotia is treated as a valid genus in the tribe Leptobotiini.

The resulting picture is that of several species groups or lineages which can be diagnosed by sets of characters (but of largely undetermined polarity). To recognise some lineages as part of the same subgenus, or as subgenera of the same genus implies that there is a phylogenetic analysis supporting this hierarchy. This analysis is not available.

An arbitrary grouping of some lineages to each other rather than to others does not make sense and the only logical way of treating these lineages pending a phylogenetic analysis is to give them an equal rank. There are three alternatives: to treat them all as genera, or all as subgenera of Botia , or as constituting a single genus. To treat them all as subgenera of Botia results in intercalating a subgeneric name between the generic and specific name, in my view an unnecessarily cumbersome notation, especially when considering that anyway Botia would be the only genus of its subfamily. To treat them as a single genus masks the already identified diversity within the family. Therefore, I treat them as distinct genera. This implies the creation of a new generic name for B. macracanthus which cannot be placed in any of the already named species-groups.

I have not had access to enough material of the East Asian botiines (Leptobotiini of Nalbant) and it would not make sense at this stage to discuss their genera on the sole basis of the literature. I treat here as valid the genera recognized by Chen (1990) and Nalbant (2002). Nevertheless, I note that here again most characters listed are uninformative, ambiguous, ambiguously worded or comparing non comparable states (i.e., body elongated vs. tall, head conical vs. elongated; posterior chamber enlarged vs. free). This does not mean that the genera and species do not exist (published illustrations suggest they do) but that suitable information is missing.