Introduction

Recent studies on the evolutionary history of lasiurine bats (Chiroptera: Vespertilionidae: Lasiurini; ^{Baird et al. 2015}, ²⁰¹⁷) have spurred discussion in the mammal literature regarding the broader implications of taxonomic revisions to long-standing nomenclature. ^{Baird et al. (2015}, ²⁰¹⁷) recommended the separation of lasiurines into three genera: Lasiurus (red bats), Aeorestes (hoary bats), and Dasypterus (yellow bats). Until ^{Baird et al. (2015}, ²⁰¹⁷), the vespertilionid tribe Lasiurini had been considered monotypic, comprised solely of the monophyletic genus, Lasiurus. At the time, some authors also recognized two subgenera: Dasypterus (yellow bats) and Lasiurus (red + hoary bats; Figure 1A). In the more distant past, other authors recognized each of these groups (red, hoary, and yellow bats) as separate genera. Throughout the taxonomic history of these bats, their status has been in flux (Figure 2). The purpose of this paper is to review the relevant literature regarding generic and subgeneric taxonomy within the Lasiurini and to address the concerns expressed by ^{Ziegler et al. (2016}), ^{Novaes et al. (2018}), and ^{Teta (2019}).

Methods

We reviewed the literature beginning with ^{Tate (1942}) who first recognized the Tribe Lasiurini and included the bats commonly referred to as red bats, hoary bats, and yellow bats. The literature of this group is extensive, but we restrict our assessment to 13 papers that we consider most influential for taxonomy (Figure 2), represent all the various taxonomic proposals, and are illustrative of the numerous changes, back and forth, between recognizing one or two genera over the course of nearly 80 years. We use this information to address the criticisms of our proposed arrangement of three genera of lasiurine bats (^{Baird et al. 2015}, ²⁰¹⁷) by ^{Ziegler et al. (2016}), ^{Novaes et al. (2018}), and ^{Teta (2019}).

Additionally, times to most recent common ancestor (TMRCA) for vespertilionid bat groups were estimated from the data provided in ^{Amador et al. (2016}). We recorded all estimated TMRCA for vespertilionid genera, tribes, and subfamilies for those taxa that were monophyletic. Two exceptions were Hypsugo/Falistrellus and Eptesicus/Histiotus. Genera represented by a single specimen in ^{Amador et al. (2016)} were not included in our analysis. Finally, dates for the TMRCA within Lasiurini were obtained from the estimates of ^{Baird et al. (2017}) because ^{Amador et al. (2016)} did not include multiple specimens of all lineages within Lasiurini. Those dates were not all specified in the ^{Baird et al. (2017)} paper, but they were extracted from the original analysis. The date for Lasiurini and the monotypic Lasiurus were included in the dates obtained from ^{Amador et al. (2016)}. We sorted the TMRCA for each taxon by date and plotted them as a histogram using R.

Results

Phylogenetic relationships of the major groups of Lasiurini and the various generic and subgeneric taxonomic arrangements are shown in Figure 1. The taxonomic changes proposed by ^{Baird et al. (2015}, ²⁰¹⁷, Figure 1B), who recognized three genera of lasiurine bats instead of the single genus Lasiurus, have been accepted by many authors (^{Alurralde et al. 2017}; ^{Amador et al. 2016}; ^{Best and Hunt 2020}; ^{Schmidly and Bradley 2016}; ^{Decker et al. 2020}; ^{Espinosa-Martínez et al. 2016}; ^{Geluso and Bogan 2018}; ^{Gimenez and Giannini 2017}; ^{Krejsa et al. 2020}; ^{Lew and Lim 2019}; ^{Tirira 2018}); however, others have not followed our taxonomy including ^{Upham et al. (2019}) who report the most extensive phylogeny of mammals. ^{Ziegler et al. (2016}), ^{Novaes et al. (2018}), and ^{Teta (2019}) have argued that changes to the taxonomy of Lasiurini are not warranted because Lasiurus was a monophyletic genus. They all suggest recognizing the red, yellow, and hoary bats as different subgenera of Lasiurus (Figure 1C).

Figure 1 Phylogenetic relationships of the major groups of lasiurine bats according to Baird et al. (2015, 2017). The taxonomy proposed by Baird et al. (2015, 2017) in B differentiates the phylogenetic node that demarcates the tribe from the node, or nodes, that demarcates the genus or genera. In the taxonomies shown in A and C, node 1 demarcates both the tribe and genus. 

Table 1 shows the estimated Time to Most Recent Common Ancestor (TMRCA) for 24 genera, five tribes, and four subfamilies of vespertilionid bats. Figure 3 shows a histogram of these estimates and illustrates how the subdivision of Lasiurini into three genera changes the monotypic Lasiurus from an outlier among genera to three genera having approximately average vespertilionid TMRCAs.

Discussion

^{Teta (2019}) asserted that a “main goal of the zoological nomenclature is to promote nomenclatural stability.” Nomenclatural stability also is a key point raised by ^{Ziegler et al. (2016}) and ^{Novaes et al. (2018}). We agree that it is good to strive for stability; however, we reject the idea that stability is the main purpose of nomenclature. The history of nomenclature is one of frequent change and this is true for lasiurine bats (Figure 2). With the current rapid rate of change in technology, including increasing computing capacity and the increased speed and decreased cost of genetic sequencing, it is inevitable that change will be rapid in our understanding of biosystematics and consequently in taxonomy. Zoological nomenclature is a powerful tool that should be utilized to its maximum descriptive potential, not simply conserved because of the status quo and to make the lives of scientists, and online searches, easier. When used to its potential, nomenclature should convey evolutionary relationships, diversity, divergence, and the potential to clarify conservation priorities. ^{Teta (2019)} promoted the use of subgenera in cases such as Lasiurini where a monophyletic genus contains multiple distinct lineages. This was considered herein, but for reasons outlined below it is apparent that full generic recognition of the lineages is warranted due to their genetic and morphological distinction, and to keep them consistent with other vespertilionid genera. Having a distinct tribe comprised of a single genus containing three morphologically diverse lineages that are genetically as distinct and old as other vespertilionid genera does not adequately reflect the true biodiversity or history of the tribe.

Figure 2 Studies showing the various generic and subgeneric taxonomic relationships of the Tribe Lasiurini prior to Baird et al. (2015, 2017). 

^{Novaes et al. (2018}) also argued against the splitting of the vespertilionid genus Lasiurus into three distinct genera (Aeorestes, Dasypterus, and Lasiurus), as proposed by ^{Baird et al. (2015}, ²⁰¹⁷). The first argument made by ^{Novaes et al. (2018)} is that genetic distance, and divergence times calculated from genetic distance, is not a useful character for defining genera because it is not comparable between different groups. They cite the examples of primate genera, Tarsius and Homo, which have vastly different divergence times from their respective sister taxa. Although this is certainly true for widely divergent taxa, ^{Baird et al. (2015}, ²⁰¹⁷, and references therein) stated that divergence times and genetic distance among genera within Vespertilionidae are generally consistent, with the notable exception of Lasiurus.

Splitting the clades within Lasiurus into three distinct genera aligns their divergence times and genetic distances more closely to most other splits within Vespertilionidae. Examples of vastly different divergence times in other mammalian taxa, such as those in Primates cited by ^{Novaes et al. (2018}), are irrelevant to the discussion of lasiurine taxonomy because those discrepancies are not seen to the same degree in vespertilionids. Evidence of the relative consistency in age of vespertilionid taxa can be seen in Figure 3 and Table 1. The TMRCA of Lasiurini, and of Lasiurus prior to our subdivision of the genus, is 20 MYA making this the oldest genus in the family (tied with Kerivoula of the monotypic subfamily Kerivoulinae). The three genera recognized by ^{Baird et al. (2015}, ²⁰¹⁷) have TMRCA estimates that range from 9 to 12 MYA and which are close to the mean TMRCA of 12.58 calculated for the 24 genera in Figure 3 and Table 1 (not including the monotypic Kerivoulinae, Kerivoula, and Lasiurini, Lasiurus, which are clear outliers).

The equivalence of taxonomic categories at the same rank and the decision as to what appropriate taxonomic level a group of species should be included have long been issues that have vexed taxonomists. ^{Schaefer (1976}, p. 2) recommended a clear and logical solution: “Should a natural group of species, clearly distinct from other groups, be treated as a genus, tribe, or family? The answer of course appears to lie in comparing the group with other genera, tribes, and families in the higher category to which the group in question belongs.” This is precisely what was done by ^{Baird et al. (2015}, ²⁰¹⁷).

Secondly, ^{Novaes et al. (2018}) disputed that morphological differences distinguish Aeorestes, Dasypterus, and Lasiurus and thus stated that the taxonomic arrangement of Baird et al. “attributes unnecessary weight to clades whose phenotypic distinction is merely superficial”. The most obvious morphological difference among the three genera is pelage color, which has given rise to the colloquial names of each group: the hoary, yellow, and red bats. Aeorestes (the hoary bats) are characterized by grayish pelage, Dasypterus (yellow bats) by yellowish pelage, and Lasiurus (red bats) by reddish pelage. Moreover, members of Dasypterus have only one premolar on each side of the upper jaw (^{Hall and Jones 1961}) compared to two in Lasiurus and Aeorestes. Aeorestes species generally are larger in size than the other two genera. Other diagnostic characters of Aeorestes include multiple unique dental and skeletal features (^{Shump and Shump 1982}). Although ^{Handley (1960}) considered the differences among the three “species groups” (as he referred to them) insufficient to warrant their distinction as different genera, he nonetheless provided a table of distinguishing characteristics for each (see Table 3 in ^{Handley 1960}).

Figure 3. Observations of time to most recent common ancestors (TMRCA) of vespertilionid taxa. The data used to create this figure are detailed in Table 1. 

The final point made by ^{Novaes et al. (2018}) was that vernacular names “cannot be used as an argument to take (sic) taxonomic decisions.” There is a reason that the vernacular names red bats, hoary bats, and yellow bats exist: it is because the names reflect morphological distinction among the groups. We do not support changing taxonomy simply because vernacular names exist; it is the basis for their existence that supports the taxonomic change.

Characters in support of taxonomic revision.^{Novaes et al. (2018}) admitted that separating a monophyletic group into different genera can be supported if “well-marked phenotypic discontinuities are detected among them” and “if supported by a suite of consistent characters, preferably from multiple datasets.” But their bias towards the use of morphology in making taxonomic decisions is evident from the first part of the title of their paper. “Separation of monophyletic groups into distinct genera should consider phenotypic discontinuities.” We argue that the split of Lasiurini into three distinct genera is supported by both morphology and, more importantly, genetics. Lasiurine bats are among the most easily distinguishable group of vespertilionids, even by those who are not experts in vespertilionid morphology. It is trivial to identify most species to genus from some distance away, without having to measure skull characters, etc. Moreover, given the new taxonomy proposed by ^{Baird et al. (2015)}, it would be useful to have a morphological revision of the tribe that could provide morphological diagnoses of the genera and, importantly, include species that were unavailable to ^{Baird et al. (2015)} for genetic analysis. As for the “suite of consistent characters” required by ^{Novaes et al. (2018)} to define genera, ^{Baird et al. (2015}, ²⁰¹⁷) certainly have defined a suite of characters that consistently group the three genera into reciprocally monophyletic groups and can be used to define them. Those characters are genetic data from multiple mtDNA and nuclear loci.

Despite the historical importance of morphological characters in taxonomy, we are now on the cusp of the genomics age in mammalogy (^{Baird et al. 2019}), and molecular markers, not morphology, are the current gold standard for conducting phylogenetic analysis. Since it is generally agreed that taxonomy must reflect phylogeny, then it follows that genetic characters are the most useful in taxonomy as well. But morphology will continue to be used to diagnose taxa because those characters are useful for identification of specimens. In fact, morphological characters are useful to diagnose living and extinct taxonomic groups because they are characters with a genetic basis. If they were not, they could not be used. Nonetheless, they are not the best genetic-based characters available to us. DNA sequences are easily understood, discrete, and quantifiable.

^{Novaes et al. (2018}) “agree that clades may be separated into different genera if well-marked phenotypic discontinuities are detected among them.” Their condemnation of our decision to recognize three well differentiated lasiurine genera is based in part on a vague and undefinable metric of morphological divergence. What exactly is a well-marked phenotypic discontinuity? No such scale exists. They went on to say that “the decision will always be arbitrary.” Our decision, however, was not arbitrary. It was based on our use of genetics to estimate percent sequence divergences and divergence time estimates based on multiple genetic loci. These metrics were compared to other genera of vespertilionid bats and determined to be comparable. Consequently, we concluded that not only are Aeorestes, Dasypterus, and Lasiurus easy to distinguish based on morphology, there is also no doubt that they are distinct, highly divergent, and easily definable groups based on genetics.

Taxonomy and Phylogeny. Prior to ^{Baird et al. (2015)}, the tribe Lasiurini was a monophyletic group consisting of a single, monophyletic genus. Therefore, the same node on the tree defined both a tribe and a genus (Node 1 in Figure 1A). The only taxonomic status given to the divergent clades within the tribe were the subgenus Lasiurus (node 2 in Figure 1A) and subgenus Dasypterus (node 5 in Figure 1A); however, the subgeneric taxonomy was not recognized by many authors (Figure 2). Even the authors who recognized subgenera would not normally use those names when referencing a particular group of lasiurines; they only used the genus name, which references node 1 in Figure 1A.

The taxonomic change proposed by ^{Baird et al. (2015)} maintains the monophyletic tribe defined by one node (node 1 in Figure 1B), but shifts the generic levels to their own nodes on the tree (nodes 3-5 in Figure 1B). This arrangement maximizes the use of taxonomy to describe the variation present in the lasiurine tree. The old arrangement (Figure 1A) did not assign taxonomic status to several important nodes in the lasiurine phylogeny. In using the ^{Baird et al. (2015)} taxonomy, researchers can now reference specific parts of the lasiurine tree by utilizing a genus name.

Critics of the ^{Baird et al. (2015)} taxonomy, including ^{Ziegler et al. (2016}), ^{Novaes et al. (2018}), and ^{Teta (2019}), all supported the following taxonomy: Tribe Lasiurini and Genus Lasiurus, with subgenera Lasiurus, Aeorestes, and Dasypterus. As shown in Figure 1C, this arrangement does not help resolve the issue of having the Tribe and Genus both defined by the same node on the phylogeny (i.e., both the genus and the tribe still reference node 1 in Figure 1C). Although the subgenera would clarify the specific part of the phylogeny, researchers do not generally use subgenera, and therefore this proposed taxonomy does not meet our criteria of maximizing the potential of taxonomy to reflect phylogenetic divergence. Naming each of the major nodes within the lasiurine phylogeny is the most powerful way to utilize taxonomy.

The fact that the three critical papers all suggested that the recognition of three subgenera of Lasiurus would be appropriate, indicates that all three studies do in fact recognize these lineages as being distinct, and that we are only arguing about the taxonomic level at which they should be recognized. ^{Patterson and Norris (2016}) faced a similar dilemma in that all chipmunks were placed in a monophyletic genus Tamias, but it included three distinct lineages recognized as subgenera. ^{Patterson and Norris (2016)} elevated the subgenera to genera based on 1) the degree of genetic differentiation among the subgenera being comparable to other genera of ground squirrels, 2) the chipmunk lineages are older than the ground squirrel lineages as indicated in the fossil record, and 3) morphological distinction. Thus, we now have three genera of chipmunks despite the original genus Tamias being monophyletic. As with the lasiurine bats addressed here, the key metric to determine if the lineages represent genera or subgenera was the level of genetic differentiation and TMRCA of the lineages as compared to others in their taxonomic group.

^{Novaes et al. (2018}), ^{Teta (2019}), and ^{Ziegler et al. (2016}) supported the status quo of recognizing a single, monophyletic genus, Lasiurus, within the tribe Lasiurini. They suggested that the names Aeorestes and Dasypterus should be used as subgenera (Figure 1C). Although we agree that splitting a monophyletic genus should not be done without strong evidence, we do not think that sub-generic taxonomy is the best way to handle the lasiurine situation. If, as ^{Ziegler et al. (2016)}, ^{Novaes et al. (2018)}, and ^{Teta (2019)} suggested, one were to recognize Aeorestes and Dasypterus as subgenera of Lasiurus, it would not solve the problem of having a single node on the tree defining both a genus and a tribe, thus still rendering the taxonomy ambiguous and lacking in resolution (Figure 1C). Additionally, and more importantly, it does not reflect the true degree of differentiation of these three highly distinctive lineages that are comparable in age and genetic distance to average vespertilionid generic lineages (Table 1, Figure 3).

Finally, ^{Novaes et al. (2018}) concluded their paper by wrongly suggesting that the taxonomic arrangement of ^{Baird et al. (2015}, ²⁰¹⁷) has not been widely accepted. Multiple papers cited above have followed ^{Baird et al. (2015)}, but it is especially worthwhile to note ^{Amador et al. (2016}). This is a comprehensive molecular systematic review of bats based on a study of 796 species using 9 nuclear and mitochondrial genetic markers. They report data for 270 species from 48 genera of vespertilionid bats. Notably, they report no subgenera.

We encourage other mammalogists to view taxonomy as we have outlined here. It should be a tool used to convey evolutionary relationships and biodiversity. A taxonomic arrangement is a hypothesis; therefore, it is subject to change when better data are available. The taxonomy of ^{Baird et al. (2015}, ²⁰¹⁷) is a hypothesis that will be tested in future studies as better methods and more samples become available. Future studies may support or falsify the hypothesis of three genera, but as it stands now, ^{Baird et al. (2015}, ²⁰¹⁷) is the most complete and modern analysis of phylogeny and taxonomy of lasiurine bats ever conducted. Thus, their phylogenetic and taxonomic hypotheses should be accepted pending studies presenting data and analyses that falsify them.