Review of Eyeless Pseudosinella Schäffer (Collembola, Entomobryidae, and Lepidocyrtinae) from Brazilian Caves

Herein, eyeless Pseudosinella species from Brazilian caves are reviewed, including the description of 23 new species, new records plus additional notes on the descriptions of P. ambigua Zeppelini, Brito, and Lima and of P. guanhaensis Zeppelini, Brito, and Lima. We also provide an identification key to 27 eyeless species recorded from Brazil. To organize the 26 Brazilian eyeless taxa analyzed in this work, we organize them in apparently artificial groups: 11 species have one larger tooth on the unguiculus outer lamella (petterseni group); one presents unguiculus outer lamella smooth or serrated (never with a larger tooth), with 9 held prelabral chaetae undivided and the last 6 held prelabral chaetae bifurcated. The Brazilian species of eyeless Pseudosinella herein described present a remarkably conservate dorsal chaetotaxy; therefore, the main diagnostic characters are related to other features like prelabral, labral, and ventral head chaetotaxy and empodial complex morphology. In addition, our study suggests that Brazilian caves possibly shelter a great diversity of Pseudosinella taxa, several of them potentially cave dependent.


Antennal Chaetae
Antennal segments with about 11 types of chaetae (see Figure 1B); a-h as sens and i-k as chaetae: type a short and thin, apically acuminate and with median region swollen, present subapically on Ant IV; type b short and thin, apically acuminate, present on Ant I and IV; type c finger-shaped gently pointed in apex, present densely on Ant I-IV; type d finger-shaped, present on Ant III-IV; type e conical, present on Ant II-IV; type f ball, present on Ant III; type g guard sens, present on Ant III; type h ms-like sens present on Ant IV; type i smooth present on Ant I-IV; type j weakly ciliate, present on Ant I-IV; type k heavily ciliate, present on Ant I-IV.

Tibiotarsal Modified Chaetae
Tibiotarsus I-III with 7 types of modified mac present on proximal half (as in Figure 10E-G); tibiotarsus I with 1 mac on inner side, tibiotarsus II-III with 1 mac on inner, anterior and outer side ( Figure 10E Type VII ( Figure 60E): Heavily ciliate and finger-shaped, abruptly dilated in the apex [43] (p. 47, Figure 169).
Legs. Subcoxa I with 3 chaetae and 2 psp; subcoxa II with "a" row of 6 chaetae, "p" row with 2 chaetae and 1 psp; subcoxa III with one row of 6 chaetae and 2 posterior psp ( Figure 15A-C). Trochanteral organ with about 6 spine-like chaetae ( Figure 15D). Tibiotarsus I-III formula with 1, 3, and 3 mac type II; tibiotarsus I-III internally with some scale-like chaetae ( Figure 15E). Unguis outer side with paired teeth developed in hook-shaped and on proximal one fourth; inner side with slender lamella and 3 teeth, basal pair unequal, b.p. tooth surpassing the m.t. apex, m.t. on a little more than the unguis distal half and gently larger to b.a. tooth length, with a.t. absent. Unguiculus with all lamellae smooth and acuminate (ai, ae, pi, and pe), except pe with outer tooth ( Figure 15F-G); ratio unguis:unguiculus in holotype = 1:0.55. Tibiotarsal smooth chaeta about 0.84 smaller than unguiculus; tenent hair acuminate and about 0.38 smaller than unguis outer edge.
Etymology. The species was named after its type locality, "Brumadinho" municipality, where a mining disaster killed 252 people (there still 18 missing) on 25 January 2019.
Etymology. Refers to presence of abundant ciliate chaetae on mouthparts of the new species ( Figure 19F).
Legs. Subcoxa I with 3 chaetae and 2 psp; subcoxa II with "a" row of 6 chaetae, "p" row with 2-3 chaetae and 2 psp; subcoxa III with one row of 6 chaetae and 2 posterior psp. Trochanteral organ with about 8 spine-like chaetae ( Figure 24A). Tibiotarsus I-III formula with 1, 3, and 3 mac type II (as Figure 10E-G). Unguis outer side with paired teeth developed in hook-shaped and on proximal one third; inner side with wide lamella and 3 teeth, basal pair unequal, b.p. tooth reaches the m.t. apex and with 1 smaller split tooth posteriorly, m.t. on distal one third and subequal to b.a. tooth length, a.t. absent. Unguiculus with all lamellae smooth and acuminate (ai, ae, pi, and pe), except ai truncate distally and pe with outer tooth ( Figure 24B,C); ratio unguis:unguiculus in holotype = 1:0.43. Tibiotarsal smooth chaeta about 1.06 longer than unguiculus; tenent hair acuminate and about 0.51 smaller than unguis outer edge.
Furcula. Only with ciliate chaetae and scales. Manubrium ventrally with 3 subapical and about 9 distal scales ( Figure 24E); manubrial plate with 4 ciliate chaetae (2 inner mac) and 2 psp (as Figure 11C). Mucro teeth subequal in size, basal spine reaches the apex of basal tooth (as Figure 15K) Etymology. Refers to presence of filaments similar to spines on anterior and median labral chaetae of the new species ( Figure 22B).
Remarks. The presence of filaments in labral chaetae and head M2 mac absent make P. labruspinata sp. nov. similar to P. keni sp. nov. and P. labiociliata sp. nov., but P. labruspinata sp. nov. easily differs from these species by the modifications in a1 and b4 appendages of labial papilla B and Th II-III with 2 (p3 and p5) and 2 (p2-3) mac, respectively. Due to this last feature, P. labruspinata sp. nov. is more similar to P. brumadinhoensis sp. nov. More comparisons are presented in the remarks of this late species and Tables 1 and 3.
Etymology. Refers to type locality of the new species, "Pará" State.
Etymology. Refers to "Serpentina" in the name of the type locality of the new species.
Etymology. Refers to labral a1-2 chaetae of the new species which resemble the horns of a bull or an ox (from Latim: taurus) ( Figure 31B).
Remarks. Pseudosinella taurina sp. nov. resembles most P. certa Christiansen and Bellinger, 1980 and P. unimacrochaetosa sp. nov. by head "M" and "S" series devoid of mac, Th II-Abd III with 1, 0, 0, 1, and 0 central mac, and unguis basal teeth unequal (see Tables 1-3). However, both new species differ from P. certa by basomedian and basolateral labial fields with A1-4 unilaterally or completely ciliate in P. taurina sp. nov and P. unimacrochaetosa sp. nov., respectively (smooth in P. certa). They also differ by Abd IV with 3 central mac (2 in P. certa), tenent hair capitate (acuminate in P. certa), unguis with m.t. (absent in P. certa), and mucro teeth subequal and with basal spine reaching the apex of basal tooth, while P. certa has basal tooth smaller and basal spine exceding the apex of it [5,7,20]. Both new species also resemble P. josemarii Soto-Adames, 2010 by the presence of 1 mac on Th II, but they differ in Pa5 mac absent and Th III devoid of mac, while in P. josemarii this chaeta is present (see Table 2) [20]. Finally, P. taurina sp. nov. differs from P. unimacrochaetosa sp. nov. by labral a1-2 thicker, similar to horns (no such morphology in P. unimacrochaetosa sp. nov.), labial papilla E with l.p. acuminate (finger-shaped in P. unimacrochaetosa sp. nov.), and collophore anteriorly with 4 chaetae and lateral flap with 5 ciliate chaetae (7 anterior and 6 smooth chaetae on lateral flap in P. unimacrochaetosa sp. nov.). They also differ by 3 unguis inner teeth, with b.p. undivided in P. taurina sp. nov. while in P. unimacrochaetosa sp. nov. there are 4 unguis inner teeth, with b.p. divided.
Etymology. Refers to the presence of a single macrochaeta on Th II of the new species (from Latim: uni-one) ( Figure 35A).
Legs. Subcoxa I with "p" row with 6 chaetae and 2 psp; subcoxa II with "a" row with 5 chaetae, "p" row with 7 chaetae and 2 psp; subcoxa III with one row of 7 plus 2 anterior chaetae and 1-2 posterior psp ( Figure 45A-C). Trochanteral organ with about 14 spine-like chaetae ( Figure 45D). Tibiotarsus I-III formula with 1, 2, and 1 mac type VI ( Figure 45E). Unguis outer side with paired teeth straight, not developed and on little more than distal half; inner side with wide lamella and 4 teeth, basal pair (b.a. and b.p.) subequal and not reaching the m.t. apex, m.t. on distal one third and gently larger to basal teeth, a.t. smaller on distal one sixth. Unguiculus with all lamellae smooth and acuminate (ai, ae, pi, and pe), except pe serrate ( Figure 45F); ratio unguis:unguiculus in holotype = 1:0.66. Tibiotarsal smooth chaeta about 1.02 longer than unguiculus; tenent hair capitate and about 0.85 smaller than unguis outer edge.
Etymology. Refers to type locality of new species, "Ceará" State.
Etymology. Refers to type locality of new species, "Diamantina" municipality.
Etymology. Refers to type locality of new species, "Mariana" municipality, where a mining disaster killed 18 people on 5 November 2015.
Etymology. The species was named after the nickname of our dear friend, Collembola researcher Dr. Nerivânia Nunes Godeiro.
Etymology. Refers to small size of this new specie (from Latin: pusill-very small).
Legs. Subcoxa I with 6 chaetae and 2 psp; subcoxa II with "a" row of 5 chaetae, "p" row with 5-8 chaetae and 2 psp; subcoxa III with one row of 8 chaetae and 2 posterior psp ( Figure 63A-C). Trochanteral organ with about 12 spine-like chaetae ( Figure 63D). Tibiotarsus I-III formula with 1, 3, and 3 mac type I ( Figure 63E). Unguis outer side with paired teeth straight and not developed on distal half, nearside the inner basal pair; inner side with wide lamella and 4 teeth, basal pair slightly unequal, b.p. larger than b.a. and reaching m.t. apex, m.t. on distal one third and smaller than basal teeth, a.t. minute on distal one fifth. Unguiculus with all lamellae smooth and acuminate (ai, ae, pi, and pe), except pe serrate ( Figure 63F); ratio unguis:unguiculus in holotype = 1.46:1. Tibiotarsal smooth chaeta subequal or slightly longer than unguiculus lenght; tenent hair capitate and about 0.78 smaller than unguis outer edge.
Etymology. Refers to mix of morphological features, resembling both P. ambigua and P. parambigua sp. nov. (from Greek: chimeraa hybrid monster).

Discussion
There are now 27 eyeless Pseudosinella species recorded from Brazil, of which 11 are from the group with one larger tooth on unguiculus outer lamella (petterseni group), one presents unguiculus outer lamella toothless, 10 hold prelabral chaetae undivided, while the other six hold prelabral chaetae bifurcate.
At this time, it is difficult to state if prelabral and labral chaetae modifications observed in part of the new species could hold phylogenetic signal and delimite Neotropical species group, since such distinct morphology is unknown to most species from North America [6,7,48]. Furthermore, these modifications are not unique to Pseudosinella, since bifurcate prelabral chaetae also occur in other genera of Entomobryoidea such as Amazhomidia, Entomobrya, Orchesella, Plumachaetas, and Pseudodicranocentrus [59][60][61][62][63]. In this same sense, filaments on labral a1 and a2 chaetae as well as ciliation on labral p0-2 chaetae and b.c. of maxillary palp are modifications also seen in Lepidocyrtus species [43,[64][65][66]. These labral modifications may indicate that Brazilian Pseudosinella could have arisen from different branches of Lepidocyrtus or that such morphology emerged independently within Pseudosinella, and similarities concerning such features with Lepidocyrtus taxa are due to parallelism [17,21,22].
The eyes reduction and presence of a large outer tooth on unguiculus pe lamella (petterseni group) seen in some Pseudosinella species are troglomorphic characteristics quite certainly homoplastic [17], since they are also present in other Entomobryoidea genera such as Alloscopus, Coecobrya, Heteromurtrella, Sinella, and Cyphoderinae as a whole [13,[67][68][69][70]. Other aggravating factors are that an unguiculus outer tooth is present in species with and without eyes besides interspecific inconsistency in the number of eyes, as observed in at least nine American species [7]. Although there are indications that eyes reduction is homoplastic within Lepidocyrtinae, unguiculus morphology (e.g., presence/absence of external tooth) should be further investigated phylogenetically to reveal if it may hold significance in splitting Pseudosinella into natural subgroups.
It is also not clear if the dorsal pattern of macrochaetotaxy is useful to support ingroups of Pseudosinella since the same patterns occur in different species regardless of the shape of the unguis and unguiculus [24,32,50,52,56]. Even though such patterns are similar among remarkably different species, Holarctic taxa tend to have a larger number of macrochaetae [7,52,56] while the Holotropical species hold a smaller number of them, as seen in most of the Brazilian new species. Curiously, the dorsal macrochaetae reduction in tropical species was also reported to other Entomobryidae genera like Seira and Lepidocyrtus [16,36,71] and it may represent some trend in the adaptation to tropical areas, possibly related to high temperature.

Conclusions
Due to this difficulty in validating species groups using morphology in disjoint regions without a proper phylogenetic study, for now, we cannot assume if they are natural groups or that the diagnostic characters that we are using are homoplastic for troglomorphic springtails [17]. Therefore, a phylogenetic analysis is in need to test whether the genus is really polyphyletic, and if so, how many branches of "Pseudosinella" there are and which species groups (inner taxa) are valid. Also, it is important to understand which morphological features can support the lineages of Pseudosinella. So far, phylogenetic studies performed with Lepidocyrtinae had few representatives of the genus and were not intended to verify its internal relationships [2,[21][22][23].