Designation of a new family group name, Tonzidae fam. nov., for the genus Tonza (Lepidoptera, Yponomeutoidea), based on immature stages of Tonza citrorrhoa

1 urn:lsid:zoobank.org:author:5A3A485F-95E4-4CA0-9693-E996C9B2A9D2 2 urn:lsid:zoobank.org:author:56AE5FE9-77C9-4FA4-9F4A-31A9450A9F3F 3 urn:lsid:zoobank.org:author:F6265053-D7FD-4497-910E-AA9B0F2AB568 4 urn:lsid:zoobank.org:author:84687C58-E0A2-4706-9A17-20B589243993 5 urn:lsid:zoobank.org:author:50B2933B-ABF8-4C6A-B8F8-A68FE823998F 6 urn:lsid:zoobank.org:author:ADC0E307-B009-4BE5-A3BD-EEFBF43A132D


Introduction
Tonza Walker, 1864 is a small genus of moths from Australia and Asia (Sri Lanka, Taiwan, Japan, and Solomon Is.). This genus was originally assigned to Plutellidae Guenée, 1845 (Walker 1864). Since then, the majority of authors (e.g., Common 1990;Heppner 1992) has followed the original assignment. However, no previous publications provided convincing evidence supporting the plutellid association of Tonza.
Recently, Kobayashi et al. (2015) recorded T. citrorrhoa Meyrick, 1905 from Japan and described the wing venation and genital structures of Tonza for the fi rst time. Furthermore, they reevaluated the previously suggested association of Tonza with Plutellidae. The diagnostic characters of Tonza include: antenna almost the same length as forewing; forewing with only three radial plus radial sector veins; in the male genitalia, both uncal processes and the socii present; in the female genitalia, lamella antevaginalis strongly sclerotized. These morphological features of Tonza along with absence of ocelli clearly disputed its plutellid association and seemed to associate Tonza with the yponomeutoid family Attevidae Bruand, 1850 . However, a formal family transfer was pending until the immature stages of Tonza became known.
In this paper, the immature stages of Tonza are described for the fi rst time for T. citrorrhoa from Japan with photographs and drawings: the later instar larvae, pupae and life history. We propose a new family group name, Tonzidae Kobayashi & Sohn fam. nov., for the genus Tonza and discuss its relationships with other yponomeutoid families, based on larval and pupal morphology.

Material and methods
Larvae were collected from spinnings among the leaves and branches of Putranjiva matsumurae Koidz. by the third author (M. Kimura) in April, 2016 on Yonaguni Is., Okinawa Prefecture. Those were taken to the lab and reared in plastic cups (420 ml: 129 mm in top diameter and 60 mm in depth) containing wet cotton on the bottom. The cups were maintained in the laboratory at 20 ± 5ºC and at 13-16L:8-12D photoperiodic conditions. Some samples were preserved in 99% ethanol for DNA sequencing. Emerged adult specimens were preserved in the Osaka Prefecture University (OPU). Some samples of the larvae, the pupae and the adults were dried and sputter-coated with a 60:40 mixture of gold-palladium for scanning electron microscope (SEM). SEM photographs were taken using a Hitachi SU1510 (Hitachi Ltd., Tokyo, Japan) with a lanthanum hexaboride (LaB6) cathode source at an accelerating voltage of 15 kV. For genital dissections, the whole abdomen was removed, macerated for 3-4 min in 10% aqueous KOH, and brushed in 70% ethanol to remove residual scales and soft parts. Genitalia were then stained in acetocarmine for 1-2 h, dehydrated in a series of 70-100% ethanol and mounted in Canada balsam on a glass slide. For the observations of larval structure, the alcohol-fi xed larvae were macerated for 5-6 min in 10% aqueous KOH and the right lateral side of the body was opened up, followed by removal of residual inner soft parts in 70% ethanol. The larval surface skin was then stained and mounted using the same methods as the genitalia. Larvae of Atteva aurea Fitch, 1856 were collected by the fourth author (JCS) and Scythropia crataegella Linnaeus, 1767 was obtained from Dr. Ian Sims (UK). Nomenclature for genitalia follows Sohn & Nishida (2011). Nomenclature for larval characters follows Stehr (1987). Scientifi c names of plants follow the Plant List (www.theplantlist.org).

Molecular analysis
The DNA barcode region (658 bp), a part of the mitochondrial cytochrome c oxidase subunit I (COI) gene, was chosen for sequencing. This marker has been popularly used for inferring the relationships among closely related moth species and populations (Brown et al. 1994) One emerged adult of T. citrorrhoa was preserved in 99% ethanol for DNA sequencing (voucher no. SK-097). Total genomic DNA was extracted from the mid-and hind legs. Polymerase chain reaction (PCR) was carried out to obtain the DNA barcode region, using the primer LCO1490 (fwd) (5'-ggt caa caa atc ata aag ata ttg g-3') and HCO2198 (rev) (5'taa act tca ggg tga cca aaa aat ca-3') (Folmer et al. 1994) in a thermal cycler C1000 (Bio-Rad, Hercules City, CA, USA) under the following conditions: initial 120 s denaturation at 94°C, and 39 alternating cycles of 15 s at 94°C for denaturation, 30 s at 52°C for annealing, and 60 s at 72°C for extension. PCR products were purifi ed, after that they were sequenced using an automated DNA sequencer (ABI Prism 3100; Applied Biosystems, Foster City, CA, USA). The sequences were aligned using MEGA6 (Tamura et al. 2013) andGeneious v. 10.1.3 (Kearse et al. 2012). The aligned data were analyzed using RAxML installed in the CIPRES portal (https://www.phylo.org/). The outgroup sequences were obtained from GenBank (https://www.ncbi.nlm.nih.gov/genbank/) and accession numbers are shown in Table 1

Diagnosis
A . Maxillary palpi three-segmented; ocelli and chaetosema absent ( Fig. 8C-D); antennae slightly longer than or same length as forewing (Fig. 1); forewings with slightly protruding apex and tornus; forewing termen oblique or concave; only two radial sector veins present, RS1 on apex and RS2 on termen Fig. 1F); in the male genitalia Fig. 2A-D), uncus small with a pair of long processes; socii with long terminal setae; valva elongate with several small spines and plate arising from middle to base of valva; in the female genitalia ; Fig. 2E), lamella antevaginalis sclerotized, covering sternite VIII; antrum slender; inception of ductus seminalis at the middle of corpus bursae (after Kobayashi et al. 2015).

Diagnosis
For the adult: see description for the family and see Kobayashi et al. (2015).
A checklist and distribution of the species of the genus Tonza 1. Tonza purella Walker, 1864 Distribution Australia (Queensland; Moreton Bay, type locality; specimens including types in NHMUK and data on BOLD; Bold Index Number (BIN), BOLD:AAY2226).
T (F . 6A-B). T1 shield indistinct, yellow in colouration, marked with dark brown patches. Jugular gland (adenosma) absent, but like structure present ventrally on T1 (Fig. 4F), anterior to the legs and antero-medial to the long setae SV1 and SV2; the structure not eversible and a tubular or sacciform gland at inner side of body not found. D1 and D2 approximated on T2 and T3; SD1 and SD2 approximated on T2 and T3; L-group trisetose, L1 and L2 on the same pinacula on T1 and separated on T2 and T3; SV1 and SV2 on the same pinacula on T1 and on separate pinacula on T2 and T3; Seta V1 present on T1, absent T2 and T3. Thoracic legs pale ochreous to brown in colouration; pale brown claws elongate, slightly curved to inner side (Figs 4F, 6B).

Life history
The detailed biology of this species is unknown. The third author (Kimura) observed a number of late instar larvae on the young leaves of Putranjiva matsumurae. The larva is a leaf webber tying together several leaves loosely with silk threads (Fig. 3A-C). The full grown larva is suspended from the tree by a silk lifeline spun out from the head spinnerets. Pupation takes place at the intersection of some cross silken threads (Fig. 3H). A number of the mature larvae occurred on tall trees of the hostplant in Yonaguni Is., Okinawa Pref. (Fig. 3A). Young larvae were not found in our study. Given that no larval mine was observed, the young larvae are probably external feeders like the later instars.

Remarks
The resting posture of the adult moth with head end lowered and abdomen lifted is similar to certain Argyresthiidae, Ypsolophidae and several genera in Yponomeutidae (Fig. 1B). However, when at repose in nature the suspension may be diff erent and rather unusual involving only the prothoracic and mesothoracic legs in contact with the lower surface of a leaf with the wing and body tending to hang vertically below a leaf. In this position, the antennae relatively long with respect to wing length may accentuate a potential false head-like appearance.

Molecular analysis
The COI DNA barcoding region (COI-5P) was sequenced from a Japanese specimen of Tonza. We performed sequence comparison to check species variation for T. citrorrhoa, using the BOLD Identifi cation System (IDS) from the BOLD website (http://www.barcodinglife.org/) [accessed 1 Jul. 2016]. The DNA barcode sequence of T. citrorrhoa (Fig. 2, sample ID: SK-97) was clearly distinguished from that of T. purella registered in BOLD with more than 3% pairwise divergence (Fig. 2) and to a sample from Madagascar belonging to BIN BOLD:ACU1104 by 2.91% pairwise divergence. The nearest neighbour of T. citrorrhoa in the BOLD database which is 0.35% pairwise divergent is a Taiwanese conspecifi c specimen (Fig. 2), which belongs to the same BIN (BOLD:ACX8102).

Morphology
Historically, the genus Tonza had been associated with Yponomeutidae (Philpott 1927, who also noted the four-segmented maxillary palp as an exception; Common 1966) or Plutellidae (Common 1990;Heppner 1992), until Kobayashi et al. (2015) discussed its similarities with and possible association to Attevidae based essentially on male genitalic features: the presence of uncal processes and socii and the shape of the pleural lobes in the abdomen. Kobayashi et al. (2015) also mentioned that the genitalia of Stachyotis Meyrick, 1905 are similar to those of Tonza. Stachyotis is another genus whose plutellid association was defi ned merely by superfi cial similarity (Sohn 2014). However, Table 2 shows many characteristics in the genitalia are homoplastic among yponomeutoids and thus their phylogenetic value needs to be carefully evaluated. Immature stages as well as molecular data may help in defi ning the systematic position of Tonza. Such information has not been available or analysed until the present study.
We examined the immature stages and some additional adult characters of Tonza and compared it with other yponomeutoids (Table 2, Fig. 14). Our examination supports the designation of a new family group name, Tonzidae fam. nov., for Tonza. Tonzidae fam. nov. is clearly distinguished from other families of Yponomeutoidea by larval prolegs absent on A5 and A6. Larvae of other families possess prolegs on A3-A6, while several families have reduced prolegs. For example, the prolegs on A3 and A6 of Bedelliidae are shorter than those on A4 and A5 (Fig. 16A). The larval proleg modifi cations are curious and might relate to locomotion on silk threads as mentioned above. Mature larvae of Tonza have a very long seta MD1 on the head. This character also occurs in Yponomeutidae (Yponomeutinae and Saridoscelinae) and Scythropiidae Friese, 1966. Yponomeutine larvae possess a cervical gland (adenosma) on T1 that has been presumed to be an exocrine organ producing a trail-pheromone, used for orientation and homing (Povel & Beckers 1982). This character is not found in other yponomeutoid groups, including Tonza. However, the larvae of Tonza and Bedellia (Bedelliidae) possess a gland-like structure ventrally on T1 (Figs 4F, 16B, D). Furthermore, an endocrine gland-like structure is found in the larvae of Bedellia inside the body (Fig. 16), but not in the larvae of Tonza.
The pupa of Tonza is also similar to that of Bedelliidae (Table 2, Fig. 15K) in having a triangular single frontal process on the head (Figs 10A-D, 16F, 17A-E). The larva of Lyonetia Hübner, 1825 (Lyonetiidae Stainton, 1854) forms a hammock-shaped cocoon similar to Tonza and Bedelliidae, but the pupa diff ers from the latter two in possessing two frontal processes on the head and straight setae on the cremaster (Ahn et al. 2004;Kobayashi 2015).

Molecular phylogeny
Unexpectedly, our ML tree showed a close relationship between Tonza and Glyphipterix (Glyphipterigidae: Glyphipteriginae) (Fig. 13). The multigene analyses by Sohn et al. (2013) supported the family Glyphipterigidae comprising three subfamilies, Glyphipteriginae, Acrolepiinae Heinemann , 1870, and Orthoteliinae Herrich-Schaff er, 1857. Our COI phylogeny failed to recover such relationships. Tonza may be related with Glyphipteriginae and in that case, suggest that it is a member of Glyphipterigidae. We, however, could not fi nd any morphological characteristic associating Tonza with Glyphipterigidae.
Our phylogeny based on a single gene was not suffi cient to resolve the association of Tonza with Glyphipterigidae. This issue needs to be explored with a much larger molecular dataset that is currently under construction.
The larval and pupal morphological features of Tonza exclude its associations with all existing families in Yponomeutoidea. Therefore, we propose a new family group name, Tonzidae fam. nov., to accommodate these unique characteristics of Tonza. This hypothesis is now available for further phylogenetic testing using new data and any other existing members of this new taxon need to be searched for and examined. The family-level rank itself needs to be confi rmed with a much larger molecular dataset that is currently under construction.

Host association
The hostplant relationship of Tonza uncovered from this study is novel for Yponomeutoidea. Putranjivaceae seems to be the only non-Brassicaceae plant lineage containing mustard oil glucosides, a convergence with Brassicaceae Burnett (Hall et al. 2002). It would therefore be interesting to check other members of the Indo-Australian genus Putranjiva Wall., as well as the more widely distributed Drypetes Vahl for larvae of Tonza, and also other members of this family (e.g., Lingelsheimia Pax). The new relationship gives a strong clue to discover early stages elsewhere in the World. Since Tonza is now removed from Plutellidae (a group like Pierinae Swainson, 1820 that has radiated on Brassicaceae), the ability to feed on plants with mustard oils (iso-thiocyanates: Puntambekar 1950) could be a symplesiomorphy or a convergence.